delirium in the acute general medical settingdro.deakin.edu.au/eserv/du:30082790/cull-incident...iii...
TRANSCRIPT
Incident Delirium in the Acute General Medical Setting
by
Emily Jane Cull
BN Nursing (Hons), RN
Submitted in fulfilment of the requirements for the degree of
Doctor of Philosophy
Deakin University
July, 2015
iii
Dedication
Gone yet not forgottenAlthough we are apart
Your spirit lives within meForever in my heart
This thesis is dedicated to my grandmother Jane Stone. A beautiful, gentle person
who inspired me to be a better person and always encouraged me to pursue my
dreams.
iv
Acknowledgements
The completion of my doctoral thesis would not have been possible without the
support and guidance from many people. Firstly, I would like to acknowledge
Professor Bridie Kent. Thank you for encouraging me to pursue my doctoral studies
after completing my honours degree and for taking me on again as a PhD candidate.
Thank you for seeing my vision and helping it come into reality.
To my principal supervisor Professor Alison Hutchinson and my associate supervisor
Associate Professor Nikki Phillips, I am so thankful for your guidance, expertise and
encouragement throughout my candidature. Your input into the study has been so
very valuable and I am very appreciative of your contributions making it what it is
today. I am so grateful and consider it a privilege to have you both as professional
role models.
To Dr Mohammadreza Mohebbi, biostatistician at Deakin University whose valuable
contribution has helped me to undertake the statistical analysis for this thesis. I am
very thankful for his contribution and for helping me to understand and analyse the
statistics. Also thank you to Dr John Reynolds for his assistance early on in the
development of the research study. Thanks also to the librarians at Deakin
University for their valuable assistance in developing a search strategy for the
systematic review.
Special thank you to my husband Joshua for always encouraging and supporting me,
especially when I doubted myself. Thank you to my parents, John and Lexie and my
siblings for their love and support and for always being there when I needed them.
I would also like to acknowledge the Nurses Memorial Centre and the Rosemary
Norman Foundation for providing me the ‘Babe’ Norman Scholarship to undertake
my studies. Without the support of the scholarship, undertaking the research would
not have been possible.
v
Professional editor, Sophie Dougall, provided copyediting and proofreading
services, according to the guidelines laid out in the university endorsed national
‘Guidelines for editing research theses’.
vi
List of publications, presentations and awards arising from thisthesis
Publications
Cull, EJ, Kent, B, Phillips, NM and Mistarz, R 2012, Protocol: Risk factors for incident
delirium in acute medical in patients: a systematic review, JBI library of systematic
reviews, vol. 10, no. 56, Supplement, pp. 1 12.
Cull, EJ, Kent, B, Phillips, NM and Mistarz, R 2013, ‘Risk factors for incident delirium
in acute medical in patients: a systematic review’, JBI database of systematic
reviews and implementation reports, vol. 11, no. 5, pp. 62 111, University of
Adelaide, S. A.
Joanna Briggs Institute 2013, ‘Risk Factors for incident delirium in acute medical in
patients: best practice information sheet’, Vol 17, No 6.
Presentations
Oral
Cull, EJ 2012, 2013, 2014, ‘Incident delirium in the acute general medical setting’.
Oral presentation, School of Nursing and Midwifery Research School, Deakin
University, Melbourne, Australia.
Cull, EJ, Kent, B, Phillips, NM and Mistarz, R 2014, ‘Risk factors for incident delirium
in acute medical in patients: a systematic review’. Oral presentation, DECLARED
Delirium Clinical and Research Day, Melbourne Brain Centre.
Poster
Cull, EJ, Kent, B, Phillips, NM and Mistarz, R 2014, ‘Risk factors for incident delirium
in acute medical in patients: a systematic review’. Poster presentation, American
Delirium Society 4th Annual Meeting, Baltimore, Maryland.
vii
Cull, EJ, Phillips, NM, Mohebbi, M, Hutchinson, AM 2014, ‘Incident delirium in acute
medical in patients: A case control study’. Poster presentation. Research Forum.
Melbourne.
Awards
Recipient of the Deakin University, School of Nursing and Midwifery Scholarship
Award in 2012. This is an award for outstanding research scholarship to a student
enrolled in the PhD Doctoral program at the School of Nursing and Midwifery,
Deakin University, Melbourne, Australia
Recipient of an Australian Postgraduate Award (2012 2013). This is a stipend
scholarship for doctoral studies from the Australian Government.
Recipient of the ‘Babe’ Norman Scholarship (2013 2015). This is a three year
scholarship available to nurses enrolled in a higher research degree at an Australian
university from the Nurses Memorial Centre.
viii
Table of Contents
Dedication ............................................................................................................ iii
Acknowledgements .............................................................................................. iv
List of publications, presentations and awards arising from this thesis ................. vi
Table of Contents................................................................................................ viii
List of Tables ........................................................................................................ xv
List of Figures.................................................................................................... xviii
List of Abbreviations ........................................................................................... xix
Abstract .............................................................................................................. xxi
Chapter 1 Introduction.........................................................................................1
1.1 Background and problem ................................................................................... 1
1.2 Aims of the research........................................................................................... 6
1.3 Overview of the study......................................................................................... 6
1.4 Research Questions ............................................................................................ 8
1.5 Significance of the study..................................................................................... 8
1.6 Outline of the thesis.......................................................................................... 10
Chapter 2 Literature Review ..............................................................................11
2.1 Introduction ...................................................................................................... 11
2.2 Delirium definition and diagnosis.....................................................................12
2.3 Incidence and prevalence ................................................................................. 16
2.4 Complications of delirium................................................................................. 18
2.5 Causes and risk factors ..................................................................................... 20
2.5.1 Predisposing and precipitating risk factors ...............................................22
2.5.1.1 Predisposing risk factors ....................................................................22
2.5.1.2 Precipitating risk factors.....................................................................25
2.5.1.3 Relationship between predisposing and precipitating risk factors....25
2.5.2 Predictive model for delirium ...................................................................27
2.6 Delirium management guidelines ....................................................................28
2.6.1 Delirium guidelines in Australia ................................................................30
2.6.2 Delirium prevention and management strategies ....................................34
ix
2.6.2.1 Prevention of delirium .......................................................................35
2.6.2.2 Management of delirium ...................................................................38
2.7 Conclusion......................................................................................................... 41
Chapter 3 Methods ............................................................................................43
3.1 Introduction ...................................................................................................... 43
3.2 Research purpose ............................................................................................. 43
3.3 Research phases ............................................................................................... 44
3.4 Phase 1 Systematic review .............................................................................45
3.4.1 Research design......................................................................................... 45
3.4.2 Systematic review process ........................................................................45
3.4.2.1 Framing the research question ..........................................................46
3.4.2.2 Developing the aim of the systematic review....................................47
3.4.2.3 Developing the protocol.....................................................................47
3.4.2.4 Inclusion and exclusion criteria..........................................................47
3.4.2.5 Search strategy................................................................................... 49
3.4.2.6 Identifying potential studies ..............................................................51
3.4.2.7 Assessing the methodological quality of studies ...............................51
3.4.2.8 Extracting the data .............................................................................51
3.4.2.9 Data synthesis .................................................................................... 52
3.4.3 Ethical considerations ............................................................................... 52
3.5 Phase 2 Case control study: retrospective audit ............................................53
3.5.1 Research design......................................................................................... 53
3.5.2 Aims of the case control study..................................................................53
3.5.3 The research setting .................................................................................. 54
3.5.4 Study population ....................................................................................... 55
3.5.5 Sample and sampling approach ................................................................55
3.5.4.1 Patients with delirium ........................................................................55
3.5.4.2 Patients with no delirium (control group)..........................................56
3.5.4.3 Sample size power calculations..........................................................57
3.5.4.4 Sample size ......................................................................................... 59
3.5.6 Data collection........................................................................................... 60
3.5.7 Data analysis.............................................................................................. 62
3.5.8 Ethical considerations ............................................................................... 66
x
3.5.7.1 Consent............................................................................................... 66
3.5.7.2 Confidentiality .................................................................................... 66
3.5.7.3 Data storage ....................................................................................... 66
3.6 Phase 3 Delirium management survey...........................................................68
3.6.1 Research design......................................................................................... 68
3.6.2 Aim of the survey ...................................................................................... 68
3.6.3 Research setting and participants.............................................................68
3.6.4 Study sample ............................................................................................. 68
3.6.5 Data collection........................................................................................... 69
3.6.5.1 Data collection instrument.................................................................69
3.6.5.2 Data collection procedure..................................................................69
3.6.6 Data analysis.............................................................................................. 70
3.6.7 Ethical considerations ............................................................................... 70
3.6.7.1 Consent............................................................................................... 70
3.6.7.2 Confidentiality .................................................................................... 71
3.6.7.3 Data storage ....................................................................................... 71
3.7 Summary of methods ....................................................................................... 72
Chapter 4 Results ...............................................................................................73
4.1 Introduction ...................................................................................................... 73
4.2 Phase 1 Systematic review results..................................................................73
4.2.1 Search results ............................................................................................ 73
4.2.1.1. Excluded studies................................................................................ 75
4.2.2 Study characteristics ................................................................................. 75
4.2.3 Methodological quality .............................................................................77
4.2.4 Results of included studies........................................................................79
4.2.5 Meta analysis ............................................................................................ 80
4.2.5.1 Dementia ............................................................................................ 80
4.2.5.2 Functional impairment.......................................................................81
4.2.5.3 Male gender ....................................................................................... 82
4.2.5.4 Visual impairment .............................................................................. 83
4.2.5.5 Pneumonia ......................................................................................... 84
4.2.6 Narrative synthesis.................................................................................... 87
4.2.6.1 Cognitive impairment.........................................................................87
xi
4.2.6.2 Depression.......................................................................................... 89
4.2.6.3 Years of education.............................................................................. 89
4.2.6.4 Blood urea nitrogen (BUN).................................................................90
4.2.6.5 Use of indwelling urinary catheter.....................................................91
4.2.6.6 Severe illness ...................................................................................... 91
4.2.7 Systematic review results summary..........................................................93
4.3 Phase 2 – Retrospective case control study results..........................................94
4.3.1 Identification of cases ............................................................................... 94
4.3.2 Identification of controls ...........................................................................95
4.3.3 Characteristics of patients.........................................................................96
4.3.3.1 Reason for admission .........................................................................99
4.3.4 Risk factors for incident delirium ............................................................101
4.3.4.1 Predisposing risk factors ..................................................................101
4.3.4.1.1 Predisposing risk factors for possible delirium group compared
with control group ......................................................................103
4.3.4.1.2 Age............................................................................................. 105
4.3.4.1.3 Logistic regression for predisposing factors..............................105
4.3.4.2 Precipitating risk factors...................................................................109
4.3.4.2.1 Precipitating risk factors for possible delirium group compared
with control group ......................................................................110
4.3.4.2.2 Blood test results.......................................................................111
4.3.4.2.3 Logistic regression for precipitating factors..............................112
4.3.5 Outcomes for patients............................................................................. 114
4.3.5.1 Residence on admission and discharge destination ........................114
4.3.5.2 Comparison of outcomes for patients .............................................117
4.3.5.2.1 Deaths........................................................................................ 118
4.3.5.2.2 Change in functioning and continence......................................118
4.3.5.2.3 Discharge destination................................................................119
4.3.5.2.4 Falls............................................................................................ 119
4.3.5.2.5 Pressure injuries ........................................................................121
4.3.5.2.6 Medical emergency team and code grey calls ..........................121
4.3.5.2.7 Length of stay ............................................................................121
4.3.5.2.8 Logistic regression of outcomes for patients ............................122
xii
4.3.6 Delirium detection and management .....................................................124
4.3.6.1 Monitoring of cognition ...................................................................124
4.3.6.2 Delirium risk assessment..................................................................124
4.3.6.3 Time of delirium development and documentation........................125
4.3.6.4 Words used to describe delirium .....................................................125
4.3.6.5 Recognition of delirium signs ...........................................................128
4.3.6.6 Diagnosis of delirium........................................................................129
4.3.6.7 Medication management.................................................................129
4.3.6.7.1 Benzodiazepine medications.....................................................130
4.3.6.8 Non pharmacological management strategies................................132
4.3.6.8.1 Physical restraints......................................................................133
4.3.6.9 Delirium prevention strategies ........................................................133
4.3.6.10 Follow up care after discharge.......................................................135
4.3.7 Case control study results summary.......................................................136
4.4 Phase 3 Delirium management survey results .............................................137
4.4.1 Hospital characteristics and participation rates .....................................137
4.4.2 Delirium management policies ...............................................................139
4.4.2.1 Delirium management policy ...........................................................140
4.4.2.2 Awareness of the Clinical Practice Guidelines for the Management of
Delirium in Older People (Clinical Epidemiology and Health Service
Evaluation Unit and Delirium Clinical Guidelines Expert Working Group
2006). .................................................................................................. 140
4.4.2.3 Delirium policy developed using the Clinical Practice Guidelines for
the Management of Delirium in Older People as a guide ..................140
4.4.2.4 Screening and diagnosing delirium ..................................................140
4.4.2.5 Documentation of the diagnosis ......................................................141
4.4.2.6 Cognitive assessment on admission.................................................141
4.4.2.7 Risk factor assessment .....................................................................142
4.4.2.8 Pharmacological management policy ..............................................142
4.4.2.9 Medical review of patients with delirium ........................................143
4.4.2.10 Barriers to implementation or development of policies ...............144
4.4.3 Survey results summary ..........................................................................145
4.5 Conclusion....................................................................................................... 145
xiii
Chapter 5 Discussion........................................................................................ 146
5.1 Introduction .................................................................................................... 146
5.2 Risk factors ..................................................................................................... 146
5.2.1 Predisposing risk factors..........................................................................147
5.2.1.1 Dementia .......................................................................................... 147
5.2.1.2 Cognitive impairment.......................................................................147
5.2.1.3 Advanced age ...................................................................................148
5.2.1.4 Functional impairment.....................................................................150
5.2.1.5 Gender.............................................................................................. 151
5.2.1.6 Sensory impairment .........................................................................151
5.2.1.7 Level of education ............................................................................153
5.2.1.8 Illness severity and co morbidity .....................................................153
5.2.1.9 Depression........................................................................................ 154
5.2.1.10 Previous delirium............................................................................156
5.2.2 Precipitating risk factors..........................................................................158
5.2.3 Risk factor prediction assessment...........................................................159
5.3 Assessing patient cognition on admission......................................................160
5.4 Delirium recognition and diagnosis................................................................162
5.5 Delirium medication management.................................................................165
5.6 Delirium management strategies...................................................................167
5.7 Outcomes for patients .................................................................................... 170
5.7.1 Discharge to a care facility ......................................................................171
5.7.2 Falls.......................................................................................................... 171
5.7.3 Decline in functioning and incontinence.................................................173
5.7.4 Increased length of stay ..........................................................................174
5.8 Delirium follow up .......................................................................................... 175
5.9 Difficulties with implementing a delirium management policy .....................176
5.10 Strengths and limitations of the research ....................................................177
Chapter 6 Conclusion and Recommendations ..................................................182
6.1 Introduction .................................................................................................... 182
6.2 Aims of the research....................................................................................... 183
6.2.1 Aim one ................................................................................................... 184
6.2.2 Aim two ................................................................................................... 185
xiv
6.2.3 Aim three................................................................................................. 187
6.3 Recommendations and implications for practice and policy..........................190
6.4 Recommendations for future research...........................................................192
6.5 Conclusion....................................................................................................... 193
References ......................................................................................................... 194
Appendices ........................................................................................................ 224
Appendix 1 – Systematic review protocol.............................................................225
Appendix 2 – Search strategy for systematic review............................................231
Appendix 3 – Joanna Briggs Institute Critical Appraisal Instrument ....................241
Appendix 4 – Joanna Briggs Institute Data Extraction Tool .................................242
Appendix 5 – Case control study research questions, hypothesis and statistical
tests................................................................................................. 244
Appendix 6 – Case control study audit tool..........................................................248
Appendix 7 Case control study audit tool on iPad application Tap Forms ........254
Appendix 8 – Ethical approval letters...................................................................261
Appendix 9 – Delirium management survey ........................................................265
Appendix 10 – Email to Directors of Nursing and participants ............................267
Appendix 11 Plain language statements............................................................269
Director of Nursing.........................................................................269
Participant/potential respondent..................................................271
Plain language statement ..............................................................273
Appendix 12 – Reasons for study exclusion from systematic review ...................275
Appendix 13 – Joanna Briggs Institute individual study critical appraisal results for
included studies ..............................................................................280
xv
List of Tables
Table 1. Diagnostic criteria for delirium in the DSM 5 and the DSM 4......................14
Table 2. Precipitating and predisposing risk factors for delirium...............................26
Table 3. Delirium prevention strategies .....................................................................39
Table 4. Examples of research questions and hypothesis ..........................................54
Table 5. Approximate bed numbers and hospital admissions per hospital ...............55
Table 6. Sample size calculations to detect significance in dementia between cases
and controls................................................................................................. 58
Table 7. Sample size calculations to detect significance in benzodiazepine use
between cases and controls ........................................................................59
Table 8. Sample size calculations for differences in length of stay ............................59
Table 9. Research questions and statistical tests used ..............................................65
Table 10. Database search results for systematic review ..........................................74
Table 11. Characteristics of Studies Included in the Systematic Review....................76
Table 12. Risk factors examined in included studies in systematic review ................79
Table 13. Average ages of patients with and without delirium .................................86
Table 14. Average MMSE scores and t test results for patients with and without
delirium........................................................................................................ 87
Table 15. Delirium incidence vs. no delirium for cognitive impairment tests ............88
Table 16. Average years of education for patients with and without delirium .........90
Table 17. Average scores for patients using the Charslon co morbidity index with
and without delirium................................................................................... 92
Table 18. Admission characteristics of patients in the case, control and possible
delirium groups ........................................................................................... 98
Table 19. Primary and secondary admission diagnosis............................................100
Table 20. Chi square test results for possible predisposing risk factors for delirium
................................................................................................................... 102
Table 21. Chi square statistic results of predisposing risk factors for patients for
possible delirium and control groups. .......................................................104
Table 22. Initial logistic regression results for predisposing factors of incident
delirium using cases and controls .............................................................106
xvi
Table 23. Final logistic regression model for predisposing factors of incident delirium
using cases and controls............................................................................107
Table 24. Logistic regression for predisposing factors (including age >80) using cases
and controls............................................................................................... 108
Table 25. Final logistic regression model for predisposing factors (not including age)
for cases and controls................................................................................109
Table 26. Chi square statistic results of precipitating risk factors for delirium
comparing cases and controls...................................................................110
Table 27. Chi square statistic results of precipitating risk factors for delirium
comparing possible delirium with control group ......................................111
Table 28. Blood tests results comparisons between delirium and control groups...112
Table 29. Logistic regression results for precipitating factors of incident delirium
using cases and controls............................................................................113
Table 30. Final logistic regression model for precipitating factors of incident delirium
................................................................................................................... 113
Table 31. Residence on admission compared to discharged destination of patients in
case and control groups ............................................................................116
Table 32. Comparison of outcomes for cases and control patients .........................117
Table 33. Comparison of outcomes for patients with possible delirium and control
group ......................................................................................................... 118
Table 34. Initial logistic regression model for patient outcomes using cases and
controls...................................................................................................... 122
Table 35. Final logistic regression model for patient outcomes using cases and
controls...................................................................................................... 123
Table 36. Description words used for the first symptoms of delirium and possible
delirium...................................................................................................... 126
Table 37. Benzodiazepine medications patients were taking prior to admission....131
Table 38. Newly prescribed benzodiazepines administered to patients during
admission................................................................................................... 132
Table 39. Non pharmacological management strategies documented for the
management of patients with delirium ....................................................133
Table 40. Environmental prevention strategies documented for patients in case,
control and possible delirium group..........................................................134
xvii
Table 41. Clinical prevention strategies documented for patients in case, control and
possible delirium group .............................................................................135
Table 42. Hospital network data for the approximate number of patient admissions
per year and estimates of delirium incidence ...........................................138
Table 43. Medications and doses recommended for patients with agitation and
aggression in a pharmacological management policy .............................143
xviii
List of Figures
Figure 1. Phases of the research project ............................................................. 7 & 44
Figure 2. Process of a systematic review....................................................................45
Figure 3. Flow diagram of the stages of searching ....................................................74
Figure 4. Critical appraisal of included studies...........................................................78
Figure 5. Forest plot of the impact of dementia on development of incident delirium
..................................................................................................................... 81
Figure 6. Forest plot of the impact of functional impairment on the development of
incident delirium.......................................................................................... 82
Figure 7. Forest plot of the impact of male gender on development of incident
delirium........................................................................................................ 83
Figure 8. Forest plot of the impact of visual impairment on development of incident
delirium........................................................................................................ 84
Figure 9. Forest plot of the impact of pneumonia on the development of incident
delirium ....................................................................................................... 85
Figure 10. Forest plot of impact of age > 80 years on the development of incident
delirium........................................................................................................ 86
Figure 11. Identification of cases and controls...........................................................96
Figure 12. Number of falls patients had during admission ......................................120
Figure 13. Time of day falls occurred during admission...........................................120
Figure 14. Percentage of description words used during admission for patients with
delirium and possible delirium ..................................................................127
Figure 15. Health professional who first documented signs of delirium..................128
Figure 16. Health professional who first documented signs of possible delirium....128
Figure 17. Number of antipsychotics prescribed for patients with delirium............130
Figure 18. Delirium management policies and guidelines in both public networks (n =
7) and private hospitals (n = 5). (*Clinical Practice Guidelines for the
Management of Delirium in Older People) ...............................................139
xix
List of Abbreviations
ACSQHC: Australian Commission on Safety and Quality in Health Care
AF: Atrial Fibrillation
ADL: Activities of Daily Living
APA: American Psychiatric Association
APACHE: Acute Physiology and Chronic Health Evaluation
BDRS: Blessed Dementia Rating Scale
BUN: Blood Urea Nitrogen
CAM: Confusion Assessment Method
CCF: Congestive Cardiac Failure
CCI: Charlson Co morbidity Index
CCU: Critical or Coronary Care Unit
CI: Confidence Interval
CPF: Clinical Patient Folder
DI: Delirium Index
DRS: Delirium Rating Scale
DRS R98: Revised Delirium Rating Scale
DSM: Diagnostic and Statistical Manual of Mental Disorders
EEG: Electroencephalograph
GDS: Geriatric Depression Scale
HELP: Hospital Elder Life Program
HLC: High Level Care
ICD: International Classification of Diseases
ICU: Intensive Care Unit
IDC: Indwelling Catheter
JBI: Joanna Briggs Institute
LLC: Low Level Care
MI: Myocardial Infarction
MET: Medical Emergency Team
MMSE: Mini Mental State Exam
NCD: Neuro cognitive Disorder
NEECHAM: The Neelon and Champagne Confusion Scale
xx
NHMRC: National Health and Medical Research Council
PICO: Population, Interest, Comparison, and Outcome
RAMU: Rapid Access Medical Unit
RUDAS: Rowland Universal Dementia Assessment Scale
SD: Standard Deviation
SPMSQ: Short Portable Mental Status Questionnaire
SPSS: Statistical Package for the Social Sciences
TCP: Transitional Care Program
TIA: Transient Ischaemic Attack
UK: United Kingdom
USA: United States of America
WHO: World Health Organisation
xxi
Abstract
Background and aim
Delirium is a serious neuro cognitive disorder that affects many people admitted to
hospital. The overall aim of this research was to add to the current evidence
regarding risk factors, characteristics and management of acute general medical
patients who develop incident delirium during hospitalisation.
Method
Phase 1 of the research involved undertaking a systematic review to identify risk
factors for incident delirium in the acute medical setting. A meta analysis was
conducted for a number of the identified risk factors for incident delirium.
Phase 2 of the research was a retrospective case control study of patients with
incident delirium admitted to a general medical setting at a public health care
organisation in Melbourne, Australia. Extracted data included information regarding
risk factors for delirium, cognitive assessments undertaken, diagnosis of delirium,
medication management, and prevention/management strategies related to
delirium.
Phase 3 utilised survey methodology to identify delirium management policies used
in acute hospitals in Melbourne, Australia. The survey was distributed electronically
to hospital representatives and completed either via telephone or email. All public
and private health care organisations in Melbourne were invited to participate.
Results
Risk factors for incident delirium identified in both the systematic review and the
case control study were dementia, cognitive impairment and functional
impairment. Additional risk factors identified in the case control study were history
of delirium and fracture on admission. Patients with delirium in the case control
study were more likely to fall, trigger an emergency call for aggression management
(code grey), have functional decline and be discharged to a continuing care facility.
In the control group, 42 patients had documented evidence of possible delirium
that was undiagnosed. In Phase 3 of the research, a representative from each of five
private hospitals and seven public health care organisations completed the survey.
xxii
Three of the private and four of the public and hospitals were reported to have a
delirium management policy.
Conclusions
A variety of risk factors can contribute to the development of incident delirium in
the acute general medical setting. In the case control study, risk factors for delirium
were not identified on admission and delirium episodes were not recognised,
leading to lengthy delays in diagnosis of delirium. Haphazard and widely varying
treatment strategies were identified that resulted in poor outcomes overall for
patients. Despite the availability of the locally developed Clinical Practice Guidelines
for the Management of Delirium in Older People, management of delirium varied
across health care settings in Melbourne, Australia. Therefore, a consistent and
clear approach to delirium assessment, prevention and management is needed.
1
Chapter 1 IntroductionDelirium, an acute disorder of attention and cognition, is a serious condition that is
widespread among patients across all health care settings. It is unrivalled by any
other disorder in its ability to penetrate across all clinical areas (O'Hanlon et al.
2014). The incidence of delirium can be used as a measure for the quality of care
and patient safety during hospitalisation (Fong, Tulebaev & Inouye 2009). Although
not always preventable, there is evidence to suggest that incident delirium can be
prevented for some patients if appropriate recognition and prevention strategies
are implemented (Anderson 2005; Cerejeira & Taylor 2011; Cole, Primeau &
McCusker 1996). In the Australian setting, limited research has been conducted to
investigate the implementation of these strategies. As a result, research was
indicated to investigate the recognition, and management of incident delirium in
this acute health care setting. This chapter outlines the background and problem of
delirium, significance of the research as well as the purpose and aims of the study.
Furthermore, this chapter provides a brief overview of the methods used to address
the overall aims of the research and serve as an outline for the structure of the
thesis.
1.1 Background and problem
Delirium is a complicated condition that has the potential to develop in patients
during a period of hospitalisation. As defined by the American Psychiatric
Association (APA) (2013, p. 599) delirium is “a disturbance of attention or
awareness that is accompanied by a change in baseline cognition that cannot be
better explained by a pre existing or evolving neurocognitive disorder (NCD)”.
Delirium involves a generalised disturbance of all higher cognitive functions (Franco
et al. 2013, p. 228 9) and signs of delirium can be considered as impairments in
three core domains. The cognitive domain includes: impaired orientation, attention,
short term memory, long term memory and visuospatial perception. The higher
level thinking domain includes: language and thought process impairment, and the
2
circadian domain includes: sleep wake cycle and motor behaviour disturbance
(Franco et al. 2013, p. 228 9; Harrington & Vardi 2014, p. 19). Another identifying
feature of delirium is that its onset is usually short, presenting within a few hours or
days of admission (Inouye 2006).
The syndrome of delirium can present as different clinical subtypes. These subtypes
are described according to the signs of delirium the patient exhibits, in terms of
their level of alertness or cognition (Gofton 2011; Saxena & Lawley 2009; Twedell
2005). Hyperactive delirium is often characterised by psychomotor hyperactivity
being easily distracted, confusion, hallucinations and delusions. Hypoactive delirium
however is characterised by symptoms such as reduced alertness, lethargy,
decreased motivation, confusion and sluggishness (Boettger & Breitbart 2011;
Gofton 2011; O'Keeffe 1999; Saxena & Lawley 2009). The mixed subtype of delirium
is a combination of these symptoms.
Delirium may also be referred to as prevalent delirium, which is present on
admission to hospital, or incident delirium, which develops during hospitalisation
(Gofton 2011). Incident delirium is an important concern for health professionals as
it has potential to be prevented in hospitalised patients. Incidence rates of delirium
can range between 3 – 29% in acute medical settings (Siddiqi, House & Holmes
2006). Recognition of delirium risk and implementation of prevention strategies
may help to reduce the incidence of delirium in health care settings (Inouye 2000).
It is therefore important to focus research on incident delirium, as health
professionals have the ability prevent its development. However, often the
fluctuating nature of delirium and the varying clinical presentations, in terms of
presenting signs, makes recognition difficult (Wong et al. 2010). As a result, despite
the clinical importance of delirium, it often goes unrecognised by clinicians working
at the bedside (Wong et al. 2010). This may in part be due to the characteristics of
delirium being closely related to dementia; unfortunately therefore patients with
delirium may be misdiagnosed in the clinical setting. In these cases, often the
underlying cause of the delirium remains untreated, resulting in deleterious effects
for the patient. Poor recognition may also be the result of a lack of knowledge
3
about altered mental state in older people and poor education about delirium
(McCrow, Sullivan & Beattie 2014). In order to implement effective prevention
strategies, clinicians should be aware of the key diagnostic features of delirium and
the patient and environmental factors that can increase the likelihood of developing
delirium. As the predominant bedside caregivers nurses therefore play a vital role in
delirium prevention. They should be able to recognise subtle changes in patient
behaviours and aid in the recognition of delirium in acute care settings. A limited
amount of research has been undertaken to investigate the identification and
recognition of delirium in Australian hospitals.
A link has been established between delirium and a number of complications and
poor outcomes for patients who develop the syndrome. The development of
delirium has the potential to dramatically complicate hospitalisation for a patient
(McCusker, Cole, Dendukuri, et al. 2001) and may also increase their risk of
experiencing many long term adverse problems. Research has shown that the
development of delirium can result in functional decline, including reduced ability
to perform daily living tasks, an increased likelihood of complications associated
with longer hospital stays, such as increased falls and development of pressure
sores, increased risk of admission to a care facility post discharge and also high
mortality and morbidity rates (Twedell 2005). Little is known, however, about the
outcomes for patients in the Australian setting who develop delirium.
Delirium is a complex syndrome and may be due to the interaction of physiological
illness and pre existing risk factors (Elie et al. 1998; Elmore 2002; Inouye,
Westendorp & Saczynski 2014). Risk factors known to increase a patient’s
vulnerability to delirium include dementia, cognitive impairment, severe medical
illness, functional impairment, hearing or vision impairment and some
environmental factors, such as use of restraints (Bjoro 2008; Dahl, Rønning &
Thommessen 2010; Dasgupta & Hillier 2010; Davis et al. 2012; de Castro et al. 2014;
Elie et al. 1998; Faezah, Zhang & Yin 2008; Inouye 1999). Knowledge and awareness
of risk factors for delirium enable nurses to be proactive in implementing
prevention strategies (Voyer et al. 2007). Thus, the identification of risk factors
4
during a patient’s admission to hospital is an essential step in implementing
strategies to reduce the incidence of delirium.
Risk prediction models such as those developed by Inouye et al. (1993) are one
strategy that can be used to identify patients at high risk of developing delirium.
Causes of delirium and the presence of risk factors can vary significantly between
patients. As a result it is difficult to predict development of delirium among patients
in a particular setting without first investigating causes and risk factors commonly
occurring in patients admitted to that setting. Differentiation between hospital
settings (for example medical or surgical) can assist in the development of risk
prediction models specific to particular settings. Patients in the medical setting may
be particularly vulnerable to developing delirium because they are often afflicted
with a number of co morbidities, increasing their risk of developing delirium. A
systematic review of current research examining specifically risk factors for delirium
in themedical patient population did not exist. There was a need, therefore, to
ascertain risk factors for delirium specific to medical patients.
The most effective strategy for delirium management is the implementation of
appropriate risk screening and prevention interventions (Inouye 2006; Inouye,
Schlesinger & Lydon 1999). Studies have shown that proactive strategies in acute
hospital settings to implement risk screening and prevention interventions have
resulted in a reduced incidence of delirium (Inouye et al. 2000; Lundstrom et al.
2005; Marcantonio 2007). Furthermore, when active and appropriate management
has been implemented, the severity of incident delirium has been reduced and the
duration shortened (Inouye et al. 1999). The most effective prevention
interventions include daily orientation to surroundings, cognitively stimulating
activities, non pharmacological sleep enhancers such as relaxation recordings, early
ambulation and motion exercises, encouraging use of sensory aids, maintaining
fluid balance and nutrition and avoidance of benzodiazepines (Anderson 2005;
Fong, Tulebaev & Inouye 2009; Inouye 2000).
5
In order to inform the appropriate management of delirium in acute care settings,
the Clinical Practice Guidelines for the Management of Delirium in Older People
were developed in Australia (Clinical Epidemiology and Health Service Evaluation
Unit and Delirium Clinical Guidelines Expert Working Group 2006). Implementation
of delirium management guidelines is an important strategy to encourage the
appropriate diagnosis and management of delirium. These guidelines, along with
other initiatives such as the Delirium Care Pathways (Australian Health Ministers
Advisory Council [AHMAC] 2011), provide a series of recommendations to assist in
the assessment of delirium as well as its prevention and management. Since the
development of these guidelines, a limited number of studies have been
undertaken to investigate the implementation of guideline recommendations in
hospitals throughout Australia (these studies will be discussed further in the
literature review chapter). However, as guidelines alone do not improve care,
strategies and processes for implementation and organisational change, such as
education, are necessary (Young & George 2003).
The management of delirium in hospital settings can be inconsistent (O'Hanlon et
al. 2014). There is currently limited data to show if and how any of these
management strategies have been implemented into health care organisations and
if they have been effective. Thus, there is limited data regarding how patients with
delirium are being treated in terms of addressing the cause of delirium, the
medications they are receiving, and the strategies being incorporated into their care
in the Australian health care setting. This information is necessary in order to
provide greater understanding of the management strategies currently used in
hospitals and to indicate where education and delirium management plans need to
be implemented.
Many aspects of delirium research still need to be addressed. The development of
risk prediction models, diagnostic tools and management strategies have
contributed significantly to the understanding of delirium and the dramatic impact
it can have on hospitalised patients. While research highlights the benefits of
6
assessment tools and strategies, there remains a gap in knowledge in terms of the
integration of these into clinical practice in acute care settings.
1.2 Aims of the research
Following a review of the literature it was determined that the overall purpose of
this research was to add to the current evidence regarding the clinical risk factors,
characteristics and management of hospitalised generalmedical patients who
develop incident delirium in Australia. The specific aims of this research were to:
1. Systematically review the evidence for risk factors related to the
development of incident delirium in generalmedical patients.
2. Describe the characteristics of patients who develop incident delirium
during hospitalisation in Australia. These include the demographic
characteristics (age, gender, residency prior to admission, functional and
cognitive status prior to admission), potential risk factors (predisposing
and precipitating), and outcomes for patients including discharge
destination, length of stay in hospital and medication treatment.
3. Examine and describe the current state of delirium management in an
acute hospital setting in Victoria.
1.3 Overview of the study
A multi phase study design was used to address the overall aims of this research.
The research was conducted in three phases. Phase 1 involved a systematic review
of the risk factors for development of incident delirium in general medical patients
in the acute care setting. Findings from this systematic review contributed evidence
regarding risk factors for incident delirium in acute general medical patients.
Furthermore, the findings of this review informed the second phase of the research,
including risk factors that were likely to be observed in the medical in patient
population. Phase 2 of the research involved a retrospective case control clinical
7
audit of medical records of medical in patients who developed delirium during
hospitalisation as well as a control group of patients who did not develop delirium.
The case control study focused on identifying the characteristics of patients who
developed delirium, including evidence of the risk factors identified in the
systematic review and outcomes experienced by those patients, compared to a
control group. Data gathered as part of the case control study audit were also used
to examine how health professionals managed patients with delirium during
hospitalisation. The third and final phase of the study involved surveying key
informants from hospitals in Melbourne, Australia to identify the delirium
management policies and procedures that are currently in use in the respective
organisations. This information was used to determine if guidelines for health
professionals existed at a policy level to guide their practice for caring for patients
with delirium. Figure 1 illustrates the phases of the research and how each of the
aims of the individual studies are linked.
Phase 1 Phase 2 Phase 3
Figure 1. Phases of the research project
Identified Risk
Factors
Survey of deliriummanagement practices
in hospitals
AIM: Identify currentpractice in relation todelirium management
AIM: Compare currentpractice with clinicalpractice guidelines
AIM: Identify clinical characteristics ofpatients who develop delirium.Including medical related risk factorsfor delirium
Retrospective case controlclinical records audit of
medical patients
Systematic review ofrisk factors for
incident delirium inmedical patients
Clinicalrecordsaudit
Policy andProtocolSurvey
8
1.4 Research Questions
The research questions developed to address the aims of the research were:
1) What predisposing and precipitating risk factors are associated with the
development of incident delirium in the acute general medical setting?
2) What predisposing factors are likely to predict the possibility that a patient
will develop incident delirium?
3) What precipitating factors are likely to predict the possibility that a patient
will develop incident delirium?
4) Do patients with incident delirium have worse outcomes than patients with
no delirium during hospitalisation?
5) Which health professional first recognise and document the signs of
delirium?
6) Are patient’s cognitive status assessed on admission to hospital?
7) How do health professionals manage delirium during hospitalisation?
8) Is there a policy for the management of patients with delirium during
hospitalisation in hospitals in Melbourne?
1.5 Significance of the study
This research is significant nationally for the identification, prevention and
management of delirium in acute care organisations in the Australian setting.
However, this research also has international significance as similar guidelines and
management strategies are in place in other developed counties, all of which have
patients admitted to hospital who develop delirium. This research has contributed
to current evidence identifying risk factors for delirium in general medical patients
by undertaking a systematic review of the evidence. This knowledge can be used to
determine if patients in different hospital settings are exposed to different risk
factors. This information should inform the future development of risk prediction
models specific to the medical patient population. Little previous research
investigating risk factors for delirium has been conducted in the Australian setting.
9
The evidence regarding risk factors produced in the systematic review has added to
the existing body of knowledge and informed comparisons of risk factors for general
medical patients in the case control study.
Delirium is one of the most serious complications of hospitalisation; it is frequently
misdiagnosed, poorly managed or not recognised by health professionals. This
research has contributed evidence regarding how delirium is managed and
recognised by health professionals in an acute general medical setting. Examining
medical records of patients who develop incident delirium has provided evidence
regarding the process of screening, identification and prevention strategies
documented in the medical records. The research has also further highlighted areas
where education could be targeted in order to improve the recognition and
management of delirium in the clinical setting.
This is the first study to investigate the management of delirium in multiple acute
hospitals in Melbourne, Australia. The research has examined some of the key
strategies in the Clinical Practice Guidelines for the Management of Delirium in
Older People (Clinical Epidemiology and Health Service Evaluation Unit and Delirium
Clinical Guidelines Expert Working Group 2006) and examined if these have been
implemented widely into the clinical setting. In particular, this study has provided
information regarding the policies used by organisations in terms of screening,
recognition, medication management, and non pharmacological management
strategies for delirium. This provides new evidence about how the guidelines for
delirium management have been implemented into health care organisations and
highlights the areas where education should be implemented or where policies
need to improve or be more widely implemented.
This study has also provided evidence regarding the outcomes for patients who
develop incident delirium in the Australian setting. This contributes further
evidence that delirium can have detrimental effects for patients and that the
implementation of screening and prevention interventions is important.
10
1.6 Outline of the thesis
The thesis is presented in a number of chapters. This chapter has outlined the
background, purpose, aims, research questions and the methods used to undertake
the study. Chapter two, the literature review, provides a comprehensive review of
related research literature on delirium, including causes and risk factors,
recognition, and management. The review also identifies gaps in the evidence base
and situates the current study in the related literature. Chapter three, the methods
chapter, describes the methods used to collect and analyse the data. The focus of
the methods chapter is on the research design, participants, and data collection.
The results chapter (chapter four) presents the findings of all three phases of the
study, including the statistical analyses using forest plots, chi square, logistic
regression and descriptive statistics. In chapter five, a discussion of the findings and
limitations of the study is presented. Finally, in chapter six the concluding
statements of the study, recommendations for clinical practice and further research
are presented.
11
Chapter 2 Literature Review
2.1 Introduction
‘I was so afraid, so afraid, and I cried and shouted…. I was certainly very afraid…and
every time I got that injection I thought I got worse, and I was convinced that they
were going to kill me. I was very angry with the cleaner, doctor and everybody’
(Duppils & Wikblad 2007, p. 814). This is an account of a patient who has
experienced delirium. Delirium is one of the most serious complications that a
patient can experience during hospitalisation. Patients who develop delirium, not
only recall frightening experiences of their delirious episode but are also at risk of
longer hospital stays and becoming more functionally dependent (McCusker et al.
2003). Alarmingly though, delirium is often overlooked by physicians and nurses and
therefore continues to be a neglected clinical problem. Consequently, delirium is a
major contributor for poor patient outcomes and is therefore an important area of
health research.
In this chapter, a thorough review of delirium research will be presented, including
research from the Australian setting. Firstly, the definitions and diagnostic criteria
of delirium will be described in detail. Then evidence regarding prevalence and
incidence of delirium will be examined, as well as the problems related to poor
outcomes that patient’s experience. The economic toll arising from delirium on the
health care system will also be discussed. Potential causes and the risk factors for
delirium will be identified and the development of a predictive model of delirium
will be discussed. The literature associated with the development of delirium
management guidelines and their implementation in Australia will be appraised and
the evidence based management and prevention strategies recommended in these
guidelines will be examined. Finally, gaps in the academic literature on delirium will
be highlighted and the case for conduct of the current study will be will be
presented.
12
2.2 Delirium definition and diagnosis
The word delirium or delirious derives from the Latin word ‘delirare’ which means
‘going off the ploughed track’ (Ayto 2005). Hippocrates first described delirium as
‘phrenitis’,meaning an acute inflammation of mind and body (Lipowski 1990). The
term delirium is used interchangeably with terms such as acute brain failure, acute
organic brain syndrome, acute confusional state and post operative psychosis
(Saxena & Lawley 2009), although more recently, ‘acute confusion’ has become a
familiar descriptor for many clinical nurses (Milisen et al. 2002; Schofield 2008).
Variation in the use of diagnostic terms and descriptions is problematic in that it
contributes to the health professional’s failure to identify patients with delirium
(Sendelbach & Guthrie 2009). In order to maintain consistency in diagnosis and
research, Lipowski (1987) suggested that ‘delirium’ is the appropriate term to
describe the syndrome. As such, for the remainder of this thesis the term delirium
will be used.
After working closely with patients who developed delirium and examining their
electroencephalograph (EEG) results, Engel and Romano (1959) described delirium
as ‘a derangement in functional metabolism…and that this is reflected at the clinical
level by the characteristic disturbance in cognitive function’ (p. 262). The
publication of the ‘Diagnostic and Statistical Manual of Mental Disorders’ (DSM) by
the American Psychiatric Association (APA) in 1987 brought together definitions and
diagnostic features of delirium. More recently in the DSM–IV, delirium has been
defined as ‘a disturbance of consciousness that is accompanied by a change that
cannot be better accounted for by a pre existing or evolving dementia’ (APA 2000,
p. 136). This definition is widely recognised and used by researchers (Sendelbach &
Guthrie 2009). In 2013, the APA published the fifth edition of the DSM and updated
the definition of delirium. Delirium is now featured in a newly included chapter
titled ‘Neurocognitive Disorders’. In the previous edition it was included in the
chapter ‘Delirium, Dementia, and Amnestic and other Cognitive Disorders’. This new
edition of the DSM presents variations on the diagnostic criteria of delirium. Table 1
13
lists the diagnostic criteria presented in the new edition of the DSM V compared to
the previous edition, the DSM IV.
The new edition places a stronger emphasis on the disturbance of attention and
also states that delirium refers to altered cognition that cannot be better explained
by other pre existing neurocognitive disorders and has included this as an additional
diagnostic criterion. Similarly, the World Health Organisation (WHO) developed the
International Classification of Disease (ICD 10) in 1992 (World Health Organisation
1992). This is the tenth revision of the ICD and is the result of substantial alterations
made to the ICD 9, which was published in 1979 (Caraceni & Grassi 2004). The ICD
10 classification provides a list of diagnostic guidelines for delirium not induced by
alcohol or other psychoactive substances. For a definitive diagnosis the patient
must exhibit the following signs:
a) Impairment of consciousness and attention
b) Global disturbance of cognition: illusions, hallucinations, and impairment of
immediate recall and of recent memory
c) Psychomotor disturbances: hypo/hyperactivity and shifts between the two
d) Disturbance of the sleep wake cycle: including insomnia, total sleep loss,
reversal of sleep
e) Emotional disturbances: depression, anxiety, fear, wondering perplexity
The ICD 10 describes delirium as having a rapid onset with a total duration of less
than six months. The criteria produced by both of these organisations are relatively
similar but have some variations. The ICD 10 suggests additional criteria for
diagnosis: ‘disturbance in the sleep wake cycle’, ‘psychomotor disturbances’, and
‘emotional disturbances’ (World Health Organisation 1992; Mattoo, Grover & Gupta
2010). Although multiple sources of diagnostic criteria exist, a study by Kazmierski
et al. (2010) revealed the DSM criteria were more inclusive and the ICD 10 criteria
were more restrictive in establishing a diagnosis of delirium. This may be why the
DSM criteria are more commonly cited in the literature and used in research.
14
Table 1. Diagnostic criteria for delirium in the DSM 5 and the DSM 4
DSM IV
(American Psychiatric Association 2000, p.
143)
DSM V
(American Psychiatric Association 2013, p.
595)
A Disturbance of consciousness (i.e.
reduced clarity of awareness of the
environment) with reduced ability to
focus, sustain or shift attention
A disturbance in attention (i.e.
reduced ability to direct, focus, sustain
and shift attention) and awareness
(reduced orientation to the
environment).
B A change in cognition (such as
memory deficit, disorientation,
language disturbance) or the
development of a perceptual
disturbance that is not better
accounted for by a pre existing,
established, evolving dementia.
The disturbance develops over a short
period of time, represents a change
from baseline attention and
awareness and tends to fluctuate in
severity during the course of the day.
C The disturbance develops over a short
period of time and tends to fluctuate
during the course of the day.
An additional disturbance in cognition
(e.g. memory deficit, disorientation,
visuospatial ability or perception)
D There is evidence from the history,
physical examination and laboratory
findings that the disturbance is a
direct physiological consequence of a
general medical condition.
The disturbances in A and C are not
better explained by another pre
existing established or evolving
neurocognitive disorder.
E None There is evidence from the history,
physical examination and laboratory
findings that the disturbance is a
direct physiological consequence of
another medical condition.
15
As delirium can manifest in a variety of ways and present differently in each patient,
there are often challenges with recognition and diagnosis (Wong et al. 2010).
Physicians individually can interpret the diagnostic criteria for delirium and
differences in opinion may arise due to the eligibility to fit a certain criteria. This is
further highlighted in the large range of diagnostic tools that have been developed
to help health professionals diagnose delirium. The lack of uniformity in delirium
diagnosis continues. More than 24 delirium diagnostic instruments have been used
in published studies, including the NEECHAM confusion scale (Neelon et al. 1996),
the delirium symptom interview (Albert et al. 1992) and the delirium observation
screening scale (DOS) (Schuurmans, Shortridge Baggett & Duursma 2003). One tool,
the Confusion Assessment Method (CAM) developed by Inouye et al. (1990), has
been used extensively (Inouye, Westendorp & Saczynski 2014). This instrument is
based on the diagnostic criteria found in the DSM III R and was developed to allow
non psychiatric physicians to quickly and accurately diagnose delirium (Wei et al.
2008). The diagnostic algorithm of the CAM is based on four features: acute onset
and fluctuating course, inattention, disorganised thinking, and altered level of
consciousness (Inouye et al. 1990). In their original study, Inouye et al. (1990)
concluded that the CAM was a sensitive, reliable and easy to use tool to recognise
and diagnose delirium.
Following a systematic review on the use of the CAM, Wei et al. (2008) concluded
that it has helped improve identification of delirium in clinical and research settings.
However, formal training to use the CAM accurately is highly recommended.
Research investigating an educational intervention for nursing staff on 187 elderly
patients (aged 65 years and over) found a significantly lower sensitivity of the CAM
when used by those not trained in its application (Rockwood et al. 1994). Prior to
the educational intervention, delirium was recognised in 3% of patients. The
number of patients diagnosed with delirium following the intervention increased to
9%. This suggests that time spent educating nurses to undertake the CAM is
necessary. In addition, use of the CAM must be validated in settings other than
those in which it was originally developed. Currently the use of the CAM has not
been validated in the Australian setting (Tropea et al. 2008). Delirium diagnostic
16
criteria and assessment tools are widely accessible, although it is unclear how
Australian health professionals are diagnosing delirium and which tools are
currently being used to assist them make this diagnosis. There is currently no
research that has shown how health professionals in clinical practice in Australia are
diagnosing delirium.
This section has defined delirium and the diagnostic criteria according to the DSM
and the ICD. It has also highlighted the number of tools developed to assist in
making a delirium diagnosis. The next section will discuss the reported numbers of
patients that develop delirium. Research identifying a gap between the number of
patients diagnosed with delirium and the number of patients that potentially could
be diagnosed, if appropriately assessed, will also be discussed.
2.3 Incidence and prevalence
Firstly, prevalence of delirium refers to the number of patients admitted to hospital
with a delirium while incidence of delirium refers to the number of patients that
develop delirium during hospitalisation. The incidence of delirium is important as it
indicates the scale of a possible preventable condition. A number of studies have
shown that the incidence of delirium can vary across health care settings (such as
Intensive Care Unit (ICU), surgical or medical unit) and patient groups (Tropea et al.
2008). There is no clear reason why this occurs. However, at any one time there will
be at least one patient with delirium in a general medical, surgical or orthopaedic
ward (Schofield 2008). Most estimates of delirium incidence have originated from
studies of overseas populations and currently in the Australian setting there are
limited data regarding the prevalence and incidence of delirium (Travers et al. 2013;
Tropea et al. 2008). In Germany, Galanakis et al. (2001) found that delirium
developed in up to 55.9% of patients who had undergone hip fracture surgery.
Furthermore, in the United States of America (USA), McNicoll et al. (2005) identified
around 70% of all patients over 65 years in the ICU with delirium, whilst Inouye et
al. (1993) found that approximately 25% of medical patients developed delirium
17
during hospitalisation. This varying incidence highlights how delirium occurrence
changes across different settings.
Focusing only on medical in patients, a systematic review of studies of the
development of delirium found an overall incidence rate ranging between 3 – 29%
(Siddiqi, House & Holmes 2006). Studies included in the review originated from a
range of countries including Canada, United States of America (USA), United
Kingdom (UK), France, Italy, and Finland. Only one Australian study was included in
this review, highlighting the limited number of Australian studies on delirium
incidence in the medical setting and suggesting the need for additional Australian
research on delirium. Since the systematic review there have been a small number
of Australian prevalence and incidence studies conducted. Travers et al. (2013)
undertook an observational study of delirium prevalence in acute hospitals in
Queensland. The authors profiled 493 patients from surgical, medical and
orthopaedic units and concluded that delirium was a common problem in Australian
hospitals and was likely to increase with an ageing population. The authors
identified an overall delirium prevalence of 9.7% and a 7.6% delirium incidence
across all three settings (Travers et al. 2013). Data specific to the general medical
population in this study indicated an incidence rate of 3.6% (n = 9) in 70 79 year
olds, 3.4% (n = 7) for the 80 89 year olds, and 5.3% (n = 2) for patients over 90
years. Furthermore, Iseli et al. (2007) conducted a cohort study during 2005 – 2006
at an acute hospital in Melbourne, Victoria and found that approximately 18% (n =
19 out of 104) of general medical patients aged over 65 years had delirium on
admission to hospital and a further 2% (n = 2 out of 85) developed incident delirium
after admission. However, the researchers argued that this might be an
underestimation of delirium incidence due to the small number of participants
screened, and therefore not a true representation of the incidence of delirium.
Further evidence indicating an underestimation in delirium prevalence rates was
found in a study conducted by Speed et al. (2007) who undertook an audit of
patients who displayed signs of delirium in medical and surgical wards in Western
Australia. They found that, of the 132 patients (10.9%) who displayed symptoms
18
suggestive of delirium, only 48 (36%) had a confirmed diagnosis. This indicates that
a further 64% could potentially have had an undiagnosed and untreated delirium.
Typically, in everyday practice across all settings, as many as two thirds of delirium
cases are diagnosed late or are missed (O'Hanlon et al. 2014). These findings,
specifically the under reporting and non diagnosing of delirium are supported by
several other studies (Foreman et al. 1995; Inouye 2006; Potter & George 2006;
Siddiqi, House & Holmes 2006; Treloar 1998; Tropea et al. 2008) and suggest a need
for screening and appropriate diagnosis of delirium in clinical settings.
The ageing population in many western countries has raised concerns about the
increase in the number of older patients who have the potential to develop delirium
(Holden, Jayathissa & Young 2008; Ski & O'Connell 2006). The Department of Health
in Victoria, Australia (Department of Health 2012) estimated that the number of
individuals aged between 70 – 84 years will rise by 59% and individuals over the age
of 85 will increase by 79% by the year 2021. As hospital use tends to increase with
age, inevitably there will be an increase in the number of individuals with the
potential to develop delirium. Thus there is an urgent need for research examining
the prevalence and incidence of delirium in the acute care setting in Australia.
Epidemiological data of delirium prevalence and incidence in the Australian health
care context are necessary to provide an overall view of the impact delirium is
having on hospitalised patients, and form the first step towards dealing with the
cost of delirium in terms of patient outcomes and hospitalisation costs. These data
would lead to a better understanding of the extent of delirium and its far reaching
implications. The next section will discuss in more detail the problems that can
occur when a patient develops delirium, including the development of
complications and increasing their length of stay.
2.4 Complications of delirium
The misdiagnosis and under reporting of delirium may be contributing to significant
harmful effects on hospitalised patients. Moreover, research has shown that
19
delirium could be prevented in around 30 – 40% of cases (Inouye, Westendorp &
Saczynski 2014) and has the potential to be reversed with early recognition and the
right treatment (Inouye & Ferrucci 2006). Consequently, patients may be
experiencing poor outcomes unnecessarily, which further signifies the importance
of accurate diagnosis. Patients experience a range of complications as a result of
delirium. Some of these complications include: functional decline (McCusker et al.
2003), an increased likelihood of complications associated with longer hospital stays
(McCusker et al. 2003), an increased risk of admission to a long term care facility
post discharge (Voyer et al. 2006; Witlox et al. 2010), an increased risk of falls
(Lakatos et al. 2009), and higher mortality and morbidity rates (Twedell 2005;
Witlox et al. 2010). Furthermore, patients who develop delirium can temporarily
lose their ability to reason, their usual self care skills and personhood (Young &
Inouye 2007). These outcomes are devastating for any patient who may have been
functionally independent and living in their own home prior to developing the
delirium. Whilst many physical problems occur, patients, their families or carers can
also experience long term psychological complications (Breitbart, Gibson &
Tremblay 2002). After six months, approximately 50% of patients can recall the
delirious episode and in many cases are still distressed by their recollections
(O'Hanlon et al. 2014).
Following an episode of delirium, the patient is at increased risk of developing long
term cognitive impairment or dementia (Jackson et al. 2004; MacLullich et al. 2009;
Witlox et al. 2010). Some studies have shown that about 40% of patients who had a
delirium during hospitalisation developed cognitive impairment following
hospitalisation (Jackson et al. 2004; MacLullich et al. 2009). However, since most of
the research into patient outcomes has been conducted in overseas settings, there
remains uncertainty about the outcomes for patients who have experienced a
delirium in the Australian context. This information is important to be able to target
appropriate services for patients. As a result, patients may not be adequately
followed up or receiving the care they need to reduce the likelihood of developing
long term problems. Further research is therefore needed to address this issue and
20
determine what care should be provided post hospitalisation to patients that
develop incident delirium.
The economic cost of a patient developing a delirium is immense. In the USA, the
national financial burden of delirium on the health care system ranges from $38
billion to $152 billion per year, matching the health care costs of falls and diabetes
(Leslie et al. 2008). This also includes an average cost of approx. $60,000 per
delirious patient post the delirious episode, resulting from increased care needs and
possible admission to a long term care facility (Leslie et al. 2008; Tune & DeWitt
2011). However, there are no accurate economic cost data for the Australian
population (Ski & O'Connell 2006; Tropea et al. 2008). This absence of information
highlights a lack of insight into how much the Australian health care system is
spending on delirium. The next section will outline the possible causes and risk
factors that increase the risk of developing delirium.
2.5 Causes and risk factors
Delirium is a complex syndrome and may be caused by a number of physiological
factors (Inouye 2006; Inouye, Westendorp & Saczynski 2014). As health
professionals can implement delirium prevention interventions it is important to be
aware of the factors associated with delirium (Voyer et al. 2007). Numerous
researchers have attempted to explain what is occurring physiologically to cause
delirium (Choi et al. 2012; Flacker & Lipsitz 1999; Inouye & Ferrucci 2006;
MacLullich et al. 2008; Maldonado 2008; Sanders 2011; Simone & Tan 2011; Tune
2000; van Munster 2009; White 2002). As a result, various hypotheses and multiple
interacting theories have been proposed. Some of these theories include: reduced
blood flow to the brain due to ageing (Flacker & Lipsitz 1999), a dysfunction in the
metabolism of the brain or an overall insufficiency of the cerebral cortex (Engel &
Romano 1959; Romano & Engel 1944), cholinergic deficiency (Flacker & Lipsitz
1999), and an overreaction of the body’s natural stress response, with a
corresponding increase in systematic inflammation (MacLullich et al. 2008). Despite
21
these theories, the specific neurological processes that occur in the brain remain
unclear (Rigney 2010). This may be because a number of clinical conditions such as
sepsis or severe illness can cause a delirium. However, the clinical conditions may
not necessarily occur in the same combinations for each case of delirium (Kamholz
2010). That is, for two patients with similar clinical characteristics, one may develop
delirium, while the other may not. Although delirium may be caused by just one
factor, in older people the cause of delirium can be multifactorial (Inouye,
Westendorp & Saczynski 2014). That is, certain predisposing and precipitating risk
factors such as co morbid illness may increase a patient’s vulnerability to delirium.
Development of delirium is a complex process, which often involves a complex
multi factorial relationship between predisposing factors and exposure to
precipitating factors (Inouye 2006; Inouye, Westendorp & Saczynski 2014; Schofield
& Hasemann 2011). These risk factors for delirium appear to interact in intricate
ways, thus pinpointing any particular risk factor is difficult (Kalisvaart et al. 2006;
Villalpando Berumen et al. 2003).
One of the challenges of discovering the pathophysiological mechanisms of the
syndrome is that populations in which delirium has been studied are varied and
reflect significant heterogeneity (Rigney 2010; Voyer et al. 2007). Therefore, the
ability of researchers to draw definitive conclusions based on this evidence is
limited. Although the syndrome presents consistently across a range of settings in
terms of signs, the mechanisms behind its development may be completely
different (Rigney 2010). Inouye (1998a, 2006) stated that delirium could be the
common endpoint for a number of different conditions. As such, there is a need to
focus delirium research in a specific hospital setting and not to compare across
multiple settings because the causes may be different.
Delirium research focused in specific patient populations (such as surgical, ICU or
medical) will help isolate and examine setting specific factors related to delirium. A
number of studies have investigated delirium in the ICU setting (Divatia 2006;
McNicoll et al. 2003; Ouimet et al. 2007), and the surgical setting (Adunsky et al.
2003; Balas 2005; de Castro et al. 2014; Marcantonio et al. 1994; Pandharipande et
22
al. 2008), but few studies specific to the general medical setting, especially in
Australia, have been undertaken despite the observed but still anecdotal evidence
of a high incidence of delirium within this population. The present study will
therefore focus on delirium that develops in patients in the acute medical setting to
help address this gap. Understanding the factors that contribute to delirium in the
medical setting can help provide insight into the potential mechanisms that underlie
the syndrome for these patients. It is important to know more about these setting
specific risk factors so that patients who are admitted to a particular setting can be
screened for the risk factors identified as high risk in that population. For example:
patients admitted to ICU can be screened for risk factors specific to the ICU setting
such as medication use or severity of illness and patients admitted to a medical
setting can be screened for risk factors specific to the medical setting. This is
important as risk prediction models can aid in delirium prevention.
2.5.1 Predisposing and precipitating risk factors
2.5.1.1 Predisposing risk factors
Predisposing risk factors for delirium are present on a patient’s admission to
hospital. Dementia, or a pre existing cognitive impairment, is the most common
predisposing risk factors for delirium in older people (Inouye 2006). Dementia is
associated with increasing age and as the population ages, the incidence of
dementia increases. The association between cognitive impairment, dementia and
delirium is well documented in the literature (Ajilore & Kumar 2004; Arnold 2005;
Ciampi et al. 2011; Fick, Agostini & Inouye 2002; Foreman et al. 2001). Despite this,
the nature of this association remains poorly understood (Inouye 2006). The clinical
features of delirium and dementia remain closely intertwined and it is difficult for
clinicians to distinguish the two conditions (Treloar 1998). Inouye (2006) states that
around two thirds of patients who develop a delirium have a prior diagnosis of
dementia. Thus, such patients need to be closely monitored for delirium. It is for
this reason that screening for cognitive impairment on admission to hospital for all
adults over the age of 65 is recommended in the Clinical Practice Guidelines for the
Management of Delirium in Older People (Clinical Epidemiology and Health Service
23
Evaluation Unit and Delirium Clinical Guidelines Expert Working Group 2006).
Performing cognitive assessment screening using tools such as the Mini Mental
State Exam (MMSE) for a patient on admission can provide clinicians with a baseline
measurement of a patient’s cognition and assist in the prediction of patients’ level
of risk for delirium. Furthermore, Jones et al. (2010) found evidence to suggest that
people with no prior cognitive impairment possessed a reserve factor that acts to
delay the onset of neurodegenerative conditions, such as dementia, which impair
intellect and level of functioning. This information adds further evidence that
impairment in cognition increases patients’ vulnerability to delirium.
A study conducted by Schor et al. (1992) on delirium risk factors in general medical
and surgical wards found the most important factors that increased patients’
vulnerability for delirium were already present on admission. Existing co morbid
illnesses included cognitive impairment, advanced age greater than 80 years and an
admission diagnosis of a fracture (Schor et al. 1992). A range of predisposing factors
have been investigated in the literature including, stroke (Caeiro et al. 2004; Sheng
et al. 2006), frailty (Quinlan et al. 2011), depression, functional decline, sensory
impairment, and dehydration (Inoyue 2006). Most of these risk factors have been
examined across a number of hospital settings.
A systematic review conducted by Elie et al. (1998) focused on risk factors
associated with the development of delirium in hospitalised older patients. The
authors identified sixty one different risk factors. The most significant factors were
advanced age (greater than 80 years), dementia and medical illness. These factors
were the most studied and also had the strongest association with delirium (Elie et
al. 1998). Physical status, such as fever and hypotension had a low level association
with delirium, while pain was not associated with delirium (Elie et al. 1998). The
systematic review conducted by Elie et al. (1998) does have some limitations that
may reduce the generalisability of the results. In particular, the researchers did not
differentiate between the various hospital settings. The review examined studies
sampling patients from medical, surgical and psychiatric settings and did not
separate the results based on these hospital settings. Consequently, setting specific
24
risk factors have not been differentiated. The researchers also failed to consider the
distinction between prevalent and incident delirium. Separation of the two is
required to determine differences in risk factors that may have developed as a
direct result of hospitalisation. The authors made a number of recommendations
based on their findings including the recommendation that further research into
delirium risk factors should investigate only patients from one particular setting.
This will enhance the ability to recognise which factors that increase the risk of
delirium in different populations (Elie et al. 1998; Villalpando Berumen et al. 2003).
More recently, Mattar, Chan and Childs (2012) conducted a systematic review that
investigated the evidence concerning predisposing risk factors for the critically ill
patient in the ICU. Twenty four studies that examined factors causing delirium in
critically ill patients were included in the review. Medications administered in the
ICU, such as benzodiazepines, had the greatest association with delirium in this
review (Mattar, Chan & Childs 2012). Due to heterogeneity of the included studies,
meta analysis could not be undertaken. A narrative summary of the studies showed
that advanced age and an elevated C reactive protein (CRP) were also common risk
factors for delirium (Mattar, Chan & Childs 2012). It is clear from the results of these
two systematic reviews there are differences in risk factors for delirium in different
populations, further highlighting the need to study delirium in specific hospital
settings.
The review of studies conducted in the ICU setting has produced evidence for
delirium predisposing risk factors that are common in patients in the ICU (Mattar,
Chan & Childs, 2012). Yet, a systematic review examining risk factors for delirium in
medical in patients has not yet been conducted. Additionally, evidence relating to
risk factors for delirium has predominantly been generated from studies conducted
outside Australia. Consequently, there is limited evidence regarding risk factors for
medical in patients in the Australian population. To bridge this gap, a systematic
review examining patients in the general medical setting is required. The findings of
the review could then be used to develop risk prediction models that are specific to
25
patients admitted to medical settings and could be tested in the Australian
population.
2.5.1.2 Precipitating risk factors
Precipitating factors occur during the hospitalisation of a patient and can be more
easily controlled or treated by the health professionals caring for patients. These
factors include: abnormal blood results, such as abnormal blood urea
nitrogen/creatinine ratio (Elie et al. 1998; Inouye et al. 1993); taking benzodiazepine
medications (Santos et al. 2005); adding more than three new medications during
admission; indwelling catheter (IDC) use and restraints (Inouye and Charpentier
1996). A study conducted by McCusker et al. (2001a) also identified some
potentially modifiable risk factors that produced a greater severity of delirium.
These factors included multiple room changes, use of chemical/physical restraints
and the lack of a clock, a watch and reading glasses. Table 2 presents a summary of
potential predisposing and precipitating risk factors for delirium.
2.5.1.3 Relationship between predisposing and precipitating risk factors
The interplay between predisposing and precipitating factors is complex. Inouye
and Charpentier (1996) proposed a multi factorial model of delirium based on the
relationship between these risk factors. An example of how the model works is as
follows: a patient admitted with relatively low vulnerability or minimal predisposing
factors (low baseline risk) is at a higher risk of developing delirium if exposed to
multiple precipitating factors such as use of an indwelling catheter and adding more
than three medications (high precipitating risk). On the other hand, a patient
admitted with high vulnerability or many predisposing factors (high baseline risk)
may still develop delirium with minimal precipitating factors (low precipitating risk)
(Inouye & Charpentier 1996). In other words, the greater the number of
predisposing factors, the greater the sensitivity to increases in noxious insults
(Schreier 2010). This model highlights the importance of assessing patients’ risk
level on admission to hospital and in turn implementing prevention strategies that
can reduce the amount of precipitating factors that increase the likelihood of the
26
patient developing delirium. The next section will further explore how the
investigation of risk factors can help to develop risk prediction models to identify
patients at greatest risk of developing delirium.
Table 2. Precipitating and predisposing risk factors for delirium
(Table adapted from Inouye 2006)
Precipitating Factors Predisposing Factors
Alcohol/drug withdrawal Age > 65years
Admission to ICU Alcohol abuse
Dehydration Cognitive impairment
Environmental Decreased oral intake
Fever Dehydration
Hypoxia Dementia
Intracranial bleeding Depression
Inter current illnesses Functional status
Infection Fracture or trauma
Iatrogenic complications History of delirium
Low serum albumin History of falls
Meningitis Immobility
Narcotics Low level of activity
Primary neurologic disease Malnutrition
Prolonged sleep deprivation Male
Pain Severe illness
Sedative hypnotics Sensory impairment (visual and hearing)
Shock Terminal illness
Stroke Treatment with multiple psychoactive drugs
Surgery (orthopaedic, cardiac and
prolonged cardiopulmonary bypass)
Severe acute illness
Treatment with multiple drugs
Use of physical restraints
Use of bladder catheter
27
2.5.2 Predictive model for delirium
As previously discussed, risk factors for the development of delirium have been
identified in a number of studies (Inouye 1998a&b, 1999; Inouye et al. 1993;
Kalisvaart et al. 2006; Khurana, Sharma & Avasthi 2002; Korevaar, van Munster & de
Rooij 2005; McCusker et al. 2001a; Mentes et al. 1999; Schor et al. 1992).
Understanding predisposing and precipitating factors, as well as the multifactorial
relationship of delirium, has led to the development of predictive models. These risk
prediction models are based on the risk factors that have been independently
associated with delirium, and can help to identify a patient’s potential for
developing delirium during admission to hospital (Inouye & Charpentier 1996). One
such model was first developed by Inouye et al. (1993) and is based on predisposing
factors for delirium. Inouye et al. (1993) identified that screening patients using
previously identified risk factors can be effective in informing strategies to prevent
delirium. Factors identified in the study population were vision impairment, severe
illness, pre existing cognitive impairment, and dehydration (blood urea nitrogen
(BUN) level >18) (Inouye et al. 1993). These factors were then analysed to create a
prediction model. The presence of more than three of these factors results in an
assessment of high risk, one to two factors signifies intermediate risk, and no
factors indicates low risk. Performance of the predictive model was tested in a
validation cohort and rates of delirium were 3% in the low risk group, 16% in the
intermediate and 32% in the high risk group, indicating the model successfully
predicted the likelihood of a patient developing delirium and highlighted the
importance of implementing a risk stratification system (Inouye et al. 1993).
During the course of hospitalisation precipitating factors that can result in a high
risk of delirium have also been identified. Inouye and Charpentier (1996) developed
a delirium risk prediction model for precipitating factors, the most significant of
which were: use of restraints, malnutrition, introduction of more than three new
medications, use of an indwelling catheter, and any iatrogenic event (Inouye &
Charpentier 1996). Again, the presence of more than three of these factors results
in an assessment of high risk, one to two factors are associated with intermediate
risk, and no factors indicates low risk. Application of the predictive model in the
28
validation cohort found overall delirium rates of 4% in the low risk, 20% in the
intermediate and 35% in the high risk groups. Inouye and Charpentier (1996)
reported that this corresponded to an 8.2% rate of delirium per day in the high risk
group. These results further highlight the importance of assessing older medical
patients’ risk of developing delirium on admission to hospital and assessing the
possibility that delirium risk may increase if they are exposed to further
precipitating factors.
The use of prediction models can be valuable in identifying patients that may
require additional monitoring and implementation of basic intervention strategies.
If used together, the predisposing and precipitating factor models can provide a
solid basis with which to target at risk patients and develop preventive
interventions. Many researchers have adopted the use of predisposing risk factor
predication models and identified their importance in the clinical setting to
detection of patients at greatest risk of developing delirium (Kalisvaart et al. 2006;
Marcantonio et al. 1994; Moerman et al. 2012; O'Keeffe & Lavan 1996; Pompei et
al. 1994; Rudolph et al. 2011). Despite this supporting evidence, albeit from outside
Australia, the use of delirium risk prediction models has not been extensively
studied in the Australian setting (Tropea et al. 2008). The next section will discuss
the development and use of management guidelines to help guide clinical practice
in relation to delirium care.
2.6 Delirium management guidelines
Delirium is one of the most serious complications a patient can experience during
hospitalisation, yet it remains under recognised, inappropriately evaluated and
poorly managed in a large number of patients who develop the syndrome
(Marcantonio 2007). The most effective strategy for delirium management is
appropriate screening and prevention (Inouye 2006; Inouye et al. 1999). Nurses in
acute settings are in a unique position to detect delirium because they work closely
with older patients to be able to evaluate their cognition and are responsible for
29
identifying acute changes in cognitive status (Cheah et al. 2011). Nurses are
therefore at the forefront of clinical decision making regarding implementation of
prevention strategies. Research studies have shown that proactive intervention
strategies implemented by nursing staff to promote delirium screening and
prevention has reduced the incidence of delirium (Inouye et al. 2000; Lundstrom et
al. 2005; Marcantonio 2007), although this may not be routinely done in clinical
practice.
Guidelines have been developed to assist clinical staff in hospitals to provide
effective and proactive management to patients with delirium. Michaud and
colleagues (2007) systematically identified and evaluated all available guidelines,
systematic reviews, randomised control trials and cohort studies in order to provide
a set of recommendations for delirium management. The Delirium: Guidelines for
General Hospitals outline a set of recommendations based on research for
screening, prevention and management of delirium in hospitalised patients
(Michaud et al. 2007). The guidelines provide advice on risk factors, and
recommendations regarding prevention of delirium, screening and diagnosis, and
pharmacological and non pharmacological treatment of delirium (Michaud et al.
2007).
There are also a number of international guidelines addressing the management of
delirium: the Guidelines for the Prevention, Diagnosis and Management of Delirium
in Older People in Hospital (British Geriatric Society 2006) and the ‘Practice
Guideline for the Treatment of Patients with Delirium’ (American Psychiatric
Association 1999). Implementation of guidelines such as these is an important step
to improve the management of delirium. However, research has provided evidence
of the inconsistency in not only the recommended guidelines but in the actual
clinical practice across disciplines (O'Hanlon et al. 2014). However, it is unclear if
nurses have been made aware of these guidelines and how they are using the
guidelines to influence their clinical practice. The next section will discuss guidelines
that have been developed in Australia and evidence regarding their use in the
Australian setting.
30
2.6.1 Delirium guidelines in Australia
In Australia, the National Standards on Safety and Quality Health (Australian
Commission on Safety and Quality Health Care [ACSQHC]) were developed to
provide a set of measures that are nationally consistent across the range of health
care settings (ACSQHC 2011). Standards are developed to provide strict guidelines
to ensure safe high quality care. There are ten standards that address various
aspects of hospitalisation including: medication safety, patient identification, clinical
handover, blood and blood products, preventing pressure injuries, recognising
clinical deterioration and preventing falls. The first standard states clinical practice
services should be ‘adopting processes to support the early identification and
appropriate management of patients at increased risk of harm’ (ACSQHC 2011 p.
18). Patients who develop delirium are at greater risk of harm including those
discussed earlier such as falls, increased length of stay and increased risk of
functional decline. Yet, a standard relating to identification or management of
people with cognitive impairment and delirium, specifically, does not yet exist. The
ninth standard, which applies to recognising and responding to clinical deterioration
in acute health care, is deemed to ‘not apply to deterioration of a patient’s mental
status’ (ACSQHC 2011, p. 61).
The standards currently offer no guidance on what to do when there is a sudden
change in a patient’s mental status, which is usually due to delirium, despite this
being a serious threat to their health. Many researchers have emphasised that a
change in brain function is an early marker of a serious underlying condition
(Flaherty et al. 2007) and mental status assessment has been suggested as a sixth
vital sign (Flaherty et al. 2009). Researchers have suggested that delirium be used as
a marker for the quality of patient care in hospitals (Inouye, Schlesinger & Lydon
1999). Despite delirium not being included in the National Standards on Safety and
Quality in Health Care, Australian guidelines for delirium management do exist to
guide practice; for example, the Clinical Practice Guidelines for the Management of
Delirium in Older People (Clinical Epidemiology and Health Service Evaluation Unit
and Delirium Clinical Guidelines Expert Working Group 2006). Yet, compared to the
National Standards on Safety and Quality Health, which are rigirously implemeted in
31
health care settings through accreditation processes, the implementation of the
Clinical Practice Guidelines for the Management of Delirium in Older People is
neither monitored nor compulsory for health services. Recently though, and since
conducting this research, because there are currently no mechanisms for requiring
best practice in the Australian setting, ACSQHC have released information regarding
the development of a Delirium Clinical Care Standard (ACSQHC 2015) and resources
for providing safe and high quality care for patients with cognitve impariment, titled
‘A better way to care’ (ACSQHC 2014). These resources have been developed to
guide clinicians to improve the care of people with delirium.
The Clinical Practice Guidelines for the Management of Delirium in Older People
(Clinical Epidemiology and Health Service Evaluation Unit and Delirium Clinical
Guidelines Expert Working Group 2006) were developed after a thorough search of
the literature and are the result of the synthesis of a number of previous studies
that investigated delirium screening, prevention and management. The
recommendations are based on research and expert opinion. The guidelines
therefore represent the best available evidence for delirium management at the
time of development. However, the authors emphasise that most of their
recommendations for practice have emanated from research conducted in overseas
settings and they advise caution in generalising them to the Australian setting
(Clinical Epidemiology and Health Service Evaluation Unit and Delirium Clinical
Guidelines Expert Working Group 2006). Furthermore, as research remains ongoing
into the best strategies for delirium management, the authors state that these
guidelines are not a definitive statement on screening for delirium and the
management of the syndrome and only provide a general guide (Clinical
Epidemiology and Health Service Evaluation Unit and Delirium Clinical Guidelines
Expert Working Group 2006).
In order to build upon the Clinical Practice Guidelines for the Management of
Delirium in Older People, the Australian Health Ministers Advisory Council (AHMAC)
developed the Delirium Care Pathways (AHMAC 2011). The Delirium Care Pathways
were developed to provide clinicians with examples of different patient journeys
32
and how these patients could be managed in these scenarios. The two documents
should be used in conjunction to provide the most effective care for patients with
delirium.
The Clinical Practice Guidelines for the Management of Delirium in Older People and
the Delirium Care Pathways reflect the best available evidence for the screening,
prevention and management of delirium. However, since development of these
guiding documents little research has been undertaken to evaluate their uptake in
health organisations in Australia, especially their use by nursing staff and the
implementation of these guidelines has not been widely reported (Mudge et al.
2012). It is important to investigate how organisations have implemented the
guidelines because the guidelines themselves will not improve the process of care
or change the outcomes for patients who develop delirium (Young & George 2003).
The implementation of the Clinical Practice Guidelines for the Management of
Delirium in Older People needs to be supported through the policies of
organisations. One study aimed to review the presence of policies for the
management of behavioural signs and presence of delirium in hospitals in
Melbourne, Australia (Watson et al. 2009). The authors surveyed all the public and
private hospitals in Melbourne to determine the extent of use of a policy or
procedure to inform management of delirium in hospital settings. The study further
investigated the availability of a policy regarding restraint use, models of care for
delirium and education for staff regarding delirium detection and management
(Watson et al. 2009). The authors found that 83% of the hospitals surveyed had a
management policy for aggression or severe agitation (which is often caused by
delirium) and a specific delirium management protocol was available at less than a
third of the hospitals surveyed (Watson et al. 2009). Although this study addressed
a range of gaps in the evidence regarding delirium management, uncertainties
regarding the content of the policies used by hospitals remain. The researchers did
not investigate screening policies for delirium and cognitive impairment. Therefore,
a paucity of evidence about many aspects of delirium management outlined in the
Clinical Practice Guidelines for the Management of Delirium in Older People and the
33
Delirium Care Pathways exists. Relatively little knowledge exists regarding the tools
or processes used to detect and diagnose delirium in Victorian hospitals. Research is
needed to establish whether structured processes are used to screen and diagnose
delirium and whether processes are used for the screening of cognitive impairment
generally. This information is important for developing an understanding of the
current policies and practices of delirium management and to provide an insight
into where hospital processes and systems may need to improve.
Only one study was identified that aimed to implement the Clinical Practice
Guidelines for the Management of Delirium in Older People in a general medical
service in Queensland in order to improve the screening, prevention and
management of delirium (Mudge et al. 2012). The researchers implemented many
of the recommendations contained in the Clinical Practice Guidelines for the
Management of Delirium in Older People including screening processes and
improving education for delirium. The authors stated there was difficulty in
effectively implementing some of the environmental strategies. The researchers
reported no incidence cases of delirium, which is in contrast with many studies
investigating delirium incidence. The screening for delirium was only conducted
twice a week and as a result incident cases of delirium may have been missed.
Implementation of the Clinical Practice Guidelines for the Management of Delirium
in Older People recommendations did help to significantly reduce persistent
delirium, which may be a result of better recognition by health professionals in the
care team. The authors also noted reductions in falls rates during the
implementation period. This study has shown some early improvements to
practices that resulted from the implementation of the clinical guidelines. More
research is needed to investigate the extent to which these guidelines have been
effective.
Information regarding current medication management guidelines for patients who
develop delirium is also scarce. Medication management and knowledge of the
most appropriate medications to use for patients who are agitated with delirium is
important. Serious complications and medication mismanagement can have
34
detrimental effects on older patients (Clegg & Young 2011). The availability of a
medication management policy for aggressive and agitated patients with delirium is
integral to encouraging the use of appropriate medications. A limited quantity of
research has investigated the prescribing patterns of medications for patients with a
delirium. An Australian study conducted by Tropea et al. (2009) investigated the
pharmacological management of delirium as a baseline measure prior to the
publication of the Clinical Practice Guidelines for the Management of Delirium in
Older People (Clinical Epidemiology and Health Service Evaluation Unit and Delirium
Clinical Guidelines Expert Working Group 2006). The authors audited medical
records to determine which medications, and dosages, were being administered to
patients with delirium. Since the release of the Clinical Practice Guidelines for the
Management of Delirium in Older People however, no further research has been
undertaken to investigate the medication management hospital policies and the
patterns of medication prescribing for management of delirium. Developing an
understanding of the policies and procedures available to clinicians in hospitals and
examining medical records to determine how medication is being used in relation to
delirium is important. A difference between current and recommended clinical
practice may be found, so one aim of this research is therefore to address this gap
and to investigate the existence and use of delirium management protocols in
hospital settings.
The Clinical Practice Guidelines for the Management of Delirium in Older People are
based on research into the prevention and management of delirium. The following
sections will discuss the delirium prevention and management strategies outlined in
the Clinical Practice Guidelines for the Management of Delirium in Older People.
2.6.2 Delirium prevention and management strategies
Patients who develop delirium are at greater risk of experiencing poor outcomes;
there is, therefore, a need to determine and implement the most effective and
appropriate prevention and management strategies. Early identification of risk
factors for delirium as well as appropriate management strategies can help to
35
reduce the potential for a patient to develop delirium (Burns, Gallagley & Byrne
2004). A number of studies have been undertaken to determine the appropriate
methods of delirium management (Day, Higgins & Keatinge 2011; Kamholz 2010;
Young et al. 2008). According to Flaherty (2006), the most important areas for
delirium management include the prevention of delirium, avoidance of the
outcomes of delirium, and reduction in the need to physically restrain the patient.
2.6.2.1 Prevention of delirium
This section will discuss in detail research that supports use of the prevention
strategies recommended in the Clinical Practice Guidelines for the Management of
Delirium in Older People (Clinical Epidemiology and Health Service Evaluation Unit
and Delirium Clinical Guidelines Expert Working Group 2006). Delirium is a serious
condition that needs to be well understood by health professionals looking after
patients with the potential to develop the syndrome (Burns, Gallagley & Byrne
2004). Delirium is a syndrome that can be caused by a number of factors which
necessitates broad intervention strategies (Inouye et al. 2006). Therefore, multi
component interventions for delirium prevention are important in delirium
management.
The multifactorial nature of delirium makes the targeting of delirium prevention
interventions complex and requires several strategies. Inouye et al. (1999) identified
multi component interventions for delirium, aimed to reduce exposure to
additional risk factors and prevent delirium from developing during hospitalisation.
Using the risk predictive model discussed previously (developed by Inouye et al.
1993), participants were assigned a level of risk, based on four risk factors for
delirium: visual impairment, cognitive impairment, high blood urea/creatinine and
severe illness (Inouye et al. 1999). The intervention strategy called the ‘elder life
program’ involved a team of nurse specialists and also trained volunteers (Inouye et
al. 1999). Additional risk factors for the development of delirium were targeted,
with specific interventions implemented for each risk factor. Risk areas that were
targeted were: cognitive impairment, sleep deprivation, immobility, visual
impairment, hearing impairment and dehydration (Inouye et al. 1999). These risk
36
factors were addressed using the following interventions: an orientation board,
early mobilisation strategy, encouraging the use of glasses and hearing aids and
encouraging oral intake. The intervention resulted in a reduced incidence of
delirium and a reduction in the number of days’ delirium persisted. However, this
benefit was only seen in patients who were identified as at intermediate risk at the
beginning of the study (Inouye et al. 1999).
Further studies investigating the elder life program have been conducted since its
development in 1999. In 2000, Inouye adapted the original program to become the
‘hospital elder life program’ (HELP), not only aimed to prevent delirium but also to
reduce the factors that lead to cognitive decline (Inouye et al. 2000). The hospital
elder life program provided comprehensive care for the patients in order to
maintain them at a regular level of functioning (Inouye et al. 2000). Several years
later, following the success of the original design, Inouye et al. (2006) investigated
the use of HELP, which had been implemented in 13 sites across the USA and
Canada. The HELP was reported to have had many beneficial outcomes across the
sites (Inouye et al. 2006). This included provision of appropriate training, improving
patient outcomes, enhancing nursing education, both patient and family
satisfaction and appeared to be cost effective (Inouye et al. 2006).
A systematic review conducted by Milisen et al. (2005) included seven studies using
multi component interventions and aimed to determine beneficial and efficient
characteristics of the interventions (Bogardus et al. 2003; Cole et al. 2002; Cole et
al. 1994; Inouye 1999; Marcantonio et al. 2001; Milisen et al. 2001; Wanich et al.
1992). The elements of multicomponent interventions that appeared to be most
effective in delirium prevention were those that would be considered basic
elements of daily care (Milisen et al. 2005); for example, encouraging the use of
hearing aids and glasses if applicable, and also maintaining adequate hydration
(Milisen et al. 2005). This finding suggests the need for education for health
professionals on the importance of maintaining a basic level of care. The review
concluded that as nurses are in frequent contact with patients, they play a major
37
role in implementing these basic interventions, including early recognition and
treatment of delirium (Milisen et al. 2005).
In a study conducted following the systematic review on multi component
interventions, Vidán et al. (2009) carried out further investigations on the
implementation of a multicomponent intervention in daily practice in a hospital in
Madrid, Spain. Vidán et al. (2009) stated that the HELP intervention (Inouye et al.
2000; Inouye et al. 1999) was difficult to implement in another country. There were
costs associated with the implementation of the intervention such as copyright on
protocols and the costs for other staff members, for which is often difficult to
obtain funding in public health care (Vidán et al. 2009). Vidán et al. (2009)
implemented an intervention similar to the principles of the HELP intervention but
it did not require the assistance of extra staff. The study excluded people with
delirium on admission and included people with the presence of at least one risk
factor: cognitive impairment, acute disease severity, visual impairment or
dehydration (blood urea/creatinine >40) (Vidán et al. 2009). An assessment was
conducted within the first 24 hours of admission to hospital. A specialist geriatric
nurse, trained in delirium took responsibility for implementation of the
intervention. The intervention consisted of simple or multi component actions that
targeted specific risk factors for delirium and were repeated daily (Vidán et al.
2009). Patients were also assessed daily, using validated instruments, for the
presence of delirium. The intervention successfully prevented the development of
delirium in patients with no prior delirium. However, in patients that had developed
delirium, the intervention had no effect on the severity or duration of the episode
(Vidán et al. 2009). Therefore, prevention interventions appear to be more effective
than interventions implemented only once delirium has developed.
Education for health professionals is an important component of delirium
prevention and management. Education of medical and nursing staff can help to
increase the awareness of risk factors for delirium and also symptoms experienced
by patients (Rockwood et al. 1994; Tabet et al. 2005). An intervention strategy
investigated by Tabet et al. (2005) aimed educational sessions at medical and
38
nursing staff in order to reduce the incidence of delirium and increase recognition
of the syndrome. Two acute wards were included in the study. Staff on one ward
received the intervention and the other did not receive the intervention and had
only the usual care. The intervention did not impact on the day to day management
of delirium but increased awareness in staff on the ward. Staff on the intervention
ward underwent teaching sessions, were provided with written information and
participated in group discussions (Tabet et al. 2005). The point prevalence of
delirium was significantly reduced in the intervention group compared to the
control group (p <0.05), indicating an overall positive effect on the prevention of
delirium (Tabet et al. 2005).
The Clinical Practice Guidelines for the Management of Delirium in Older People
(Clinical Epidemiology and Health Service Evaluation Unit and Delirium Clinical
Guidelines Expert Working Group 2006) summarise these prevention strategies to
make the evidence based strategies clear and accessible to health professionals.
Table 3 outlines the prevention strategies identified in Clinical Practice Guidelines
for the Management of Delirium in Older People. Little research has been
undertaken by researchers in Australia on prevention interventions and it is not
known if these strategies are appropriate in this setting. There is currently limited
data to show if and how any of these prevention strategies have been implemented
into health care organisations in Australia and if they have been effective. The next
section will discuss the management strategies that are outlined in the guidelines
and the evidence to support these strategies.
2.6.2.2 Management of delirium
This section will discuss the importance of effectively diagnosing, documenting and
developing a management plan for patients that develop delirium. Once delirium
has been identified prompt attention is required (Burns, Gallagley & Byrne 2004).
One of the most important aspects of delirium management is to actively seek and
treat the cause of the delirium (Burns, Gallagley & Byrne 2004). These may be as
simple as eliminating all the precipitating factors, for example treatment of
39
infections or other inflammatory conditions, removal of harmful medications and
correction of metabolic disturbances (Kamholz 2010). It has been identified by
previous research that removal of the underlying cause will usually result in
resolution of symptoms of delirium (Casey et al. 1996). Other strategies, such as
preventing complications and supporting the patient’s functional needs, are also
important aspects of delirium management. Medication use may be one of the
underlying causes of delirium development but because of the behavioural
symptoms of delirium, medications are often used. However, medications should
only be used when other strategies have been ineffective in controlling behavioural
symptoms (Miller 2008).
Table 3. Delirium prevention strategies
Delirium Prevention Management Strategies
Ensure appropriate lighting
Provide a clock and a calendar
Encourage and assist the patient with eating and drinking (to reduce risk of dehydration
and under nutrition)
Ensure that patients who usually wear hearing and visual aids are assisted to use them
Optimise communication (for example, use interpreters and liaison staff).
Avoid psychoactive drugs
Encourage family or carer and friends to visit and be involved in patient care.
Promote relaxation and sufficient sleep and discourage daytime napping.
Avoid use of mechanical restraints
Avoid use of indwelling catheters
Avoid room changes (may increase disorientation)
Ensure that pain relief is adequate
Promote cognitive stimulation
(Clinical Epidemiology and Health Service Evaluation Unit and Delirium Clinical Guidelines Expert
Working Group 2006)
Medications to manage delirium need careful assessment for the risks and benefits
to the individual patient before being prescribed and administered. There is
insufficient evidence to support the use of clear pharmacological approaches in the
40
treatment of delirium (Kamholz 2010). However, use of some medications to
stabilise behaviour can be effective. There are a number of systematic reviews on
the use of different types of medications in the treatment of delirium (Clegg &
Young 2011; Lacasse, Perreault & Williamson 2006; Seitz, Gill & Zyl 2007). The use
of benzodiazepines is contraindicated in delirium as it may worsen symptoms
(Kamholz 2010; Lonergan et al. 2009). Haloperidol is one of the most studied drug
treatments for delirium and appears to be effective (Kamholz 2010). Although, a
recent systematic review conducted by Wang et al. (2012) found that despite
consistent recommendations for the use of haloperidol in the treatment of
delirium, minimal data exists to support its efficacy in critically ill patients.
Furthermore, there is evidence to suggest that haloperidol is no more effective than
other antipsychotics such as olanzapine and rispiridone in the management of
delirium (Lonergan et al. 2007).
A study conducted by Boettger et al. (2011) compared the use of aripiprazole and
haloperidol in the treatment of delirium symptoms. Results indicated that these
drugs were both equally effective in relieving symptoms (Boettger et al. 2011).
However, around 19% of patients treated with haloperidol had some
extrapyramidal side effects. These side effects occurred in patients with hyperactive
delirium because the behaviour associated with hyperactive delirium meant these
patients required higher doses of haloperidol. In contrast, patients with hypoactive
delirium required less haloperidol and did not experience any side effects (Boettger
et al. 2011). The authors concluded that aripiprazole might be just as effective as
haloperidol in the treatment of delirium (Boettger et al. 2011). Recommendations
for delirium medication management in older people support the theory to ‘start
low, go slow’, and low doses of haloperidol, rispiridone and olanzapine have been
shown to be appropriate (Burns, Gallagley & Byrne 2004; Lonergan et al. 2007).
The Clinical Practice Guidelines for the Management of Delirium in Older People
(Clinical Epidemiology and Health Service Evaluation Unit and Delirium Clinical
Guidelines Expert Working Group 2006) summarise the medication management
strategies. However, minimal investigation by Australian researchers has been
41
undertaken to explore medication management and treatment strategies, in the
Australian setting. This is striking as this indicates existence of little data regarding
how patients with delirium are treated in terms of addressing the cause of delirium,
the medications that they are receiving and the strategies being incorporated into
their care. This information would provide greater understanding of the treatment
and management strategies currently used in hospitals in Australia and could give
an indication as to where education may need to be implemented.
2.7 Conclusion
This literature review has provided an outline and discussion regarding the
definition of delirium as well as delirium diagnostic criteria. The evidence outlined in
the review has highlighted the importance of accurately recognising and identifying
patients with a delirium. The tools developed to assist clinicians make this diagnosis,
especially the use of the Confusion Assessment Method, was also examined.
Limited knowledge regarding how clinicians diagnose delirium and the applicability
of using diagnostic tools in the Australian setting was also identified as a gap in the
literature.
Delirium prevalence and incidence estimates from overseas settings have been
presented and the limited amount of Australian prevalence and incidence data
discussed. The evidence presented in the review has highlighted that accurate data
is required to uncover the true incidence of delirium as well as the full social and
economic cost of delirium on patients and the Australian health care system.
Evidence regarding the complications associated with delirium and the subsequent
poor outcomes for patients that develop delirium was outlined. Patients in the
Australian health care setting who develop delirium may be unnecessarily
experiencing these complications as a result of poor recognition and diagnosis.
Patients can experience both physical and psychological effects as a result of
delirium, including falls, pressure injuries and functional decline. The lack of
42
Australian data on the outcomes of patients who developed delirium was identified
in the review and requires further research.
Evidence relating to the causes and risk factors related to delirium has been
outlined in this literature review. A number of hypotheses about how delirium
develops in older patients were examined. Risk factors that have been associated
with the development of delirium, including both predisposing and precipitating risk
factors were discussed. The importance of synthesising evidence of risk factors in
any one particular setting was also identified. However, knowledge of risk factors in
individual hospital settings is an important aspect of delirium care that needs
further research. The development and use of risk prediction models was also
reviewed, but limited data of the use of risk prediction models in Australia has
highlighted a gap in evidence related to their applicability in this setting.
The development of the Clinical Practice Guidelines for the Management of Delirium
in Older People, and more recently the Delirium Care Pathways, was discussed. The
importance of these guidelines due to the lack of a delirium care standard was also
identified. The lack of data regarding the use and implementation of clinical practice
guidelines in Australian hospitals has been identified. Only a few studies have
investigated the implementation of policy and procedures based on information
available in the guidelines. There is a need to investigate the current policies used
by hospitals as well as the clinical practice that is occurring in hospitals. This
information will provide valuable insights into how delirium is currently managed
and where education and training may need to be targeted. There is a need to
provide greater cohesion in the everyday clinical management of delirium.
To summarise, it is clear that there are significant gaps in research in Australia
regarding: the use of diagnostic tools for delirium diagnosis, identification of
predisposing and precipitating risk factors, use of risk predication models,
measurement of outcomes for patients, use of prevention and management
strategies, and medications administered during an episode of delirium. These are
gaps that require further research in the Australian context.
43
Chapter 3 Methods
3.1 Introduction
This chapter presents the methods used for this research. The purpose of the study
and aims are described. The methods used in each of the three phases of the
research will be outlined in detail. For the first phase of the research, a systematic
review methodology was used to identify the risk factors for incident delirium in
medical in patients. A case control retrospective audit was undertaken in the
second phase of the research to examine the characteristics of patients who
developed incident delirium as well as describe current care practices. Finally, in the
third phase of the research, a survey was conducted to gather data about the
policies and procedures used in hospitals in Melbourne, Australia, to identify and
manage patients who develop delirium. A description of each of the study settings,
the study populations and research questions addressed, as well as the individual
tools used for data collection in each phase are also presented.
3.2 Research purpose
The overall purpose of this research was to contribute to the limited evidence base
about the risk factors, clinical characteristics and management of incident delirium
in hospitalised general medical patients in Australia.
The following are the overall aims of the research.
1. Systematically review the evidence for risk factors related to the
development of incident delirium in general medical patients.
2. Describe the characteristics of medical in patients who develop incident
delirium during hospitalisation, including demographic characteristics,
potential risk factors (predisposing and precipitating), residency prior to
admission and outcomes for patients (including discharge destination,
length of stay in hospital and medication treatment).
3. Examine and describe the current state of delirium management in the
acute hospital setting.
44
3.3 Research phases
A multi phase design, involving three phases, was used to address the overall aims
of the research. Phase 1 involved a systematic review of the evidence to identify risk
factors for the development of incident delirium in medical in patients in the acute
care setting. The evidence identified from this review was used to inform the
second phase of the research.
Phase 2 was a case control retrospective audit of medical records of patients who
developed incident delirium during hospitalisation, and a matched control group of
patients that did not develop delirium. The audit focused on identifying the
characteristics of patients who developed delirium, including evidence of the risk
factors identified in the systematic review.
The third and final phase of the study involved surveying key informants from
hospitals in Melbourne, Australia to identify the delirium management policies and
procedures currently in use in health care organisations. Refer again to Figure 1 for
an outline of the phases of the study and how each phase address the study’s
overall aims.
Phase 1 Phase 2 Phase 3
Figure 1. Phases of the research project
Identified Risk
Factors
Survey of deliriummanagement practices
in hospitals
AIM: Identify currentpractice in relation todelirium management
AIM: Compare currentpractice with clinicalpractice guidelines
AIM: Identify clinical characteristics ofmedical patients who develop delirium.Including medical related risk factors fordelirium
Retrospective Case controlclinical records audit
Clinicalrecordsaudit
Systematic review ofrisk factors for
incident delirium inmedical patients
Clinicalrecordsaudit
Policyand
ProtocolSurvey
Systematic review
45
3.4 Phase 1 Systematic review
3.4.1 Research design
A systematic review was conducted as part of this research because there were no
existing systematic reviews that specifically examined risk factors for incident
delirium in general medical patients. The aim of a systematic review is to identify,
appraise and synthesise empirical evidence in order to answer a particular research
question (Higgins & Green 2011). Systematic reviews assist researchers to consider
not only the results of one or more studies but to produce more reliable findings of
multiple studies that can be used to inform decision making (Joanna Briggs Institute
2011). Meta analysis can also be used when studies are sufficiently similar. Meta
analysis is a statistical process that allows the combination of results to produce an
overall statistic (Higgins & Green 2011). Systematic reviews can be particularly
useful in examining the relationship between risk factors. The review therefore
helped to examine the relationship between incident delirium and its risk factors in
the medical in patient population. Figure 2 represents the process of a systematic
review to filter, synthesise and produce quality findings specific to a particular topic.
Figure 2. Process of a systematic review
3.4.2 Systematic review process
The process of a systematic review is structured and involves many replicable
procedures undertaken in a series of steps. The Joanna Briggs Institute (JBI) review
methodology was used in this review (Joanna Briggs Institute 2011). Prior to
undertaking the review, the author underwent training through the Joanna Briggs
Pooled data
Systematic review
Individual studies
Systematicreviewprocess
46
Institute on how to undertake a JBI systematic review. Consistent with JBI
systematic review methodology, a systematic review protocol was developed and
submitted to JBI for approval (Cull et al. 2012) (Appendix 1). The steps undertaken
to conduct the systematic review will now be outlined.
3.4.2.1 Framing the research question
The first step in conducting the systematic review involved framing the research
question. Researchers can use tools to help develop a descriptive and answerable
research question. PICO is a mnemonic used to develop a research question by
dissecting the problem into component parts and restructuring it so it is easy to find
the answers (Joanna Briggs Institute 2011). The PICO mnemonic was utilised to
assist in the development of the research question for this review. Below is each of
the elements of the PICO tool used to assist in development of the research
question.
P – Population. The population that is to be examined needs to be closely
considered and based on information gained from previous literature. As age has
been found to be a risk factor for delirium, this systematic review assessed studies
that examined adults (defined as being 18 years or above) admitted to an acute
general medical setting. The age of patients to be included was not restricted to
older adults in order to be inclusive of all potential studies that may have
investigated the incidence of delirium in younger adults.
I – Interest (phenomena of interest). The interest (or the phenomena of interest) is
the subject that is being explored. For this research, the phenomenon of interest
was risk factors that may contribute to development of incident delirium.
C Comparison (or Control). The comparison component of PICO examines the
comparison between groups on the dependant variable or intervention outcomes.
For this research, studies that compared patients who developed delirium during
hospitalisation to patients who did not develop delirium were examined.
47
O – Outcome. The outcome component relates to the particular outcome being
assessed. The incidence of delirium as related to individual risk factors was the
outcome to be examined in this systematic review.
Therefore, after considering all the components of the PICO mnemonic, the
research question was formulated as:
What risk factors are associated with incident delirium in adult patients during an
acute medical hospitalisation?
More specifically, the review objective was: to identify the best available evidence
regarding the factors associated with delirium in adult patients admitted to acute
medical facilities.
3.4.2.2 Developing the aim of the systematic review
Following development of the research question, the overall aims and objectives of
the review were developed. The aim of the systematic review was:
To explore and describe risk factors related to the development of delirium in acute
general medical patients. That is, to identify which factors (both predisposing
and/or precipitating) contributed to incident delirium in hospitalised adults in an
acute general medical setting.
3.4.2.3 Developing the protocol
Before commencing the systematic review, a protocol consistent with JBI
methodology was developed and submitted to JBI for review and approval. The
protocol for this review was approved by JBI prior to commencement of the
systematic review and was published in the JBI Library of Systematic Review
Protocols (Cull et al. 2012) (Appendix 1).
3.4.2.4 Inclusion and exclusion criteria
Inclusion and exclusion criteria can vary among reviews and are dependent on the
review question. Most importantly, inclusion and exclusion criteria are used to
48
determine the studies that may be eligible for inclusion in the review. Included
articles should clearly state the types of designs, the population under investigation,
the phenomena of interest and the outcomes to be considered for the review.
Types of participants
This review included studies that investigated adults (defined as 18 years and
above) who were admitted to an acute medical setting (e.g. general medical units,
stroke units, short stay units and neuro medical units) who were not delirious on
admission as assessed using a valid assessment method on admission (in order to
differentiate incident delirium) but who developed incident delirium during
hospitalisation.
The review excluded patients who were:
Critically ill and admitted to a specialist unit e.g., ICU or CCU
Admitted for any type of surgery (including patients who had a surgical
intervention during hospitalisation)
Admitted for alcohol related reasons
Admitted to a psychiatric facility
These patients were excluded in order to determine factors that may be exclusive
to the medical in patient setting.
Phenomena of interest
This review included studies that evaluated any risk factors (including predisposing
and precipitating factors) that may have contributed to the development of incident
delirium.
Types of outcomes
This review included studies that investigated incidence of delirium as related to
individual risk factors.
Types of studies
This review included both experimental and epidemiological study designs,
including: non randomised controlled trials, quasi experimental, before and after
49
studies, prospective and retrospective cohort studies, case control studies and
analytical cross sectional studies. Only studies published in English were considered
for inclusion. Studies published from January 1996 until July 2012 (date of literature
search) were considered for inclusion in this review. The review included studies
published from January 1996 in order to synthesise relevant quantitative studies
published after Elie et al.’s (1996) systematic review on delirium risk factors in
medical, surgical and psychiatric settings in order to have the most recent evidence
and to enable comparison in findings.
3.4.2.5 Search strategy
Consistent with JBI methodology, a three step search strategy was utilised in this
review. The search strategy aimed to identify both published and unpublished
studies. The three step process is outlined below:
1. An initial search of MEDLINE and CINAHL was undertaken to identify and
understand the main text words contained in the titles and abstracts, and of
the index terms used to describe articles.
2. A second search using all identified keywords and index terms was then
undertaken across relevant databases.
3. The reference lists of all identified reports and articles were then searched
for additional studies.
Below is a list of the key words used and the databases searched.
Key words used to search databases were:
risk factor OR risk factors
predisposing factors
precipitating factors
dementia or cognitive impairment
urinary tract infection
pneumonia
sepsis
50
delirium
acute confusion
acute confusional state
confusion
medical
hospital in patient
medical in patient
medical admission
hospitalisation or hospitalization
The databases searched using these terms were:
Medline
CINAHL
PsycInfo
Cochrane Library
Joanna Briggs Institute
Informit Health collection
Proquest Health and Medical
Embase
Scopus
The search for unpublished studies included:
Proquest Dissertation and Thesis
Mednar
JBI Library of Systematic Reviews and the Cochrane Library were searched for
similar systematic reviews that could be potential sources of primary studies.
University librarians assisted in the development of the search strategy for the
systematic review. The final search strategy is detailed in Appendix 2.
51
3.4.2.6 Identifying potential studies
The list of potential studies identified in the search strategy was then assessed
based on the title and abstract, against the inclusion and exclusion criteria. A large
number of studies were excluded in this process. Studies that potentially fit the
inclusion criteria were then retrieved as full text and the study was reviewed in
further detail. Studies that did not meet the inclusion were then excluded. The
remaining studies were included in the review.
3.4.2.7 Assessing the methodological quality of studies
Studies that met all inclusion criteria were then assessed for methodological quality.
This is an important step in the review process because in order to produce the
most reliable findings, studies with poor quality should not be included (Joanna
Briggs Institute 2011). For this systematic review the standardised critical appraisal
instruments from the Joanna Briggs Institute Meta Analysis of Statistics Assessment
and Review Instrument (JBI MAStARI) (Appendix 3) was used to determine study
quality. Specifically, the tool helps to assess study quality by determining the risk of
bias in the study design, the way the study was conducted, and the analysis of the
results. Prior to inclusion in the review two independent reviewers assessed
retrieved studies for methodological quality. Findings regarding quality of the
studies included in the review are presented in the results chapter.
3.4.2.8 Extracting the data
Data extraction refers to the process used by the researcher to source and record
relevant results from the original research study (Joanna Briggs Institute 2011). Data
extracted from the original research studies included participant information, the
phenomenon being examined and the results of the studies. Data must be extracted
systematically using standardised data extraction instruments. Data for this review
were extracted from included studies using the standardised JBI MAStARI data
extraction tool (Appendix 4). The JBI data extraction tool for comparative cohort
and case control studies was used. Data were extracted from studies that reported
possible risk factors for patients that developed delirium, as well as patients with no
52
delirium. Even if the results for possible risk factors were not significant, data were
extracted so that it could be compared to similar studies. When data were reported
in percentages only, it was necessary to calculate exact numbers using the reported
number of participants in each group in order for the data to be entered into the
meta analysis software. When reported outcome data were insufficient, authors
were contacted to obtain the original data set.
3.4.2.9 Data synthesis
Synthesis of data can either be descriptive (narrative synthesis) or statistical (meta
analysis). Meta analysis is a statistical technique for combining the findings from
independent studies (Higgins & Green 2011). The overall goal of meta analysis is to
combine the results of previous studies in order to arrive at a summary conclusion
about a particular body of research. It can increase the precision of the estimate
and provides a greater chance of detecting a statistically significant real effect
(Joanna Briggs Institute 2011). However, in order for generalisation of results to be
valid, studies included in the meta analysis should be similar to each other. The
main areas that should be comparable include: clinical (similar patient
characteristics), methodological (outcomes measured the same way), and statistical
(outcomes measured using comparable scales) (Higgins & Green 2011).
For this research, the quantitative data extracted from the studies were (where
appropriate) pooled in statistical meta analysis using JBI MAStARI. All results were
subject to double data entry. Effect sizes were expressed as odds ratio and their
95% confidence intervals were calculated for analysis. Heterogeneity was assessed
statistically using the standard chi square and also explored using subgroup
analyses based on the different study designs included in the review. Where
statistical pooling was not possible the findings are presented in narrative form
including tables and figures to aid in data presentation.
3.4.3 Ethical considerations
Ethics approval was not required in order to conduct Phase 1 of the research.
53
3.5 Phase 2 Case control study: retrospective audit
3.5.1 Research design
A case control study design was used for Phase 2 of the research. A retrospective
analysis of medical records of patients aged 18 years and over admitted to an acute
medical setting and coded for delirium at the study site over a 2 year period (1st Jan
2012 – 31st December 2013) was undertaken. Case control studies are important for
helping to yield important scientific findings in a short period of time (Rothman,
Greenland & Lash 2008; Schulz & Grimes 2002), which was necessary due to the
time constraints of this doctoral research. In case control studies, study groups are
defined by outcome. Researchers look back in time to ascertain each person’s
exposure status (Rothman, Greenland & Lash 2008). For this research the outcome
examined was the patient’s development of delirium, and the possible risk factors
that led to this development. Furthermore, an audit of medical records specifically
examining recommendations in the Clinical Practice Guidelines for the Management
of Delirium in Older People (Clinical Epidemiology and Health Service Evaluation Unit
and Delirium Clinical Guidelines Expert Working Group 2006) provided an
understanding of current practices and processes used in the management of
patients with delirium. The clinical records audit examined the processes of care for
patients who develop delirium. Risk factors identified in the systematic review
helped inform decisions about which risk factors to examine in the case control
clinical records audit.
3.5.2 Aims of the case control study
The aims of the case control clinical records audit were to:
1. Describe the characteristics of medical patients who develop delirium during
hospitalisation, including demographic characteristics, potential risk factors
(predisposing and precipitating), residency prior to admission, outcomes for
patients including discharge destination, length of stay in hospital and
medication treatment.
54
2. Describe delirium management of medical patients in an acute hospital
setting.
Table 4 provides examples of the research questions and hypotheses addressed in
the case control retrospective audit. An extensive list of the research questions
examined, variables, hypotheses and statistical technique used is provided in
Appendix 5.
Table 4. Examples of research questions and hypotheses
Research question Hypotheses
What risk factors are most
commonly associated with
incident delirium in the medical
patient population?
A patient with a diagnosis of dementia will have a
greater likelihood of developing delirium than a
patient with no dementia.
A patient with a cognitive impairment will have a
greater likelihood of developing delirium than a
patient with no cognitive impairment.
What is the average age of
patients who develop delirium
compared to patients who do not
develop a delirium?
Patients who develop delirium will be older than
patients who do not develop delirium.
What are the clinical outcomes
for patients who experience a
delirium compared to patients
who do not?
Patients who develop delirium will experience more
adverse events compared to those who do not
develop delirium.
3.5.3 The research setting
The medical records of patients admitted to any medical setting at the health care
organisation (network) research site were audited. The study site was a public
health tertiary health care organisation (network) comprised of seven hospitals.
Three of the seven hospitals have an emergency department and as a result only
these three hospitals were included in this study. The sites are public, acute care
hospitals that service an urban region of Melbourne, Australia. Table 5 outlines the
55
approximate number of patient beds available at each site and the number of
medical patients that stayed overnight for the year 2011/12 (the closest year to that
of data collection).
Table 5. Approximate bed numbers and hospital admissions per hospital
(National Health Performace Authority 2013) *bed numbers vary with seasonal demand
3.5.4 Study population
The target case population were patients aged 18 years (in order to determine age
range of adult patients in the medical setting who develop delirium) and over who
had been admitted to a medical unit (for example: medical wards including
respiratory, gastrointestinal, renal, neurological, cardiac as well as short stay and
Rapid Access Medical Units (RAMU)) with no evidence of a delirium, and who
developed incident delirium during hospitalisation. A matched control group of
patients was also selected and these were patients that did not develop incident
delirium.
3.5.5 Sample and sampling approach
3.5.4.1 Patients with delirium
After discharge, patient medical records were reviewed by support services and
coded according to the diagnoses made during hospitalisation. The report of
patients who were coded for delirium during hospitalisation was obtained from the
health care organisation’s decision support services department. The report
contained the patient record numbers, the dates of discharge and the type of
admission (e.g. medical) enabling the relevant records to be retrieved via the online
patient record system called Clinical Patient Folder (CPF). CPF is an electronic
Hospital Number of hospital beds Medical overnight admissions 2011/12
Hospital 1 Approx. 200 500* 11,903
Hospital 2 Approx. 200 500* 16,546
Hospital 3 Approx. 100 200* 7,869
56
storage system for medical records; following a patient’s discharge, the records for
that admission are scanned electronically and filed in separate electronic folders for
the specific admission. Each admission to hospital is sorted into separate folders
and dated. Records scanned into CPF include all progress notes, emergency
progress notes, admission forms, assessment tools, medication charts and discharge
summaries. The access to each record is monitored and only available to be viewed
by staff members.
Records of the patients coded for delirium were then assessed to confirm whether
or not the patient met the inclusion criteria. All records meeting the following
inclusion were closely reviewed and audited:
Inclusion criteria
Patients aged 18 years or over on admission (to determine the appropriate
age range of patients whom develop delirium).
Patients admitted to a general medical unit between 1st January 2012 and
31st December 2013.
Patients diagnosed with delirium during admission and/or a discharge
episode code for delirium.
Exclusion criteria
Patients diagnosed with delirium in the emergency department or showed
signs of delirium (such as confusion)
Patients with delirium tremens or drug and alcohol intoxication.
Patients admitted to ICU or CCU (from the emergency department), a
psychiatric or sub acute facility, surgical patients, and patients who had
surgery during admission.
3.5.4.2 Patients with no delirium (control group)
Retrospective analysis of a random sample of medical records of patients not coded
for delirium and admitted to an acute medical setting over a 2 year period (1st
January 2012 – 31st December 2013) was also undertaken. Record numbers for
57
these patients were also obtained from the health care organisation. The decision
support services department at the health care organisation used a random
number generator to identify the patients in the control group. The random
function used is called NEWID. This function is used to select a random sample of
cases based on one or more variables. These variables included only patients
admitted to a medical setting, and not coded for delirium.
In order to determine criteria for matching the groups, records of control patients
were retrieved following completion of the audit of delirium patients. The age range
of patients with delirium was ultimately used as the criterion for the selection of the
non delirium control group. Patients in the delirium group were all over the age of
42, and therefore a random sample of patients greater than or equal to 42 years of
age without a code for delirium was retrieved.
After generation of the random sample, controls were assessed against the same
inclusion and exclusion criteria in order to have a matched control group of
patients. Once the required number had been retrieved, sampling was complete.
3.5.4.3 Sample size power calculations
On the advice of the University statistician, the approximate ratio of patient medical
records to be sampled was two control records per delirium record, in order to
achieve power of approximately 80% to detect a difference in the two groups. The
number of patients needed for the study was calculated using a power analysis with
the assistance of a statistician at Deakin University. Using some risk factors that
have been previously identified as significant (e.g., dementia) for the development
of delirium in different populations, the minimum numbers of participants required
to detect a significant difference between groups was calculated. The risk factors
used were; dementia, because it is a common risk factor identified in all settings,
and use of benzodiazepines in order to determine the impact of medicine use on
delirium incidence. In order to assess outcome differences between the two groups,
previously identified length of stay difference was also used to calculate sufficient
sample size.
58
Dementia
In the systematic review, dementia was found to be a significant risk factor for the
development of incident delirium. Prior data indicates the prevalence of dementia
among controls is 0.15 (Fick, Agostini & Inouye, 2002). If the true odds ratio for
disease in exposed subjects relative to unexposed subjects is two, sample size
estimates suggest studying 151 case patients and 302 control patients to be able to
reject the null hypothesis that this odds ratio equals one with probability (power)
0.8. The Type I error probability associated with the test of this null hypothesis is
0.05. An uncorrected chi squared statistic will be used to evaluate this null
hypothesis. Table 6 presents the sample size calculations to detect significance in
dementia.
Table 6. Sample size calculations to detect significance in dementia between casesand controls.
DementiaOdds Ratio Cases Controls2.0 151 302
2.2 113 226
2.5 81 162
power 0.8, p = 0.05 for null hypothesis, prevalence of dementia among controls0.15
Benzodiazepines
Data in a previously identified study (Marcantonio, 1994) indicate that the
prevalence of benzodiazepine use among controls was 0.08. If the true odds ratio
for delirium in exposed subjects relative to unexposed subjects is three, sample size
estimates suggest studying 84 case patients and 168 control patients to be able to
reject the null hypothesis that this odds ratio equals one with probability (power)
0.8. Table 7 presents the sample size calculations to detect significance in use of
benzodiazepines.
59
Table 7. Sample size calculations to detect significance in benzodiazepine usebetween cases and controls
BenzodiazepinesOdds Ratio Cases Controls3 84 168
2.5 129 258
power 0.8, p = 0.05 for null hypothesis, prevalence of benzodiazepines among controls0.08
Length of stay
In a previously identified study (McCusker, Cole, Dendukuri & Belzile, 2003) the
length of stay within each subject group was normally distributed with standard
deviation 14. If the true difference in the case and control means is 7, sample size
calculations suggest studying 14 case subjects and 28 control subjects to be able to
reject the null hypothesis that the population means of the case and control groups
are equal with probability (power) 0.8. The Type I error probability associated with
the test of this null hypothesis is 0.5. Table 8 presents the sample size calculations
to detect significance in length of stay.
Table 8. Sample size calculations for differences in length of stay
Length of stayTrue difference ofmeans
Cases Controls
7 14 285 27 543 75 1502 169 338power 0.8, p = 0.05 for null hypothesis, standard deviation within each subjectgroup 14
3.5.4.4 Sample size
The above calculations were used to determine the approximate number of
patients needed for the study. Patient records were retrieved until approximate
numbers were reached. The sample size obtained from one year (2012) of patients
that developed delirium was 79 cases. As determined by the sample size
calculations, this was not sufficient to detect a significant difference in the two
60
groups. In order to fulfil the required numbers to detect significant differences in
the case and control groups it was necessary to include an additional year of
patients (2013). Final numbers are reported in the results chapter.
3.5.6 Data collection
A structured audit tool (Incident delirium clinical records audit tool) was used to
extract data from medical records (Appendix 6). The tool was adapted, with
permission, from a tool originally developed by Tropea, Slee, Holmes, Gorelik and
Brand (2009) for delirium research that focused on the medications administered to
patients. The development of the incident delirium clinical records audit tool was
informed using the Clinical Practice Guidelines for the Management of Delirium in
Older People (Clinical Epidemiology and Health Service Evaluation Unit and Delirium
Clinical Guidelines Expert Working Group 2006). Questions in the tool related to
cognitive assessment, delirium detection and management of delirium were based
on recommendations outlined in the guidelines. For example: guideline 1.5.1 states
‘A formal cognitive function assessment, (which may include the use of a standard
cognitive screening tool) should be performed on all older people (aged 65 years or
older) as part of the routine admission process to all health care settings.’ On the
basis of this recommendation the question included in the audit tool was ‘Was a
cognitive test performed on admission? If so what test was used?’. If the patient
was younger than 65 years, ‘not applicable’ was recorded for this question.
To aid in data collection the audit tool was customised into a digital form, created
using the ‘Tap Forms’ (©2013 Tap Zapp Software Inc) application on an iPad. ‘Tap
Forms’ is a mobile personal information manager tool for tracking personal data
and ideas. ‘Tap Forms’ uses SQLite to store and manage data in a safe and secure
environment. The application allows the user to create and change tools used for
the user’s requirements. The structured audit tool was developed into a form on
the iPad for ease of use and to assist with data output (Appendix 7). Tap forms
collates data taken from each of the collected records and generates the
61
information into a .csv data file that can be exported into Microsoft Excel and
subsequently into IBM Statistical Package for the Social Sciences (SPSS).
In order to assess usability and appropriateness of the tool, it was pre tested on a
random sample of 10 patients discharged prior to January 2012. The tool was then
further refined and changed where appropriate in order to reflect the information
available in the medical records. This included adding questions relating to the
patient’s prior falls history, mobility and continence.
Data collected from each of the medical records included:
Demographic data: age, gender, diagnosis on admission, residency prior to
admission, presence of functional impairment (as assessed by the Katz Index
of Independence in Activity of Daily Living using information in the medical
history), length of stay in hospital, place of discharge.
Patient’s level of cognition as described by admitting physician/nurse (or as
reported by patient’s family members), and/or recent Mini Mental State
Exam Score, if available.
Risk factors (including predisposing and precipitating), as informed by the
results of the systematic review, all known risk factors for delirium and their
presence for each patient. This included precipitating factors that may have
occurred prior to delirium development.
For patients not diagnosed with delirium: any suspected signs of delirium
were noted.
Evidence of whether current practice was consistent with the Clinical
Practice Guidelines for the Management of Delirium in Older People. The
audit was used to examine the following clinical practice:
o The assessment of risk factors for delirium. Was a tool used to assess
delirium risk?
o Detection of delirium: How was delirium diagnosed, was a valid tool
used? Detection of cognitive impairment: Was a baseline cognitive
function assessment such as the Abbreviated Mental Test (AMT) or
62
the Mini Mental State Exam (MMSE) performed on admission? Was
there documentation of a recent change in cognitive function?
o If delirium was diagnosed, was a care plan implemented for delirium
management?
o Prevention of delirium: Were preventative strategies put into place?
o Management of delirium: identification of cause, management of
symptoms, and prevention of complications. Was cause of delirium
identified? What strategies were used to manage delirium?
o Discharge planning and follow up: Was information regarding
delirium made available for patients and families? What follow up
monitoring was implemented? Was counselling considered for
people who had delirium?
The data were obtained using the entire patient medical record for the admission
episode under examination. The student researcher undertook data collection for
this phase of the research over an eight month period, from August 2013 to March
2014. Data collection ceased when the required number of participants had been
reached. Following data collection, data files were exported from the audit form on
the iPad application ‘Tap Forms’ (©2013 Tap Zapp Software Inc). The data were first
checked for errors, specifically looking for values that fell outside the range of
possible values for a variable. Errors or missing information were subsequently
corrected if detected. Evidence of missing data prompted a re review of the medical
record to obtain the missing information. Every effort was made to ensure missing
data were collected. As a result, there is minimal missing data for this study. For a
small number of patients certain documentation was missing from their file in
which case, for the respective variable, a code to indicate the data were not
available was entered into the field.
3.5.7 Data analysis
Data from the iPad were exported as a .csv file into Microsoft Excel © and then
imported into the statistical program IBM SPSS (Statistical Package for the Social
63
Sciences) Version 22.0. The data were extracted using the incident delirium clinical
records audit tool and were summarised and analysed using a number of statistical
methods. Using IBM SPSS the data were reviewed for missing data. Variables in
SPSS were examined and adjusted based on the program requirements for
statistical analysis. Variables such as the diagnosis of delirium were defined and
assigned labels. If a patient had been diagnosed with delirium during admission
they were labelled with a 1. If the patient did not have a delirium they were labelled
with a 0. This was used for all variables that had dichotomous Yes = 1 /No = 0
answers.
Data were summarised and analysed using a number of statistical methods. Firstly,
the data were explored utilising descriptive statistics to analyse age and gender
data as well as frequencies of other demographic characteristics of patients in both
the case and control groups. Bivariate analysis was then conducted to explore the
relationship between categorical variables. Chi square was used to explore the
relationship and the differences among categorical variables. The Yates’ Correction
for Continuity value was calculated, which compensates for the overestimation of
the chi square value, when used with a two by two table (Pallant 2013). The
relationships between potential predisposing and precipitating factors of incident
delirium were examined, as well as the outcomes patients experienced. Variables
were initially explored using cross tabulation to determine if the proportion of
patients with a variable (e.g., dementia) was significantly different between the
case and control groups. Results of the chi square tests included the odds ratio and
the corresponding p value with the significance level set at p 0.05. The 95%
confidence interval was also expressed. The 95% confidence interval is the range of
values that we can be 95% confident encompasses the true value of the odds ratio
(Pallant 2013). The odd ratio represents the change in odds of being in one of the
categories of outcome when the value of a predictor increases by one unit
(Tabachnick & Fidell 2013). If the confidence interval does not contain 1, then the
result will be significant. If the confidence interval does contain 1, the odds would
not be significant as we could not rule out that the true odds ratio was 1, indicating
equal probability of being a case or control.
64
When exploring differences in mean scores between groups the independent
samples t test was used. This was used to compare the mean ages of patients in the
case and control groups. Results of the t tests were interpreted according to
Levene’s Tests for equality as either equal variance assumed, or equal variances not
assumed. If the significance value for equal variances assumed is larger than 0.05,
then equal variance is assumed between the two groups. If the significance reading
is less than 0.05, then the variances for the two groups are not the same, and then
equal variance is not assumed (Pallant 2013). The corresponding results based on
equality were then used as the outcome. The significance value, mean difference
and 95% confidence interval of the difference for the independent samples t tests
are reported in the results.
Multivariate logistic regression modelling was used to analyse the relationship
between multiple independent variables and incident delirium. Logistic regression
provides an indication of the relative importance of each predictor variable or the
interaction among the predictor variables (Pallant 2013). Thus, in order to
determine the overall interaction between independent variables and incident
delirium, as well as to determine the relative importance of each predictor variable,
logistic modelling was used to test variables and identify factors independently
related to incident delirium. Selection of predictors for the model was based upon
the results of the chi square analysis and only variables that showed a significant
relationship with delirium were used. Logistic regression was therefore performed
on all variables shown in the bivariate analysis using chi square/t test to have a p
value <0.1 (Abbott 2014).
The multivariate analysis generated factors that remained significantly related to
incident delirium when adjusting for possible interactions between all of the
factors. The model was then adjusted by removing, one at a time, factors that were
not related to incident delirium with a p value > 0.1. The logistic regression was
then re performed with the factor removed. Again, the logistic regression results
showed which factors remained significantly related to incident delirium when
65
adjusting for possible interactions between the remaining variables. Non significant
factors were subsequently removed one at a time. This step of removing one
variable that was not significantly related to incident delirium at a time was
undertaken until all remaining factors reached a significant value p < 0.1, giving the
final logistic regression model.
The adjusted odds ratios, 95% confidence intervals and p values for the logistic
regression modelling are presented in the next chapter. Table 9 presents a sample
of the research questions, hypotheses, the variables assessed and the statistical
tests that were used for each question. A comprehensive table containing these
elements is presented in Appendix 5.
Table 9. Research questions and statistical tests used
Question Variables Hypothesis Statistical test
Is there arelationshipbetween age anddevelopment ofdelirium?
Age – continuous: age inyears (mean age ofgroups)Delirium diagnosis –categorical: Yes or No
Advanced age isrelated to thedevelopment ofdelirium.
Independentsamples t tests
Is there arelationshipbetween pasthistory and deliriumdiagnosis?
Past History categorical:e.g. Dementia,Hypertension: Yes/NoDelirium Diagnosis –categorical: Yes/No
There may be arelationshipbetween some pasthistory diagnosisand delirium butnot others.
Chi squaretest forindependence
What precipitatingfactors predict thelikelihood that apatient will developdelirium duringadmission?
Dependent Variable:Delirium – categoricalYes/NoIndependent variables:Categorical Use of IDC,
Use of Restraints, andGiven Benzodiazepineduring admission.Continuous variable:
Sodium on admission andSodium Day 3 ofadmission
The use of IDC,restraints and beinggivenbenzodiazepines issignificantly relatedto incidentdelirium.
Logisticregression
66
3.5.8 Ethical considerations
The ethical guidelines and principles as set out in the National Statement on Ethical
Conduct in Human Research (National Health and Medical Research Council
[NHRMC] 2007) and the Australian Code for the Responsible Conduct of Research
(the Code) (Australian Government 2007) were adhered to in this study. In
conducting this study, a number of ethical issues were considered and addressed.
Formal ethics approval was obtained from both the Human Research Ethics
Committee at Deakin University and the health organisation where medical records
were obtained. Ethical approval letters are available in the Appendix 8.
3.5.7.1 Consent
Consent to view medical records was not obtained from individual patients. A
waiver of consent from the ethics committee was sought and approved for the
audit of medical records. Due to the nature of retrospective data collection, it was
impractical to seek consent from all patients whose medical records were
examined.
3.5.7.2 Confidentiality
During data collection, the name of the patient whose medical record was being
examined was omitted. Patient identification numbers were assigned to each
patient during data collection. Only the allocated record number was available with
each patient’s extracted data. The patient identification numbers obtained from the
health organisation were stored digitally in a password protected file. In order to re
identify records from which data had been extracted, the allocated record numbers
were stored in a separate file with the corresponding patient identifier number. This
helped when it was necessary to go back to a particular record if information was
missing. Furthermore, the patients and also the health organisation have not been
identified through the reporting of the results.
3.5.7.3 Data storage
All data or materials collected and processed have been securely archived and
stored in accordance with Deakin University policies and procedures. As per Deakin
University's "code of good practice research” the data will be kept for a minimum of
67
six years after publication. Any information in digital form will be deleted after this
six years (this includes: audit forms, SPSS spread sheets, other spread sheets,
graphs, and descriptive data) and hard copy information will be shredded using a
secure disposal service. Data collected using the iPad was protected during data
collection by using a password to gain access to the iPad. This password was only
accessible by the student researcher. The iPad was also stored securely in a locked
cabinet following data collection. Data files containing information collected on the
iPad during the audit were stored in a password protected folder on the student
researcher’s university computer in a locked office at Deakin University.
68
3.6 Phase 3 Delirium management survey
3.6.1 Research design
In Phase 3 of the research, a survey was conducted to investigate the policies and
procedures currently utilised in hospitals in Melbourne in relation to delirium
screening and management. As previously discussed in the review of the literature,
the screening processes for delirium and cognitive impairment was not addressed in
the survey conducted by Watson et al. (2009). Consequently to address this gap,
screening processes were explored in the present study survey. The survey method
was chosen for this phase because it is an effective way to gather information about
a particular topic (Bruce, Pope & Stanistreet 2008) and data were required from a
number of health care organisations. The survey was used to gather data related to
how hospitals are addressing delirium management at a procedure and protocol
level.
3.6.2 Aim of the survey
To examine and describe delirium management in acute hospitals in Melbourne,
including the use of the Clinical Practice Guidelines for the Management of Delirium
in Older People.
3.6.3 Research setting and participants
Key representatives from all public and private hospitals in Melbourne, identified
from the Victorian Department of Human Services public website “Melbourne
Metropolitan Hospitals and Health Services Locations” (2013), were invited to
participate in the study. Hospitals primarily providing services for paediatrics,
obstetrics, gynaecology, psychiatry, and day procedure and palliative care units,
were excluded.
3.6.4 Study sample
The target population were Executive/Directors of Nursing at each of the public and
private hospitals in Melbourne. The details of the relevant personnel were
identified from hospital websites or by contacting the hospital to obtain the details
of the relevant personnel.
69
3.6.5 Data collection
3.6.5.1 Data collection instrument
A structured survey tool (Appendix 9) was developed following a review of the
literature and was further informed by the Clinical Practice Guidelines for the
Management of Delirium in Older People (Clinical Epidemiology and Health Service
Evaluation Unit and Delirium Clinical Guidelines Expert Working Group 2006). The
survey built on the tool conducted by Watson et al. (2009) to explore areas not
covered in Watson et al.’s survey. For example, the Watson et al. (2009) survey did
not include questions regarding the assessment of patients’ cognition on admission.
The current survey addresses this by asking if there is a policy or part of an existing
policy that states a cognitive assessment should be performed on admission.
The areas of inquiry included the following:
Do you currently have a policy regarding delirium management in your
organisation?
Was this policy developed with the help of the Clinical Practice Guidelines for
the Management of Delirium in Older People or the Delirium Care Pathway
developed by the Victorian and Australian Government?
Does the policy outline the need to conduct a baseline cognitive function
assessment on all patients over the age of 65? If so, what tool is suggested?
What tools, if any, are recommended for use to assess for delirium?
Are these tools readily available to staff members to use? And is training
provided to encourage their use?
What pharmacological management is recommended by the protocol?
Are risk factors for delirium assessed in all older patients admitted to the
acute setting?
3.6.5.2 Data collection procedure
The Executive/Director of Nursing at each hospital in Melbourne was emailed to
explain the study and to seek approval for inclusion of the hospital/health service.
The email (Appendix 10) contained a copy of the plain language statement and the
70
survey. In the email, an invitation was included for the Director of Nursing to
nominate key informants within the organisation to respond to the survey during a
phone interview with the researcher, or via the return of the survey by email.
Contact was then made by the researcher via email with the nominated key
informant/s to supply them with the plain language statement, a copy of the survey,
invite participation and, if they were agreeable, to negotiate a suitable time to
conduct a phone interview. Formal verbal consent was obtained from the
participant at the beginning of the phone interview and was recorded using an
audio recording device. Participants were also given the option of completing the
survey on the computer and submitting it via email or sending it via postal services.
Follow up emails and phone calls were undertaken to remind participants to
complete the survey.
3.6.6 Data analysis
Response data for the survey were entered into the statistical program IBM SPSS
Version 22.0. These responses were summarised using descriptive statistics,
including frequencies and percentages of hospitals that had a delirium management
protocol as well as other procedures for delirium management.
3.6.7 Ethical considerations
The ethical guidelines and principles as set out in the National Statement on Ethical
Conduct in Human Research (NHMRC) and the Australian Code for the Responsible
Conduct of Research were adhered to. In conducting this study, a number of ethical
issues were considered and addressed. Formal ethics approval was obtained from
both the Human Research Ethics Committee at Deakin University, and when
requested, at the health organisations taking part in the survey. Ethical approval
letters are available in the Appendix (Appendix 8).
3.6.7.1 Consent
Survey participants were provided with a participant information document in the
form of a plain language statement (Appendix 11) prior to consenting to undertake
the survey in order to provide them with information necessary to make a decision
regarding their involvement in the research. Participants gave verbal consent to
71
participate in the survey when the interview was being undertaken over the phone.
If the survey was returned by email, consent was implied.
3.6.7.2 Confidentiality
Participant’s names were not recorded when completing the survey, only their role
within the organisation. The individual responses of participants or health
organisation are not identifiable through the aggregated reporting of the results.
3.6.7.3 Data storage
All data or materials collected and processed have been securely archived and
stored in accordance with Deakin University policies and procedures. Data collected
for the survey has been securely stored in a password protected folder on a Deakin
University computer. Hard copies of the mailed surveys are stored in a locked filing
cabinet in a locked office at Deakin University. As per Deakin University's "code of
good practice research” the data will be kept for a minimum of six years after
publication. Any information in digital form will be deleted after this time (this
includes SPSS spread sheets and descriptive data) and hard copy information will be
shredded using a secure disposal service.
72
3.7 Summary of methods
This chapter has provided a description of the methods used to undertake this
research. In order to address the aims of the research a multi phase research design
was necessary. All three phases of the research have been outlined. A detailed
description of the systematic review process was presented with sufficient detail to
allow the review to be replicated in future research. The steps used to undertake
the case control study have been outlined. The calculation of sample sizes, how the
sample including cases and controls was identified, the development of the tool
used for data extraction and the statistical methods used for data analysis have
been presented. Finally, the development and distribution of the delirium
management survey has been described. Ethical considerations for each of the
three phases of the study have also been discussed. The following chapters will
report on the results of the three phases, provide an in depth discussion of the
findings of the research and the implications of the findings for practice, education
and research.
73
Chapter 4 Results
4.1 Introduction
This chapter presents the results of the three phases of the study. Firstly, the results
of the systematic review will be reported including the meta analysis and narrative
synthesis, followed by the results of the case control study, and finally the results of
the delirium management survey.
4.2 Phase 1 Systematic review results
The search strategy, the characteristics of included studies, a review of the
methodological quality of the included studies, and the results of the meta analysis
and narrative synthesis will be presented.
4.2.1 Search results
The initial search yielded 6,056 results including duplicates (Table 10). After removal
of duplicates (n = 2,978), 3,020 references were reviewed and excluded based on
information contained in the title and abstract. The remaining fifty eight articles
were retrieved and the full text of each was assessed for eligibility by two
researchers; 48 articles were excluded, because they did not meet the inclusion
criteria. Nine publications met all the inclusion criteria and were included for
analysis. Two publications (Wakefield 1996 & 2002) were included as one study;
one was a PhD thesis and the other the published article from the study. Refer to
Figure 3 for a flow diagram of search results. An outline of the search strategy is
presented in the Appendix 2.
74
Table 10. Database search results for systematic review
Database Records
Medline 967
Cinahl 336
PsycInfo 350
Informit Health collection 3
Proquest Health and Medical 400
Embase 1760
Scopus 2173
Proquest Dissertation and Thesis 33
Mednar 34
Total 6056
Duplicates 2978
Total (with duplicates removed) 3078
Figure 3. Flow diagram of the stages of searching
Records identified throughdatabase searching
(n = 6,056)Duplicate records removed
(n = 2,978)
Title and abstract recordsscreened(n = 3,078)
Records excluded(n = 3,020)
Full text articles assessed foreligibility(n = 58)
Full text articles excluded(n = 48)
Studies included in quantitativesynthesis(n = 10)
75
4.2.1.1. Excluded studies
The main reason for study exclusion was failure to assess the presence of delirium
on admission. It was unclear if patients had been admitted with prevalent delirium.
This resulted in an inability to differentiate between incident and prevalent delirium
in the reporting of results. Another reason for exclusion was the admission of
patients to surgical or intensive care units. Appendix 12 details the reason for
exclusion of each study for which the full text was retrieved.
4.2.2 Study characteristics
Characteristics of the included studies are summarised in Table 11. Most studies
originated from North America, two were from the United Kingdom and one from
Colombia. All studies investigated medical populations using either cohort studies
or case control study designs. All of the included studies screened for and excluded
patients with delirium on admission. The instruments used to diagnose delirium as
well as some of the main risk factors investigated are presented. Most studies used
the Confusion Assessment Method (CAM) to assess delirium. Only one study used
the NEECHAM confusion scale to assess delirium. Scales used to measure each risk
factor are also presented in Table 11. For example, the Mini Mental State Exam
(MMSE) was used to assess cognitive impairment in four of the studies.
76
Table 11. Characteristics of Studies Included in the Systematic Review
Authors Year Location Samplesize (n)
Deliriumassessmenttool
Risk Factors Riskfactormeasure
Alagiakrishnan
et al.
2009 Canada 132 CAM Cognition
Functioning
Mini Cog
ADL scale
Campbell et al. 2011 USA 147 CAM Cognition SPMSQ
Franco et al. 2010 Colombia 291 CAM & DRS
R98
Cognition MMSE
Inouye and
Charpentier
1996 USA 196 CAM Use of IDC
Use of
Restraints
Dehydration
Jones et al. 2006 USA 491 CAM Cognition
Illness
severity
MMSE
APACHE
McAvay et al. 2007 USA 416 CAM Depression GDS
O’Keeffe and
Lavan
1996 UK 100 DAS & DSM
3
Depression
Dementia
GDS
BDRs
Wakefield 2002
1996
USA 332 NEECHAM Cognition
Depression
MMSE
GDS
Wilson et al. 2005 UK 100 CAM &
DSM 3
Cognition
Depression
Illness
Severity
MMSE
GDS
APACHE
Note. CAM – Confusion Assessment Method. DSM – Diagnostic and Statistical Manual of Mental
Disorders. MMSE – Mini Mental State Exam. IDC – Indwelling Catheter. GDS – Geriatric Depression
Scale. DAS Delirium assessment scale. APACHE Acute Physiology and Chronic Health Evaluation,
ADL scale – Katz Idex of independence in Activities of Daily Living, SPMSQ – Short Portable Mental
Status Questionanaire, BDRS – Blessed Dementia Rating Scale
77
4.2.3 Methodological quality
All of the included studies were assessed for methodological quality using the
Joanna Briggs Institute (JBI) Critical Appraisal Tool (Appendix 3). This tool addresses
study quality in terms of minimisations of bias in relation to selection of cases and
controls, identification of confounding factors, assessing outcomes using objective
criteria, and using appropriate statistical analysis. To be considered for inclusion, a
study should meet most of the nine questions. Figure 4 presents the frequency with
which studies met the criteria on the JBI Critical Appraisal Tool.
All studies included in the review were prospective cohort or case control studies.
Of these, two studies involved secondary analysis of previous studies (Jones et al.
2006; McAvay et al. 2007). Eight studies (Alagiakrishnan et al. 2009; Campbell et al.
2011; Franco et al. 2010; Inouye & Charpentier 1996; Jones et al. 2006; McAvay et
al. 2007; O'Keeffe & Lavan 1996; Wilson et al. 2005) were representative of the
population as a whole as they investigated all patients that were admitted into the
hospital during a particular time period (Question 1 JBI appraisal tool). One study
investigated only male patients and thus was difficult to compare to other studies
and the results were not representative of the population as a whole (Wakefield
2002). All included studies used convenience sampling, recruiting patients admitted
to hospital and acutely ill. None of the studies provided a justification for sample
size using a power analysis. Study samples ranged from 100 491.
All of the studies used a valid measure of delirium (including CAM and NEECHAM
confusion scales) to select the cases (Question 8 JBI appraisal tool – Are outcomes
measured in a reliable way?). Most articles reported that risk factors for delirium
were investigated, but did not document possible confounding factors (Question 4
JBI appraisal tool). Only two articles reported that confounding factors were
identified and outlined the strategies used to address these (Inouye & Charpentier
1996; McAvay et al. 2007).
In most studies possible risk factors for delirium were assessed using valid and
reliable tools including: the Mini Mental State Exam (MMSE) to test for cognitive
78
impairment (Franco et al. 2010; Jones et al. 2006; McAvay et al. 2007; Wakefield
1996 & 2002); the Katz Index of Independence in Activities of Daily Living (Katz et al.
1970) for functional impairment (Alagiakrishnan et al. 2009; O'Keeffe & Lavan 1996;
Wakefield 1996 & 2002; Wilson et al. 2005); the Acute Physiology and Chronic
Health Evaluation (APACHE) for severity of illness (Inouye & Charpentier 1996; Jones
et al. 2006a; McAvay et al. 2007; Wilson et al. 2005); the Geriatric Depression Scale
(GDS) for depression (McAvay et al. 2007; O'Keeffe & Lavan 1996b; Wakefield 1996
& 2002; Wilson et al. 2005); and the Blessed Dementia Rating Scale (BDRS) to assess
for dementia (Jones et al. 2006; O'Keeffe & Lavan 1996). Common use of scales
enabled comparisons to be made between studies. In cases where these tools were
not used the authors relied on documentation of a condition by a medical
practitioner in the medical records. None of the included studies reported the
outcomes of patients who had withdrawn from the study (Question 6 JBI appraisal
tool) and follow up beyond the hospitalisation of the patient was not reported
(Question 7 appraisal tool). Tables reporting the critical appraisal results for each
study can be found in Appendix 13.
Figure 4. Critical appraisal of included studies
1 1 3 5 7 9
123456789
Number of studies that had a Yes response
JBIcriticalap
praisalque
stion
numbe
r
Critical appraisal of included studies
79
4.2.4 Results of included studies
Overall, there were 1,990 participants of which 320 developed delirium. This gives
an overall incidence rate of 16%. The risk factors investigated in the studies are
presented in Table 12. However, only risk factors identified by multiple studies were
pooled for analysis.
Table 12. Risk factors examined in included studies in systematic review
Study Risk factors identifiedAlagiakrishnanet al. (2009)
AgeCritical illnessHearing impairment
Abnormal mini cogscoreDementiaMale gender
Cognitive impairmentFunctional impairmentStrokeVisual impairment
Campbell et al.(2011)
AgeMale gender
Anticholinergicmedication
Cognitive impairment
Franco et al.(2010)
Acute renal failureMale genderYears of educationVisual impairment
AgePneumoniaUrinary tract infection
Cognitive impairmentUse of a bladdercatheter
Inouye andCharpentier(1996)
Concurrent illnessMedications
Iatrogenic eventsLength of hospitaladmission
Immobility
Jones et al.(2006)
AgeYears of education
Gender Dementia
O’Keeffe andLavan (1996)
Abnormal temperatureDementia (hx cognitiveimpairment)*Elevated serum urea *Severe Illness *
AlcoholAbnormal serumsodium*Dependent in >2 ADLsHypo – albuminaemiaLong term care patientMale gender
Age >80 yearsAbnormal white cellcountDepressionVisual Impairment(*Reached significance, usedin prediction model)
McAvay et al.(2007)
AntipsychoticmedicationDepressionFunctional statusMale gender
Blood urea nitrogen>18DementiaPrinciple diagnosis (lungdisease, pneumonia, heartfailure, heart disease,cancer, diabetes)
Cognitive impairmentHearing impairmentYears of educationVision impairmentSevere illness (APACHEscore >16)
Wakefield(1996 & 2002)
Activity levelsDementia Years ofeducationFunctional impairment
Cognitive impairmentMedicationsPainBlood urea nitrogen
DepressionInfection on day ofadmissionVisual impairment
Wilson et al.(2005)
AgePlasma levels of IGF 1
Cognitive impairmentMale Gender
DepressionSevere illness (APACHE11 >8)
Note. APACHE – Acute Physiology and Chronic Health Evaluation, ADLs – Activities of Daily Living
80
4.2.5 Meta analysis
Following assessment for heterogeneity of the included studies, it was possible to
conduct meta analyses on the following risk factors: dementia, functional
impairment, male gender, age > 80 years, visual impairment and pneumonia. These
factors were measured with the same measurement tools and thus were easily
comparable. Each factor was examined in terms of its impact on the incidence of
delirium in the medical population.
4.2.5.1 Dementia
Four articles reported data regarding the impact of dementia on the development
of delirium (Alagiakrishnan et al. 2009; McAvay et al. 2007; O'Keeffe & Lavan 1996;
Wilson et al. 2005). Studies were only included if they examined the number of
patients with dementia who developed a delirium compared to the number who did
not develop delirium. The chi square test result for heterogeneity was not
significant (p = .84), indicating studies were statistically homogenous and
appropriate to undertake meta analysis. Figure 5 presents the forest plot
incorporating fixed effects Mantel Haenszel Odds Ratio for dementia. The weight
for each study is indicated in the figure and the sum of all the weights equals 100%.
The study conducted by McAvay et al. (2007) had the greatest contribution
(43.63%) to the meta analysis. The results for all four studies are presented in the
forest plot. The overall analysis indicated that a diagnosis of dementia is
significantly related to development of incident delirium (Z = 5.49, OR 4.06, p
<.0001), with a relatively narrow confidence interval, suggesting greater accuracy of
the effect size. Findings suggest that patients with dementia had greater odds of
developing delirium compared to patients with no dementia.
81
Figure 5. Forest plot of the impact of dementia on development of incident delirium
4.2.5.2 Functional impairment
Three of the studies reported data regarding the impact of functional impairment
on the development of delirium. Studies were only included if they examined
patients with a functional impairment, as measured by Katz Index of Independence
in Activities of Daily Living (Katz et al. 1970), who developed a delirium compared to
patients who did not develop delirium. The chi square test result for heterogeneity
was not significant (p = .57), indicating studies were statistically homogenous and
appropriate to undertake meta analysis. Figure 6 presents the forest plot
incorporating the fixed effects Mantel Haenszel Odds Ratio for functional
impairment. The weight for each study is indicated in the figure and the sum of all
the weights equals 100%. The study conducted by O’Keeffe and Lavan (1996) had
the greatest contribution (43.75%) to the meta analysis. The results for all three
studies are presented in the forest plot. Overall analysis indicated that functional
impairment is significantly related to development of incident delirium (Z = 2.02, OR
1.75, p < .04). Patients with functional impairment have a 75% increased likelihood
of developing incident delirium compared to patients with no functional
impairment.
Another study (Wakefield 1996 & 2002) examined the relationship between
functional impairment and delirium. However, Wakefield (1996 & 2002) only
82
included male patients, thus it could not be included in the meta analysis.
Wakefield (1996 & 2002) also used the Katz Index of Independence in Activities of
Daily Living (Katz et al. 1970) to measure functional impairment and identified that
patients who developed incident delirium were dependent on admission for an
average of 2.5 Activities of Daily Living. This is compared to those who did not
develop delirium and were dependent for 0.4 Activities of Daily Living (p < .0005).
These data indicate that patients who developed delirium had greater functional
impairment and required more functional assistance.
Figure 6. Forest plot of the impact of functional impairment on the development of
incident delirium
4.2.5.3 Male gender
Six studies (Alagiakrishnan et al. 2009; Campbell et al. 2011; Franco et al. 2010;
McAvay et al. 2007; O'Keeffe & Lavan 1996; Wilson et al. 2005) reported data
regarding the impact of male gender on the development of delirium. Studies
reported the number of male patients who developed a delirium compared to the
number who did not develop delirium. Female comparison was not reported in the
studies. The chi square test result for heterogeneity was not significant (p = .07),
indicating studies were statistically homogenous and appropriate to undertake
meta analysis. Figure 7 presents the forest plot incorporating fixed effects Mantel
Haenszel Odds Ratio for male gender. The weight for each study is indicated in the
figure and the sum of all the weights equals 100%. The study conducted by McAvay
et al. (2007) had the greatest contribution (26.89%) to the meta analysis. The
83
results for all six studies are presented in the forest plot. The overall analysis
indicated that male gender was not significantly related to the development of
incident delirium (Z = 0.60, OR 1.11, p = .55). This suggests that male gender is not a
risk factor for delirium.
Figure 7. Forest plot of the impact of male gender on development of incident
delirium
4.2.5.4 Visual impairment
Three studies (Franco et al. 2010; McAvay et al. 2007; O'Keeffe & Lavan 1996)
reported data regarding the impact of visual impairment on the development of
delirium. Studies were only included if they examined the number of patients with a
visual impairment who developed a delirium compared to the number who did not
develop delirium. The chi square test result for heterogeneity was not significant (p
= .51), indicating studies were statistically homogenous and appropriate to
undertake meta analysis. Figure 8 presents the forest plot incorporating fixed
effects Mantel Haenszel Odds Ratio for visual impairment. The weight for each
study is indicated in the figure and the sum of all the weights equals 100%. The
study conducted by Franco et al. (2010) had the greatest contribution (46.42%) to
the meta analysis. The results for all three studies are presented in the forest plot.
Overall analysis revealed that visual impairment is approaching significance in the
development of incident delirium (Z = 1.89, OR 1.62, p = .059). Those with visual
impairment appear to have some risk of developing incident delirium.
84
Figure 8. Forest plot of the impact of visual impairment on development of incident
delirium
4.2.5.5 Pneumonia
Two studies (Franco et al. 2010; McAvay et al. 2007) reported data regarding the
impact of the presence of pneumonia on admission on development of incident
delirium. Studies were only included if they examined the number of patients who
had pneumonia on admission and developed delirium compared to the number
who did not develop delirium. The chi square test result for heterogeneity was not
significant (p = .46), indicating studies were statistically homogenous and
appropriate to undertake meta analysis. Figure 9 presents the forest plot
incorporating fixed effects Mantel Haenszel Odds Ratio for pneumonia. The weight
for each study is indicated in the figure and the sum of all the weights equals 100%.
The study conducted by Franco et al. (2010) had the greatest contribution (64.24%)
to the meta analysis. The results for both studies are presented in the forest plot.
Overall analysis did not find a significant relationship between pneumonia and the
development of incident delirium (Z = 0.78, OR 1.29, p = .43), suggesting patients
are unlikely to develop delirium based only on the presence of pneumonia.
85
Figure 9. Forest plot of the impact of pneumonia on the development of incident
delirium
4.2.5.6 Age > 80 years
Two studies (Alagiakrishnan et al. 2009; O'Keeffe & Lavan 1996) explored the
impact of advanced age (>80 years) on the development of incident delirium. These
studies examined the number of patients over the age of 80 who developed a
delirium compared to the number of patients over the age of 80 who did not
develop delirium. The chi square test result for heterogeneity was not significant (p
= .48), indicating studies were statistically homogenous and appropriate to
undertake meta analysis. Figure 10 presents the forest plot incorporating fixed
effects Mantel Haenszel Odds Ratio for age greater than 80. The weight for each
study is indicated in the figure and the sum of all the weights equals 100%. The
study conducted by Alagiakrishnan et al. (2009) had the greatest contribution
(61.16%) to the meta analysis. The results for both studies are presented in the
forest plot. Overall analysis shows that advanced age (> 80 years) is not significantly
related to the development of incident delirium (Z = .96, OR 1.42, p = .33),
suggesting patients are unlikely to develop delirium based only on being over 80
years of age.
Four studies (Campbell et al. 2011; Franco et al. 2010; McAvay et al. 2007; Wilson et
al. 2005) also reported the average ages of patients who developed delirium and
those that did not. The results are presented in Table 13. Those who developed
delirium were older (M = 80.46, SD = 3.57) than those who did not (M = 78.70, SD =
86
4.53). However, independent samples t test analysis comparing the two
populations, with equal variances assumed, did not find this difference to be
statistically significant, t(6) = 0.61, p = .57.
Figure 10. Forest plot of impact of age > 80 years on the development of incidentdelirium
Table 13. Average ages of patients with and without delirium
Delirium No DeliriumStudy Mean years SD Mean years SD t df p
Franco et al. (2010) 78.35(N = 34)
8.96 73.88(N = 257)
8.65 2.824
289
.005
McAvay et al. (2007) 82.3(N = 36)
6.6 79.9(N = 380)
6.5 NA NA .04
Campbell et al.(2011)
76.7(N = 33)
8.2 76.6(N = 114)
8.0 NA NA .96
Wilson et al. (2005) 84.5(N = 12)
4.19 84.41(N = 88)
4.21 .070 98 .94
Overall Average 80.46 3.57 78.70 4.53 .61 6 .57Note. NA = not reported.
87
4.2.6 Narrative synthesis
The following outcomes were not suitable for meta analysis and are therefore
presented as a narrative synthesis.
4.2.6.1 Cognitive impairment
Six of the nine studies (Alagiakrishnan et al. 2009; Campbell et al. 2011; Franco et al.
2010; McAvay et al. 2007; Wakefield 1996 & 2002; Wilson et al. 2005) examined the
relationship between cognitive impairment and incident delirium. Four of the
studies used the Mini Mental State Exam (MMSE) (Franco et al. 2010; McAvay et al.
2007; Wakefield 1996 & 2002; Wilson et al. 2005) as an indicator for cognitive
impairment, with a score of < 24 indicating a cognitive impairment. One study used
the Short Portable Mental Status Questionnaire (SPMSQ) (Campbell et al. 2011).
The remaining study used the Mini cog (Alagiakrishnan et al. 2009).
Three of the studies that used the MMSE as a measure for cognitive impairment
reported average scores for those that developed delirium and those that did not
develop delirium. Table 14 summarises the results of the average scores for each
individual study and that of the three studies combined. T test analysis for each
study showed a significant relationship between cognitive impairment (lower score
on the MMSE) and delirium. Pooling these average scores using an independent
samples t test, with equal variance assumed, also shows a significant difference in
MMSE scores of patients who developed a delirium and those that did not, t(4) =
3.93, p = .017.
Table 14. Average MMSE scores and t test results for patients with and withoutdelirium
Delirium No Delirium
Study Mean SD Mean SD t df pFranco et al. (2010) 20.65 4.65 24.23 4.01 4.276 289 < .001
Wakefield (1996 & 2002) 22.10 4.3 25.20 3.5 3.14 115 < .005
Wilson (2005) 22.75 3.39 26.35 3.58 3.29 98 .001
Average scores 21.83 1.08 25.26 1.06 3.93 4 .017
88
The remaining study that used the MMSE to assess cognitive impairment (McAvay
et al, 2006) did not report the actual average scores of each group. Instead the
article reports the number of individuals in each group that had a MMSE score less
than 24 (indicating cognitive impairment). A significant difference was found
between patients that developed delirium (N = 36), where 28 had a score less than
24 on the MMSE compared to patients with no delirium (N = 380) where 156
patients had a score less than 24 (p = .001). This finding indicates that patients with
cognitive impairment were more likely to develop incident delirium.
The study conducted by Campbell et al. (2011) measured cognitive impairment
using the Short Portable Mental Status Questionnaire (SPMSQ). Patients who had a
score of 8 or less were considered to have cognitive impairment. Patients in the
group that developed delirium (N = 33) had a lower SPMSQ score (M = 4.7, SD =
2.7), compared to the patients with no delirium (N = 114,M = 6.1, SD = 2.3). This
represented a significant difference between the groups (p = .007), suggesting that
cognitive impairment was associated with development of incident delirium.
The study conducted by Alagiakrishnan et al. (2009) identified cognitive impairment
as an abnormal score on the mini cog test (not being able to draw a clock face with
all numbers and/or being unable to recall 3 objects). The results of the study are
presented in Table 15. A diagnosis of cognitive impairment in the participant’s
history was not significantly related to delirium development (p > .99). However,
abnormal scores on the mini cog and an abnormal clock drawing test was significant
(p = .004 and p = .003, respectively), further suggesting that cognitive impairment is
associated with incident delirium.
Table 15. Delirium incidence vs. no delirium for cognitive impairment tests
Test Delirium(N = 20)
No Delirium(N = 112)
p
Cognitive impairment diagnosis 2 12 >.99
Abnormal mini – cog score 13 32 .004
Abnormal clock drawing 17 55 .003
Alagiakrishnan et al. (2009)
89
4.2.6.2 Depression
Four of the studies included in the review investigated the relationship between
depression and incident delirium (McAvay et al. 2007; O'Keeffe & Lavan 1996;
Wakefield 1996 & 2002; Wilson et al. 2005). All of the studies used the Geriatric
Depression Scale (GDS) to measure depression. However, only Wakefield (1996 &
2002) used the long version of the scale. Wakefield (1996 & 2002) did not report
actual scores on the GDS but stated that there was not a significant difference in
scores between those that developed delirium and those that did not. O'Keeffe and
Lavan (1996) recorded depression as a score of 5 or more on the GDS. Results
indicated that 7% (N = 2) of patients with delirium and 7% (N = 5) of patients
without delirium were depressed on admission (OR 1.0). This was not a significant
finding.
The other two studies found that depression was related to incident delirium.
Wilson et al. (2005) found that those with depression scoring three or more on the
GDS were seven times more likely to develop incident delirium (Z = 2.84, OR 7.14, p
= .005). The actual score differences between the groups was not reported. Lastly,
McAvay et al. (2007) found that patients who developed incident delirium reported
on admission an average 5.7 depressive symptoms compared to 4.2 depressive
symptoms for those who did not develop delirium (Hazards Ratio (HR) 1.1, 95% CI
1.0 1.2, p = .01). Furthermore, McAvay et al. (2007) found patients who developed
incident delirium were significantly more likely to report more symptoms of
dysphonic mood (M = 2.2 symptoms for delirious group, andM = 1.3 symptoms for
non delirious group, HR = 1.3, 95% CI = 1.1 1.5, p = .001) and hopelessness (M = 1.3
symptoms for delirious group, andM = 0.8 symptoms for non delirious group, HR =
1.5, 95% CI = 1.1 – 2.0, p = .006), indicating patients who are currently depressed
may be more at risk of developing incident delirium.
4.2.6.3 Years of education
Three of the nine studies (Franco et al. 2010; Jones et al. 2006; McAvay et al. 2007)
examined the relationship between years of education and incident delirium. Table
16 presents the results for the three studies that reported average years of
90
education. Combining the results of all three studies finds that years of education
was not significantly different between the delirium (M = 7.79, SD = 3.06) and non
delirium groups (M = 9.13, SD = 3.49), t(4) = 0.50, p = .64. However, the study by
Franco et al. (2010) was conducted in Colombia where the years of education for
both the patients with and without a delirium is significantly less than in the studies
conducted in the USA by Jones et al. (2006) and McAvay et al. (2007). Taking this
into account, examining only the studies conducted in the USA, patients in the
delirium group (M = 9.55, SD = .21) had significantly less years of education
compared to the control group patients with no delirium (M = 11.15, SD = .05), t(2)
= 10.12, p = .04, suggesting that education level has some relationship with
incident delirium.
Table 16. Average years of education for patients with and without delirium
Delirium No DeliriumStudy Mean SD Mean SD t df pMcAvay et al (2007) 9.7 3.8 11.1 3.7 NA NA .01
Franco et al (2010) 4.26 3.36 5.10 4.48 NA NA NA
Jones et al (2006) 9.4 3.9 11.2 3.7 NA NA < .001
Overall Average 7.79 3.06 9.13 3.49 0.50 4 .64
Note. NA = not reported.
4.2.6.4 Blood urea nitrogen (BUN)
Three of the nine studies reported data on blood urea nitrogen (BUN) levels and the
development of delirium. O’Keeffe and Lavan (1996) state that the criteria used for
an abnormal serum urea was levels > 10 mmol/L. Sixty eight percent (n = 19) of
patients with delirium met this criteria for an elevated serum urea compared to
only 31% (n = 22) of patients who did not develop delirium (adjusted OR 5.6, 95% CI
= 1.7 – 14.9). McAvay et al. (2007) reported the number of patients with a
BUN/creatinine > 18mmol/L. Twenty five patients in the delirium group compared
to 232 patients in the control group had a BUN/creatinine >18mmol/L and this was
not a statistically significant difference (HR 1.4, p = .33). Wakefield (1996 & 2002)
reported the mean BUN/creatinine ratio. Patients in the delirium group had higher
BUN/creatinine ratio (M = 22.9, SD = 9.4) compared to patients who did not develop
91
delirium (M = 17.0, SD 7.3), t(105) = 2.88, p < .005. These studies show mixed
results with two of the studies reporting positive relationships between delirium
and high BUN/creatinine ratio.
4.2.6.5 Use of indwelling urinary catheter
Two of the nine studies reported results for the use of an indwelling urinary
catheter and the effect on development of delirium. Inouye and Charpentier (1996)
found that when an IDC was present 36% (n = 18) of patients developed a delirium
compared to 12% (n = 17) of patients who developed a delirium when an IDC was
not used (RR 3.1, 95% CI = 1.7 – 5.5). Franco et al. (2010) found that 14.7% (n = 5) of
patients in the delirium group and 8.9% (n = 23) of patients in the group with no
delirium required a bladder catheter by day seven ( 2 = 1.144, df = 1, p = .285). The
results are mixed and more evidence is needed to determine the strength of the
relationship.
4.2.6.6 Severe illness
Three studies reported data regarding severe illness and delirium (McAvay et al.
2007; O'Keeffe & Lavan 1996; Wilson et al. 2005). O’Keeffe and Lavan (1996)
defined an illness as severe based on the initial assessment by the study physician.
The definition of severe illness was subjective, based on the physician’s expertise
and what they deemed to be severe illness. The authors found that 50% (n = 14) of
patients with delirium had been assessed as having a severe illness (examples of the
types of severe illness are not reported in the study) compared to 18% (n = 13) of
patients with no delirium (adjusted OR 5.6, 95% CI = 1.7 – 18.2). McAvay et al.
(2007) used the Acute Physiology and Chronic Health Evaluation (APACHE) scoring
system, which is a classification system that measures the severity of disease and
reflects the burden of acute illness on the patient. McAvay et al. (2007) found a
significant difference in scores between patients with a delirium (M = 17.8, SD = 4.9)
compared to control patients (M = 15.5, SD = 4.1, HR 1.1, 95% CI 1.1 – 1.2, p = .001).
Wilson et al. (2005) used the APACHE II to assess physical illness. No significant
difference was found between patients with delirium (M = 10.92) compared to the
control patients (M = 10.97, Z = 0.79, p = .48).
92
There are some differences between the APACHE and the APACHE II that should be
highlighted. APACHE II is an updated version of the original scale. Infrequently
measured variables were eliminated from the APACHE and the weighting of some of
the variables have been changed as a result (Bouch & Thompson 2008). These slight
differences may change the overall result of the severity of the illness and should be
considered as a factor that contributed to the difference in results between the two
studies.
Two studies reported data regarding the impact of chronic co morbid illness on
incident delirium (Campbell et al. 2011; McAvay et al. 2007). The Charlson Co
morbidity Index (CCI) is a scale used to predict mortality by weighing or classifying
pre existing conditions. Points are assigned to different medical conditions with
more serious conditions gaining more points. The scores stated in the two studies
reporting CCI scores are shown Table 17.
Independent samples t test, assuming unequal variance, found that a statistically
significant relationship (t(2) = 0.232, p = .84) between co morbid illness or
increased illness severity and delirium does not exist. In fact, the group with no
delirium had on average a higher score on the CCI than the group that did develop
delirium
Table 17. Average scores for patients using the Charslon co morbidity index with andwithout delirium
Delirium No Delirium
Study Mean SD Mean SD t df pMcAvay et al (2007) 3.1 2.2 2.6 2.1 NA NA .20
Campbell et al (2011) 2.4 1.9 3.1 2.3 NA NA .13
Average 2.75 0.49 2.85 0.35 0.232 2 .84
Note: NA = not reported
93
4.2.7 Systematic review results summary
Nine studies were included in the systematic review that reported data regarding
risk factors for incident delirium in the medical patient population. Meta analyses
and narrative synthesis have identified factors with a relationship to incident
delirium. A diagnosis of dementia or cognitive impairment is significantly related to
incident delirium in the medical in patient population. Therefore, patients with
dementia or a cognitive impairment are at a high risk of incident delirium. A
moderate relationship was also found between delirium and functional impairment,
indicating that patients who require assistance with activities of daily living are at
increased risk of delirium. Years of education, BUN/creatinine ratio and depression
were mildly associated with incident delirium. Visual impairment needs to be
studied further as results were approaching significance. Use of an indwelling
catheter as well as illness severity also require further research to determine if they
are related to delirium. Factors that were not found to increase risk of incident
delirium were male gender, age greater than 80, and pneumonia.
94
4.3 Phase 2 – Retrospective case control study results
In the following section baseline characteristics of patients in the case and control
groups will be presented, as well as the characteristics of a separate group that was
identified to have documented evidence of a possible delirium during
hospitalisation. Results of bivariate and multivariate (logistic regression) analyses
will also be presented. Additionally, descriptive data regarding the prevention and
management strategies used and medication management of delirium will be
presented.
4.3.1 Identification of cases
A representative from the hospital decision support services (information
management services) retrieved the record numbers for all patients whose
discharge summary contained a code for a diagnosis of delirium in 2012 and 2013.
These codes included: delirium unspecified, delirium superimposed on dementia,
delirium not superimposed on dementia, and other delirium. It was not clear what
criteria were used for coding into these categories.
For the year 2012, 125 patients’ records had been coded in the discharge summary
for delirium. The codes used were: delirium unspecified (n = 119), delirium
superimposed on dementia (n = 2), delirium not superimposed on dementia (n = 2)
and other delirium (n = 2). Records of patients were excluded if the patient did not
meet inclusion criteria as previously described in the Methods chapter. Based on
these criteria, 46 patient records were excluded. The remaining 79 patient records
were included as cases for the year 2012. During the record review for the control
group (to be described below), a further four cases were identified as having a
documented medical diagnosis of delirium during their hospital stay, however those
patients had not been coded for delirium in their discharge summaries. Their
records were included as cases.
For the year 2013, 168 patients were diagnosed with delirium during admission and
on discharge were coded as either delirium unspecified (n = 154), delirium
95
superimposed on dementia (n = 7), delirium not superimposed on dementia (n = 2)
and other delirium (n = 5). Based on the exclusion criteria, 91 patient records were
excluded. The remaining 77 patient records were included as cases for the year
2013. In screening the control group (to be described below), one further case was
identified as having a documented medical diagnosis of delirium during their
hospital stay, however they had not been coded for delirium in their discharge
summary. This record was also included as a case.
4.3.2 Identification of controls
A hospital decision support service representative (information management
services) compiled a random sample of patient record numbers. Control group
patients were included if they did not have a diagnosis code for delirium on
admission, were aged 42 years and older (based on age range of delirium group)
and were admitted to a medical setting during 2012 2013. Control group records
were excluded based on the exclusion criteria outlined in the Methods chapter. In
total, 321 control group records were reviewed. During the data extraction for the
control group, a number of patients were identified as having possible signs of
delirium, and thus were unable to satisfy the criteria for having no delirium. As a
result, for the year 2012, 31 patient records were identified as including
documented evidence of possible signs of delirium and were excluded. However,
these records were included separately as a possible delirium group. For the year
2013, 17 patients were identified as having documented evidence of possible signs
of delirium during admission and were therefore excluded from the control group
and instead included in the possible delirium group. Figure 11 illustrates the process
of identifying cases and controls for this study.
96
Figure 11. Identification of cases and controls
4.3.3 Characteristics of patients
The characteristics of the patients that developed delirium during admission (case
group) and those of patients that did not develop a delirium during admission
(control group) are displayed in Table 18. Forty eight patients, who were not
diagnosed with delirium but had documented evidence of signs of possible delirium
during admission, are included as a separate possible delirium group.
The mean age of the case group was 84 years (SD = 7.4, range 42 to 100 years) the
control group was 77 years (SD = 11.8, range 42 to 98 years), and the possible
delirium group was 83 years (SD = 7.6, range 63 to 99). A relatively even distribution
of gender across the groups was evident, with a slightly higher percentage of
females in both groups. Patients in the control group were more likely to be
admitted from home with a spouse or other family members present (55.5%),
compared to patients in the delirium group (44.7%). Patients in the control group
were also less likely to live at home alone (26.8%) when compared to the patients in
the delirium group (30.4%). Patients in the ‘possible delirium’ group mostly lived at
home with family (54.2%) rather than alone (18.8%).
97
The patients’ level of functioning was assessed retrospectively using the Katz Index
of Independence in Activities of Daily Living instrument (Katz et al. 1970). Based on
admission assessments documented by nursing or allied health staff, patients were
given a score rating using the Katz scale. Areas that are assessed include needing
assistance with bathing, dressing, toileting and feeding. Patients in the control
group had a higher level of independent functioning, with almost 80% being
independent with their activities of daily living prior to admission, compared to only
54% of cases, and 45.8% of patients with possible delirium. A higher percentage of
cases (26.1%) and possible delirium patients (27.1%) required some assistance with
the performing two or more activities of daily living compared to the control group
(11.8%).
Most patients in the control group were reported as having no cognitive
impairment or normal cognition on admission to hospital (76.6%). This is contrasted
to the cases, among whom less than half were documented to have normal
cognition (47.2%), and even fewer in the possible delirium group (43.8%) were
documented to have normal cognition. On admission, a higher percentage of cases
and possible delirium patients were reported as having some memory loss (24.8%
and 27.1% respectively), compared to the control group (15.3%). Cases and patients
with possible delirium were also more likely to have a past history of dementia.
Patients in the control group were less likely to have had a previous fall. For 47.8%
of cases and 45.8% of patients with possible delirium, two falls or less in the last six
months were reported, compared to 34.6% in the control group. For four cases, a
falls risk assessment was not available and no other evidence of falls risk
assessment could be located in the medical history.
98
Table 18. Admission characteristics of patients in the case, control and possibledelirium groups
Characteristics
CaseDeliriumN = 161n (%)
Control
N = 321n (%)
PossibleDeliriumN = 48n (%)
Mean age (SD) 84.11 (7.4) 77.69 (11.8) 83.98 (7.6)GenderFemale 94 (58.4) 172 (53.6) 25 (52.1)Male 67 (41.6) 149 (46.4) 23 (47.9)Living status on admissionHome alone 49 (30.4) 86 (26.8) 9 (18.8)Home with services 9 (5.6) 14 (4.4) 1 (2.1)Living with family 72 (44.7) 178 (55.5) 26 (54.2)Low level care 18 (11.2) 27 (8.4) 8 (16.7)High level care 13 (8.1) 16 (5.0) 4 (8.3)Level of functioningIndependent with all ADL (Katz score 5 6) 88 (54.7) 255 (79.4) 22 (45.8)Independent with most ADL but require someassistance (Katz score 3 4)
42 (26.1) 38 (11.8) 13 (27.1)
Assistance required for most ADL (Katz score2)
17 (10.6) 23 (7.2) 12 (25.0)
Full assistance required (Katz score 0) 14 (8.7) 5 (1.6) 1 (2.1)Reported cognition on admissionCognition reported as normal 76 (47.2) 246 (76.6) 21 (43.8)Reported cognitive issues (no diagnosis) 13 (8.1) 6 (1.9) 4 (8.3)Mild cognitive impairment 6 (3.7) 0 2 (4.2)Some memory loss 40 (24.8) 49 (15.3) 13 (27.1)Dementia 26 (16.1) 20 (6.2) 8 (16.7)Falls history (prior to admission)No falls 48 (29.8) 186 (57.9) 19 (39.6)Less than 2 in 6 months 77 (47.8) 111 (34.6) 22 (45.8)More than 2 in 6 months 32 (19.9) 24 (7.5) 7 (14.6)No data 4 (2.5) 0 0Pressure risk score (on admission)Low 90 (55.9) 222 (69.2) 25 (52.1)Medium 47 (29.2) 61 (19.0) 12 (25.0)High 21 (13.0) 31 (9.7) 10 (20.8)No data 3 (1.9) 7 (2.2) 1 (2.1)
Note. ADL = Activities of Daily Living.
99
4.3.3.1 Reason for admission
The primary and secondary reasons for admission of the case, control and possible
delirium groups are presented in Table 19. Admission diagnoses were recorded
based on the diagnosis listed on the discharge summary. For some patients the
discharge summaries were not available. If so, the diagnosis documented during the
patient’s admission was used. Across the groups, it was common to be admitted
with a respiratory infection, a urinary infection, as well as for cardiac reasons,
including myocardial infraction, congestive cardiac failure and atrial fibrillation.
100
Table 19. Primary and secondary admission diagnosis
Admission Diagnosis
Primary diagnosis Secondary diagnosisCaseDeliriumN = 161n (%)
Control
N = 321n (%)
PossibleDeliriumN = 48n (%)
CaseDeliriumN = 161n (%)
Control
N = 321n (%)
PossibleDeliriumN = 48n (%)
Abdominal Pain 2 (1.2) 1 (0.3) 0 0 3 (0.9) 0Allergic Reaction 0 4 (1.2) 0 0 0 0Anaemia 5 (3.1) 5 (1.6) 0 1 (0.6) 1 (0.3) 1 (2.1)Back Pain 2 (1.2) 17 (5.3) 4 (8.3) 1 (0.6 5 (1.6) 0Car Accident 0 2 (0.6) 0 0 0 0Cardiac AF 4 (2.5) 10 (3.1) 0 0 0 0Cardiac CCF 10 (6.2) 24 (7.5) 2 (4.2) 2 (1.2) 7 (2.2) 1 (2.1)Cardiac MI 4 (2.5) 11 (3.4) 2 (4.2) 5 (3.1) 14 (4.4) 1 (2.1)Cardiac Other 0 1 (0.3) 0 0 0 0Cellulitis 3 (1.9) 12 (3.7) 1 (2.1) 1 (0.6) 6 (1.9) 1 (2.1)Collapse 19 (11.8) 11 (3.4) 5 (10.4) 1 (0.6) 4 (1.2) 0Constipation 2 (1.2) 3 (0.9) 0 0 0 0Dehydration 0 0 0 1 (0.6) 1 (0.3) 1 (2.1)Dizziness/Vertigo 0 1 (0.3) 0 0 1 (0.3) 1 (0.3)Diabetic Issue 1 (0.6) 2 (0.6) 0 0 6 (1.9) 0Electrolyte Imbalance 1 (0.6) 5 (1.6) 0 6 (3.7) 3 (0.9) 0Fall 48 (29.8) 51 (15.9) 17 (35.4) 3 (1.9) 2 (0.6) 0Fracture 3 (1.9) 3 (0.9) 0 27 (16.8) 28 (8.7) 10 (20.8)Generalised Weakness 1 (0.6) 4 (1.2) 2 (4.2) 2 (1.2) 5 (1.6) 1 (2.1)GastrointestinalObstruction
0 2 (0.6) 1 (2.1) 0 0 0
Haematoma 1 (0.6) 0 2 (4.2) 0 3 (0.9) 0Infection/sepsis 1 (0.6) 8 (2.5) 1 (2.1) 3 (1.9) 5 (1.6) 0Leukaemia 0 0 0 0 1 (0.3) 0Medication Related 0 1 (0.3) 0 0 0 0Pain 4 (2.5) 8 (2.5) 0 0 13 (4.0) 1 (2.1)Respiratory Exacerbation 4 (2.5) 19 (5.9) 2 (4.2) 0 3 (0.9) 0Renal Failure 2 (1.2) 7 (2.2) 0 8 (5.0) 13 (4.0) 2 (4.2)Respiratory Infection 29 (18.0) 58 (18.1) 5 (10.4) 12 (7.5) 12 (3.7) 6 (12.5)Reduced Mobility 1 (0.6) 1 (0.3) 1 (2.1) 0 0 0Respiratory Effusion 1 (0.6) 4 (1.2) 0 0 0 0Respiratory Embolism 0 3 (0.9) 0 0 1 (0.3) 0Seizure 0 1 (0.3) 0 0 1 (0.3) 0Stroke 2 (1.2) 11 (3.4) 1 (2.1) 3 (1.9) 1 (0.3) 1 (2.1)Urinary Infection 6 (3.7) 9 (2.8) 1 (2.1) 7 (4.3) 9 (2.8) 4 (8.3)Urine Retention 1 (0.6) 2 (0.6) 0 0 1 (0.3) 0Vomit/Diarrhoea 4 (2.5) 17 (5.3) 1 (2.1) 4 (2.5) 3 (0.9) 0Weight Loss 0 0 0 0 2 (0.6) 0Nil other 0 0 0 74 (46.0) 167 (52.0) 15 (31.3)
Note. AF – Atrial Fibrillation, CCF – Congestive Cardiac Failure, MI – Myocardial Infarction
101
4.3.4 Risk factors for incident delirium
Chi square tests and logistic regression models were used to examine the
relationship between possible predisposing and precipitating risk factors and the
development of delirium in patients admitted to the medical setting.
4.3.4.1 Predisposing risk factors
A chi square test for independence (with Yates Continuity Correction) was used to
examine the relationship between possible predisposing risk factor variables. Table
20 displays the predisposing risk factors tested to examine their relationship with
delirium. Some factors that have previously been identified by other researchers
and in the systematic review as risk factors for delirium were not significant in this
population. Depression (OR 1.41, p = .20), hearing impairment (OR 1.30, p = .37) and
male gender (OR 0.82, p = .37) were not significantly related to the development of
delirium in this population. Other factors such as dementia (OR 2.90, p = .001),
cognitive impairment (OR 3.01, p < .000), functional impairment (OR 3.05, p = .000),
and a previous delirium (OR 17.60, p < .000) were significantly related to the
development of delirium. Despite not being found to be significant in the systematic
review in Phase 1 of the research, bivariate analysis showed age > 80 years (OR
2.69, p < 0.000) to be significant in this population. A past history of stroke or
transient ischemic attack (TIA) (OR 2.3, p = .001), osteoporosis (OR 1.61, p = .04),
Parkinson’s disease (OR 2.79, p = .03), and hypercholesterolemia (OR 1.61, p = .03)
were also found to have a significant relationship with incident delirium.
Admission diagnoses of fall and fracture, found to be significantly different between
the groups, were further examined using a chi square test. Analysis found that
these were both significantly related to the development of incident delirium (Fall
OR 2.34, p < .000, Fracture OR 2.06, p = .013).
102
Table 20. Chi square test results for possible predisposing risk factors for delirium
Predisposing Factor*
Case Delirium(N = 161)n (%)
Control(N = 321)n (%)
Odds Ratio (95 % CI) df Pearson’s 2 value p
Age > 80 years 124 (77.0) 178 (55.5) 2.69 (1.75 – 4.13) 1 20.40 .00Anaemia 6 (3.7) 13 (4.0) 1.09 (0.41 – 2.92) 1 0.00 1.0Cancer 21 (13.0) 44 (13.7) 0.94 (0.54 – 1.65) 1 .004 .95Cognitive Impairment 61 (37.9) 54 (16.8) 3.01 (1.96 – 4.65) 1 25.05 .00COAD 27 (16.8) 77 (24.0) 1.57 (0.96 – 2.55) 1 2.89 .09Depression 35 (21.7) 53 (16.5) 1.41 (0.87 – 2.26) 1 1.63 .20Dementia 26 (16.1) 20 (6.2) 2.90 (1.56 – 5.37) 1 11.10 .001Diabetes 40 (24.8) 72 (22.4) 0.86 (0.56 – 1.36) 1 0.23 .63Functional Impairment 71 (44.1) 66 (20.6) 3.05 (2.02 – 4.60) 1 28.06 .00Fall on Admission 51 (31.7) 53 (16.5) 2.34 (1.50 – 3.65) 1 13.70 .00Fracture on Admission 29 (18.0) 31 (9.7) 2.06 (1.19 – 3.55) 1 6.12 .013Gender (Male) 67 (41.6) 149 (46.4) 0.82 (0.56 – 1.21) 1 0.82 .37Hearing Impairment 30 (18.6) 48 (15.0) 1.30 (0.79 – 2.15) 1 0.82 .37Hypertension 92 (57.1) 185 (57.6) 1.02 (0.70 – 1.49) 1 0.00 .99Hypercholesterolemia 54 (33.5) 76 (23.7) 1.61 (1.08 – 2.44) 1 4.81 .03Ischemic Heart Disease 28 (17.4) 51 (15.9) 0.90 (0.54 – 1.49) 1 0.08 .77Joint Replacement 22 (13.7) 27 (8.4) 1.72 (0.95 – 3.13) 1 2.70 .10Osteoporosis 59 (36.6) 87 (27.1) 0.64 (0.43 – 0.96) 1 4.18 .04Parkinson’s Disease 12 (7.5) 9 (2.8) 2.79 (1.15 6.76) 1 4.50 .03Previous Delirium 16 (9.9) 2 (0.6) 17.54 (3.99 – 77.55) 1 23.35 .00Recent Admission to Hospital 59 (36.6) 137 (42.7) 0.77 (0.53 – 1.15) 1 1.37 .24Renal Failure 19 (11.8) 34 (10.6) 1.13 (0.62 – 2.05) 1 0.06 .81Stroke or TIA 38 (23.6) 38 (11.8) 2.30 (1.40 – 3.78) 1 10.31 .001Visual Impairment 31 (19.3) 39 (12.1) 1.72 (1.03 – 2.89) 1 3.81 .05
Note. *Condition present (used as reference group for odds ratio), Significant p value. TIA transient ischaemic attack.
103
4.3.4.1.1 Predisposing risk factors for possible delirium group compared with
control group
While not an original intention of this phase of the research, a number of possible
delirium cases were identified during data collection. Analysis was also performed
on the possible delirium group compared to the control group to determine if
patients with possible delirium possessed similar risk factors to the cases. A chi
square test for independence (with Yates Continuity Correction) was used to
examine the relationship between predisposing factors in the possible delirium
group and the control group. Table 21 displays the chi square test results for
predisposing risk factors for the possible delirium group compared to the control
group.
The comparisons between the patients with possible delirium and control patients
for predisposing factors were similar to those for cases and control patients. Factors
found to be significantly different between the groups were: dementia (OR 3, p =
.02), cognitive impairment (OR 3.2, p < .00), functional impairment (OR 4.6, p < .00),
age > 80 (OR 3.5, p = .001), fall on admission (OR 2.8, p = .004), and fracture on
admission (OR 2.4, p = .04). This suggests that patients with these factors had
greater odds of showing signs of delirium. These factors were also significant for the
case and control comparisons as reported previously. No patients with possible
delirium had a documented past history of a delirium and bivariate analysis was not
conducted on this factor.
104
Table 21. Chi square statistic results of predisposing risk factors for patients for possible delirium and control group.
Predisposing Factor*
Possible Delirium(N = 48)n (%)
Control(N= 321)n (%) Odds Ratio (95 % CI)
df Pearson’s 2
valuep
Age > 80 years 39 (81.3) 178 (55.5) 3.48 (0.63 – 7.43) 1 10.43 .001Anaemia 2 (4.2) 13 (4.0) 0.97 (0.21 – 4.44) 1 .00 1.00Cancer 10 (20.8) 44 (13.7) 0.60 (0.28 – 1.30) 1 1.18 .28Cognitive Impairment 19 (39.6) 54 (16.8) 3.22 (1.69 – 6.25) 1 12.24 .00COAD 7 (14.6) 77 (24.0) 1.85 (0.80 – 4.29) 1 1.60 .21Dementia 8 (16.7) 20 (6.2) 3.03 (1.25 – 7.14) 1 5.08 .02Depression 10 (20.8) 53 (16.5) 0.75 (0.35 – 1.60) 1 0.29 .59Diabetes 10 (20.8) 72 (22.4) 1.10 (0.52 – 2.31) 1 .004 .95Functional Impairment 26 (54.2) 66 (20.6) 4.55 (2.44 – 8.33) 1 23.43 .00Fall on Admission 17 (35.4) 53 (16.5) 2.77 (1.43 – 5.26) 1 8.52 .004Fracture on Admission 10 (20.8) 31 (9.7) 2.44 (1.12 – 5.26) 1 4.21 .04Gender (Male) 23 (47.9) 149 (46.4) 0.94 (0.51 – 1.73) 1 0.002 .97Hearing Impairment 10 (20.8) 48 (15.0) 0.67 (0.31 – 1.43) 1 0.69 .41Hypertension 33 (68.8) 185 (57.6) 0.62 (0.32 – 1.18) 1 1.70 .19Hypercholesterolemia 17 (35.4) 76 (23.7) 0.57 (0.30 – 1.08) 1 2.46 .12Ischemic Heart Disease 11 (22.9) 51 (15.9) 0.64 (0.30 1.33) 1 1.02 .31Joint Replacement 3 (6.3) 27 (8.4) 1.38 (0.40 – 4.73) 1 .05 .82Osteoporosis 10 (20.8) 87 (27.1) 1.41 (0.68 – 2.96) 1 .55 .46Parkinson’s Disease 1 (2.1) 9 (2.8) 1.36 (0.17 – 10.95) 1 .00 1.00Recent Admission to Hospital 19 (39.6) 137 (42.7) 1.14 (0.61 – 2.11) 1 0.06 .80Renal Failure 5 (10.4) 34 (10.6) 1.02 (0.38 – 2.75) 1 0.00 1.00Stroke or TIA 9 (18.8) 38 (11.8) 0.58 (0.26 – 1.30) 1 1.23 .27Visual impairment 8 (16.7) 39 (12.1) 0.69 (0.30 – 1.59) 1 0.41 .52
Note. *Condition present (used as reference group for odds ratio), Significant p value. TIA transient ischaemic attack
105
4.3.4.1.2 Age
An independent samples t test was conducted to compare the ages of patients in
the case and control groups. There was a significant difference in mean ages of the
case (M = 84.11, SD = 7.3) and control group (M = 77.69, SD 11.8), t (457.35) = 7.30,
p < .00), suggesting that cases were significantly older than controls. The mean age
difference between the groups was 6.4 years (95% CI: 4.69 to 8.14).
Independent samples t test was also conducted to compare the ages of patients in
the possible delirium and control groups. There was a significant difference in mean
ages of the possible delirium (M = 83.98, SD = 7.8) and control group (M = 77.69, SD
11.8), t (85.71) = 4.93, p < .00), also suggesting that patients with possible delirium
are significantly older than controls. The mean age difference between the groups
was 6.29 years (95% CI: 8.83 to 3.75).
4.3.4.1.3 Logistic regression for predisposing factors
Bivariate analysis identified a number of factors that were significantly related to
incident delirium. All of these factors were placed into the logistic regression model
for predisposing factors. Table 22 presents the results of the initial logistic
regression using the bivariate analysis of the cases and controls.
Variables included in the initial logistic regression analysis were dementia, cognitive
impairment, vision impairment, functional impairment, age, previous delirium, fall,
fracture, stroke, osteoporosis, hypercholesterolemia, and Parkinson’s disease. As
described in the Methods chapter steps were undertaken to reach a final logistic
regression model. Table 23 presents the final logistic regression model.
106
Table 22. Initial logistic regression results for predisposing factors of incidentdelirium using cases and controls
Predisposing Risk factor* B S.E. Wald df Sig.
Odds
Ratio
95% CI for Odds
Ratio
Lower Upper
Age .040 .013 9.08 1 .003 1.04 1.01 1.07
Cognitive Impairment .941 .254 13.76 1 .000 2.56 1.56 4.21
Dementia 1.016 .372 7.45 1 .006 2.76 1.33 5.73
Functional Impairment .603 .259 5.42 1 .020 1.83 1.10 3.04
Fall .365 .301 1.47 1 .225 1.44 .80 2.60
Fracture .676 .372 3.30 1 .069 1.97 .95 4.08
Hypercholesterolemia .533 .250 4.55 1 .033 1.70 1.04 2.783
Osteoporosis .023 .240 .009 1 .924 1.02 .64 1.64
Previous Delirium 2.701 .786 11.80 1 .001 14.9 3.19 69.55
Parkinson’s .480 .550 .76 1 .383 1.62 .55 4.75
Stroke or TIA .548 .288 3.62 1 .057 1.73 .98 3.04
Vision Impairment .354 .293 1.46 1 .227 1.43 .80 2.53
Constant 5.152 1.07822.828 1 .000 .006
Note. *Condition present (used as reference group for odds ratio), Significant p value. TIA transient
ischaemic attack
107
Table 23. Final logistic regression model for predisposing factors of incident deliriumusing cases and controls
Predisposing Risk Factor* B S.E. Wald df Sig.
Odds
Ratio
95% CI for Odds
Ratio
Lower Upper
Age .04 .01 10.18 1 .001 1.04 1.02 1.07
Cognitive Impairment .97 .25 14.81 1 .000 2.63 1.61 4.32
Dementia .99 .37 7.26 1 .007 2.70 1.31 5.54
Fracture .90 .31 8.31 1 .004 2.46 1.33 4.53
Functional Impairment .70 .25 7.57 1 .006 2.01 1.22 3.29
Hypercholesterolemia .54 .25 4.67 1 .031 1.71 1.05 2.77
Previous Delirium 2.80 .78 12.96 1 .000 16.48 3.57 75.07
Stroke or TIA .52 .29 3.31 1 .069 1.69 .96 2.97
Constant 5.150 1.061 23.58 1 .000 .006
Note. *Condition present (used as reference group for odds ratio), Significant p value. TIA transient
ischaemic attack
Logistic regression was performed to determine the relationship between a number
of predisposing factors and the development of incident delirium. The final model
contained eight independent variables (dementia, cognitive impairment, functional
impairment, age, previous delirium, fracture on admission, history of stroke or
transient ischemia attack, and hypercholesterolemia). As shown in Table 23, seven
of the independent variables made a unique statistically significant contribution to
the model. The strongest predictor of incident delirium was having a previous
delirium, with an odds ratio of 16.48 (p < .0001). This indicated that the odds of
developing delirium were 16 times higher in patients that had a previous episode of
delirium compared to those that had not previously had a delirium, controlling for
all other factors in the model. Patients with a dementia (OR 2.7, p = .007) or a
cognitive impairment (OR 2.6, p < .001) had just over two times greater odds of
developing a delirium than those with unimpaired cognition. Patients that had a
functional impairment (OR 2, p = .006) or were admitted with a fracture (OR 2.5, p =
.004) also had greater odds of developing delirium. Age was also found to be an
independent risk factor delirium in this logistic regression model (OR 1.04, p = .001).
108
Further testing using a logistic regression model substituting ‘age’ (average age)
with ‘age greater than 80’ (categorical variable) was also performed to determine if
advanced age (greater than 80 years) was independently associated with the
development of delirium. These variables are highly correlated and could not be
assessed in the same logistic regression model. Using the same variables as in the
previous logistic regression for predisposing factors, only substituting the variable
‘age’ for ‘age greater than 80’, found that age greater than 80 was not
independently associated with incident delirium in this population. Table 24
presents the results of the initial logistic regression including ‘age greater than 80’
as a variable. Age greater than 80 does not appear to have an independent
association with delirium when compared with other variables. Table 25 presents
the final logistic regression model.
Table 24. Logistic regression for predisposing factors (including age >80) using casesand controls
Risk Factor* B S.E. Wald df Sig.
Odds
Ratio
95% CI for
Odds Ratio
Lower Upper
Age Greater than 80 .335 .249 1.806 1 .179 .72 .44 1.17
Cognitive Impairment 1.029 .254 16.389 1 .000 2.80 1.70 4.61
Dementia 1.088 .371 8.575 1 .003 2.97 1.43 6.15
Fall .396 .301 1.735 1 .188 1.49 .82 2.68
Functional Impairment .712 .255 7.789 1 .005 2.04 1.24 3.36
Fracture .698 .372 3.519 1 .061 2.01 .97 4.17
Hypercholesterolemia .526 .248 4.483 1 .034 1.69 1.04 2.75
Osteoporosis .121 .237 .263 1 .608 1.13 .71 1.80
Parkinson’s .359 .544 .436 1 .509 1.43 .49 4.15
Previous Delirium 2.748 .788 12.150 1 .000 15.62 3.33 73.25
Stroke Or TIA .621 .286 4.708 1 .030 1.86 1.06 3.26
Vision Impairment .404 .295 1.881 1 .170 1.50 .84 2.68
Constant 1.863 .245 57.678 1 .000 .155
Note: *condition present (reference group for odds ratio), Significant p value. TIA transient ischaemic attack,
109
Table 25. Final logistic regression model for predisposing factors (not including age)
for cases and controls
Risk Factor* B S.E. Wald df Sig.
Odds
Ratio
95% CI for
Odds Ratio
Lower Upper
Cognitive Impairment 1.139 .247 21.234 1 .000 3.12 1.92 5.07
Dementia 1.138 .366 9.672 1 .002 3.12 1.52 6.40
Fracture 1.009 .308 10.720 1 .001 2.74 1.50 5.02
Functional Impairment .863 .247 12.247 1 .000 2.37 1.46 3.84
Hypercholesterolemia .538 .245 4.795 1 .029 1.71 1.06 2.77
Previous Delirium 2.928 .777 14.185 1 .000 18.69 4.07 85.79
Stroke or TIA .640 .284 5.082 1 .024 1.90 1.09 3.31
Constant 1.899 .191 99.144 1 .000 .150
Note: *condition present (reference group for odds ratio), Significant p value. TIA Transient ischemic attack
The results of the logistic regression presented in Table 25 show seven factors
having an independent association with incident delirium. As previously identified in
the last logistic regression model shown in Table 23, dementia, cognitive
impairment, functional impairment, previous delirium, and fracture on admission
were also associated with the development of incident delirium. Age greater than
80 years was not identified as being independently associated with delirium when
accounting for all other factors in the model.
4.3.4.2 Precipitating risk factors
A number of precipitating risk factors for the development of incident delirium were
also examined in the process of auditing medical records. A chi square test for
independence (with Yates Continuity Correction) was used to examine the
relationship between variables. Variables were tested using cross tabulation to
determine if the proportion of patients with the risk factor (e.g. use of indwelling
catheter) was significantly different between the case and control groups.
Chi square analysis of variables (Table 26) found a significant relationship between
use of an indwelling catheter (OR 2.38, p = .000), administration of a
benzodiazepine during admission (OR 1.55, p = .034), use of physical restraints (OR
110
3.25, p < .00 *Fisher exact test), adding more than three medications (OR 3.94, p <
.0001), and an abnormal sodium level during admission (OR 1.76, p = .006) with
incident delirium, suggesting that patients exposed to these factors during
hospitalisation had greater odds of developing incident delirium. A significant
relationship between a patient being moved between wards and delirium was not
detected (OR 1.44, p = .142).
Table 26. Chi square statistic results of precipitating risk factors for delirium
comparing cases and controls
PrecipitatingFactor*
CaseDeliriumN = 161n (%)
Controls
N = 321n (%)
Odds Ratio (95 % CI) dfPearson’s Chisquarevalue
p
AbnormalSodium
72 (44.7) 101 (31.5) 1.76 (1.19 2.60) 1 7.62 .006
AdministeredBenzodiazepines
74 (46.0) 114 (35.5) 1.55 (1.05 2.27) 1 4.49 .034
Add more than 3Medications
85 (52.8) 71 (22.1) 3.94 (2.62 5.91) 1 44.70 .000
Moved Wards 39 (24.2) 58 (18.1) 1.44 (0.92 2.29) 1 2.58 .142Use of IDC 57 (35.4) 60 (18.7) 2.38 (1.55 3.66) 1 15.40 .000Use of Restraints 18 (11.2) 0 3.25 (2.83 3.72) 1 34.24 .000
Note: *condition present (reference group for odds ratio), Significant p value. IDC – indwelling catheter
4.3.4.2.1 Precipitating risk factors for possible delirium group compared with
control group
Analysis was also performed on the possible delirium group compared to the
control group to determine if patients with possible delirium possessed similar
precipitating risk factors to the cases. A chi square test for independence (with
Yates Continuity Correction) was used to examine the relationship between
precipitating factors in the possible delirium group and the control group (Table 27).
111
Table 27. Chi square statistic results of precipitating risk factors for delirium
comparing possible delirium with control group
Precipitating Factor*
PossibleDeliriumN = 48n (%)
Control
N = 321n (%)
Odds Ratio (95 %CI)
dfPearson’s
chisquarevalue
p
Abnormal Sodium 24 (50) 101 (31.5) 2.18 (1.18 4.02) 1 5.60 .02
Add more than 3Medications
24 (50) 71 (22.1) 3.52 (1.89 6.57) 1 15.55 <.00
AdministeredBenzodiazepines
21 (43.8) 114 (35.5) 1.41 (0.76 2.61) 1 0.89 .35
Moved Wards 14 (29.2) 58 (18.1) 1.87 (0.94 3.70) 1 2.61 .11
Use of IDC 17 (35.4) 60 (18.7) 2.39 (1.24 4.59) 1 6.10 .01
Use of Restraints 1 (2.1) 0 7.83 (5.99 10.23) 1 1.21 .27Note: *condition present (reference group for odds ratio), Significant p value. IDC – indwelling catheter
As with the comparison of precipitating factors between the delirium cases and
controls, comparisons between possible delirium patients and controls produced
similar results. Use of an indwelling catheter (OR 2.39, p = .01), adding more than
three medications (OR 3.52, p < .00), and having an abnormal sodium level during
admission (OR 2.18, p = .018) were all shown to have a relationship with the
possible development of delirium. That is, patients that had possible delirium were
likely to have been exposed to similar precipitating factors during admission as the
cases. The three remaining factors were found to be not significant: administered
benzodiazepines (OR 1.41, p = .35), use of restraints (OR 7.83, p = .27) and moving
wards (OR 1.87, p = .11). In the ‘possible delirium’ group only one patient was
physically restrained, therefore the sample size was not large enough to undertake
bivariate analysis.
4.3.4.2.2 Blood test results
Blood test results for patients in the delirium and control group were recorded
during audit for the day of admission and again for the third day after admission. An
independent samples t test was conducted to compare the difference between the
case and control groups (Table 28). The mean sodium level on admission was not
significantly different between the delirium (M = 137.46mmol/L, SD = 5.3) and
control group (M = 137.26mmol/L, SD = 4.5; t (480) = .43, p = .670). Sodium level on
the third day after admission was also not significantly different.
112
Urea levels on admission were not different between the delirium (M =
10.7mmol/L, SD = 7.6) and control groups (M = 9.99mmol/L, SD = 9.7; t (480) = .81,
p = .418). However, differences in average urea levels on the third day of admission
did reach significance (t (471) = 1.96, p = .05).
Table 28. Blood tests results comparisons between delirium and control groups
Blood Test Results
CaseDeliriumN = 161Mean (SD)
Control
N = 321Mean (SD)
df t Mean Difference(95% CI)
p
Sodium Level OA 137.46 (5.3) 137.26 (4.5) 480 0.43 0.20 ( 0.71 1.11) .67
Sodium Level Day 3 137.23 (12.1) 135.40 (19.3) 472 1.09 1.83 ( 1.47 – 5.12) .28
Urea Level OA 10.70 (7.6) 9.99 (9.7) 480 0.81 0.71 ( 1.01 – 2.42) .42
Urea Level Day 3 10.17 (6.3) 8.99 (6.1) 471 1.96 1.18 (0.01 – 2.36) .05
Note. OA = on admission.
4.3.4.2.3 Logistic regression for precipitating factors
Selection of predictors for the logistic regression modelling was based upon the
bivariate analyses performed with precipitating factor variables. Table 29 presents
the results of the logistic regression including all precipitating factors identified as
significantly related to incident delirium. Use of physical restraints was not included
as there was no incidence of restraint use in the control group. This is not estimable
using multivariate analysis.
113
Table 29. Logistic regression results for precipitating factors of incident delirium in
cases and controls
Precipitating Factor* B S.E. Wald df Sig.
Odds
Ratio
95% C.I. for
Odds Ratio
Lower Upper
Abnormal Sodium .398 .214 3.451 1 .063 1.49 0.99 2.27
Administered Benzodiazepine .245 .214 1.319 1 .251 1.28 0.84 1.94
Add more than 3 Medications 1.226 .214 32.825 1 .000 3.40 2.24 5.18
IDC .629 .238 7.016 1 .008 1.88 1.18 2.99
Urea Day 3 .020 .017 1.379 1 .240 0.98 0.95 1.01
Constant .770 .333 5.353 1 .021 2.160
Note: *condition present (reference group for odds ratio), Significant p value. IDC Indwelling catheter
Variables included in the analysis were use of indwelling catheter, adding more than
three medications, abnormal sodium, administered benzodiazepine, and urea level
on day three. After the initial logistic regression analysis was undertaken, the results
showed precipitating factors that remained significantly related to incident delirium
when adjusting for possible interactions between all of the factors. Table 30 shows
the final logistic regression model for precipitating factors for incident delirium.
Table 30. Final logistic regression model for precipitating factors of incident delirium
Risk Factor* B S.E. Wald df Sig.
Odds
Ratio
95% CI for Odds
Ratio
Lower Upper
IDC .694 .232 8.958 1 .003 2.00 1.27 3.15
Add more than 3 Medications 1.274 .212 36.279 1 .000 3.58 2.36 5.41
Abnormal Sodium .430 .212 4.118 1 .042 1.54 1.02 2.33
Constant 1.499 .164 83.876 1 .000 .223
Note: *condition present (reference group for odds ratio), Significant p value. IDC – Indwelling catheter
The final model contained three independent variables (use of indwelling catheter,
adding more than three medications, and an abnormal sodium during admission).
114
All three independent variables made a unique statistically significant contribution
to the model (Table 30). The strongest risk factor for incident delirium was being
administered more than three new medications during admission, with an odds
ratio of 3.57 (p < .001). This indicates that patients administered more than three
new medications during admission had 3.5 times greater odds of developing a
delirium than patients who were not administered more than three medications,
controlling for all other factors in the model. Patients who had an indwelling
catheter inserted during admission (OR 2, p = .003) and patients that had an
abnormal sodium level at any stage during admission (OR 1.5, p = .04) also had
greater odds of developing delirium.
4.3.5 Outcomes for patients
The outcomes for patients in the case, control, and possible delirium groups were
documented and analysed using descriptive statistics, bivariate analysis using the
chi square test, and logistic regression.
4.3.5.1 Residence on admission and discharge destination
Residence on admission and discharge destinations of all patients were recorded
during the medical records audit (Table 31). Patients in the case group were more
likely to live alone when compared to the other groups. However, when comparing
discharge destination, there was a difference across groups in terms of the number
of patients that were subsequently discharged back to being home alone after
hospitalisation. Of the 30% (n = 49/161) of cases that came from living at home
alone, only 1.9% (n = 3) were discharged back to home alone. There was a marginal
increase in the cases that were discharged home alone with services being
implemented from 5.6% (n = 9) prior to admission to 6.8% (n = 11) following
discharge. However, the majority of these patients were more likely to have been
discharged to rehabilitation, with around 42% (n = 68) of cases and 35.4% (n = 17) of
possible delirium patients being discharged to a rehabilitation facility compared to
24% (n = 79) of control group patients.
115
Patients in the case, control and possible delirium groups that came from high level
care establishments were likely to return there following discharge, with the same
numbers of patients in each group returning to high level care. Patients in both the
case and control groups were less likely to return home with family. Around 44% (n
= 72) of cases came from home with family, with only 18% (n = 29) of patients going
home with family on discharge. This was similar for patients with possible delirium,
with 22.9% (n = 11) of patients returning home with family from the original 54.2%
(n = 26). The number of patients that had services at home increased following
discharge from hospital for both case and control groups (Table 31).
116
Table 31. Residence on admission compared to discharged destination of patients in case and control groups
Note. TCP = Transitional Care Program, LLC = Low Level Care, HLC = High Level Care
Location
Residence on Admission Discharge Destination
CaseDeliriumN = 161n (%)
ControlN = 321n (%)
PossibleDeliriumN = 48n (%)
CaseDeliriumN = 161n (%)
ControlN = 321n (%)
PossibleDeliriumN = 48n (%)
Another Acute Facility 0 0 0 4 (2.5) 15 (4.7) 1 (2.1)
Died 0 0 0 9 (5.6) 10 (3.1) 6 (12.5)
Home Alone 49 (30.4) 86 (26.8) 9 (18.8) 3 (1.9) 34 (10.6) 1 (2.1)
Home with Services 9 (5.6) 14 (4.4) 1 (2.1) 11 (6.8) 28 (8.7) 1 (2.1)
Home with Family 72 (44.7) 178 (55.5) 26 (54.2) 29 (18.0) 112 (34.9) 11 (22.9)
HLC 13 (8.1) 16 (5.0) 4 (8.3) 13 (8.1) 16 (5.0) 4 (8.3)
LLC 18 (11.2) 27 (8.4) 8 (16.7) 10 (6.2) 19 (5.9) 4 (8.3)
Palliative Care 0 0 0 5 (3.1) 1 (0.3) 3 (6.3)
Rehabilitation 0 0 0 68 (42.2) 79 (24.6) 17 (35.4)
TCP 0 0 0 9 (5.6) 7 (2.2) 0
117
4.3.5.2 Comparison of outcomes for patients
A chi square test for independence (with Yates Continuity Correction) was used to
examine the relationship between outcomes for patients. These outcomes include
death during or after admission, the discharge destination of the patient and
possible iatrogenic events during hospitalisation, such as a having a fall (Table 32).
Table 32. Comparison of outcomes for cases and control patients
Outcome*
CasesDeliriumN = 161n (%)
Control
N = 321n (%)
Odds Ratio (95 % CI) dfPearson’schi square
valuep
Change in continence 67 (41.6) 8 (2.5) 27.88 (12.93 60.14) 1 121.95 .00
Code grey 22 (13.7) 1 (0.3) 50.6 (6.76 379.47) 1 39.19 .00
Decreased functioning 72 (44.7) 28 (8.7) 8.47 (5.15 13.91) 1 82.33 .00
Developed pressure injury 5 (3.1) 1 (0.3) 10.26 (1.19 88.54) 1 4.73 .02
Died during admission 9 (5.6) 10 (3.1) 1.84 (0.73 4.63) 1 1.14 .29
Died after admission 30 (18.6) 36 (11.2) 1.81 (1.07 3.07) 1 4.39 .04
Discharged to care facility 37 (23.0) 43 (13.4) 1.93 (1.18 3.14) 1 6.44 .01
Discharged to
rehabilitation
68 (42.2) 79 (24.6) 2.24 (1.50 3.36) 1 14.89 .00
Fall 27 (16.8) 11 (3.4) 5.68 (2.74 11.78) 1 24.48 .00
MET call 16 (9.9) 29 (9.0) 1.11 (0.59 2.11) 1 0.02 .88
Note. *Condition present (used as reference group for odds ratio), Significant p value. MET – Medical
Emergency Team
Because there was a low sample frequency for the case and control groups for some
outcomes (developed pressure injury, code grey during admission, and change in
continence), Fisher exact test results for the p – value have been reported.
In addition to the comparison of outcomes for patients in the delirium and control
groups, possible delirium and control group patient outcomes were assessed. The
results of this analysis are displayed in Table 33.
118
Table 33. Comparison of outcomes for patients with possible delirium and control
group
Outcome*
PossibleDeliriumN = 48n (%)
Control
N = 321n (%)
Odds Ratio (95 % CI) dfPearson’s
chisquarevalue
p
Change in continence 16 (33.3) 8 (2.5) 19.56 (7.77 – 49.25) 1 60.34 <.00
Died during admission 6 (12.5) 10 (3.1) 4.44 (1.54 – 12.85) 1 6.75 .01
Died after admission 10 (20.8) 36 (11.2) 2.08 (0.96 – 4.54) 1 2.71 .10
Decreased functioning 14 (29.2) 28 (8.7) 4.31 (2.07 – 8.97) 1 15.34 <.00
Discharged to care
facility
11 (22.9) 43 (13.4) 1.92 (0.91 – 4.05) 1 2.32 .13
Discharged to
rehabilitation
17 (35.4) 79 (24.6) 1.68 (0.88 – 3.20) 1 2.00 .16
Fall 8 (16.7) 11 (3.4) 8.54 (2.14 14.93) 1 12.40 <.00
MET call 7 (14.6) 29 (9.0) 1.72 (0.71 – 4.17) 1 0.90 .34
Note. *Condition present (used as reference group for odds ratio), Significant p value. MET – Medical
Emergency Team
4.3.5.2.1 Deaths
A significant difference in the frequency of deaths during admission between the
cases and the control group was not detected (p = .285); however, a significant
difference in the frequency of deaths following the patient’s discharge was detected
(p = .036). In contrast, when exploring the possible delirium cases and controls, a
significant difference between the groups for those that died during admission was
found (p = .009) and no significant difference between groups was found for those
that died after discharge (p = .10).
4.3.5.2.2 Change in functioning and continence
During the patient’s admission, if there was a change in the patient’s ability to carry
out any of their activities of daily living or a decrease in their mobility it was
considered to be a change in their functional status. Cases were more likely to have
had documented evidence of some decrease or alteration in their ability to carry
out activities of daily living compared to patients in the control group (OR 8.47, p =
.000). Patients with possible delirium were also more likely to have a change in
functional status compared to control group patients (OR 4.31, p < .000).
119
Changes in a patient’s continence during admission were also investigated. If the
patient had one or more episodes of incontinence during admission, and they were
previously described as being continent, they were considered to have had a
change in their continence. Cases were more likely to have had a change in their
continence with 41.6% of the patients having at least one episode of incontinence,
compared to only 2.5% of patients in the control group (OR 27.78, p = .000). Again,
patients with possible delirium were also more likely to have an episode of
incontinence during admission compared with the control group (OR 19.56, p <
.000). However, due to the smaller sample size the confidence interval for both
comparisons are quite wide and therefore it is difficult to predict the true effect
size.
4.3.5.2.3 Discharge destination
The discharge destination of patients in the three groups was also examined in
bivariate analysis. Cases had greater odds of being discharged to a care facility (OR
1.93, p = .008) and being discharged to a rehabilitation facility (OR 2.24, p = .000).
This finding was not the same for patients with possible delirium compared to the
control group. Being admitted to a care or rehabilitation facility was not statistically
significant for the possible delirium cases when compared with the control group. A
comparison of discharge destinations and the places of residence on admission
were presented earlier (see Table 31).
4.3.5.2.4 Falls
Patients in the delirium group had greater odds of having a fall during admission
compared to the control group (OR 5.68, p < .000). Patients in the possible delirium
group also had greater odds of having a fall during admission compared to the
control group (OR 8.54, p < .000). Figure 12 illustrates the number of falls
experienced by patients in each group during admission. Falls occurred at various
times of the day as illustrated in Figure 13.
120
Figure 12. Number of falls patients had during admission
Figure 13. Time of day falls occurred during admission
19
6
1 1
9
1
8
0
2
4
6
8
10
12
14
16
18
20
1 fall 2 falls 3 falls 4 falls
Num
bero
ffalls
Number of falls per patient
Delirium Control Possible Delirium
8
13
17
1
4
7
4
1
3
0
2
4
6
8
10
12
14
16
18
Morning (06001200)
Afternoon (12002000)
Night (2000 0600)
Num
bero
ffalls
Time of day falls occuredDelirium Control Possible Delirium
121
4.3.5.2.5 Pressure injuries
As outlined earlier in Table 18, patients that developed delirium were more likely to
be scored as high risk for pressure injuries using the Braden scale pressure risk
assessment. Patients in the control group were more likely to be scored as low risk
for pressure injury with 69.2% scored as being at low risk, compared to 55.9% of
cases. Chi square analysis revealed that cases had greater odds of developing a
pressure injury compared to the control group (OR 10.3, p = .017). However, there
was an overall low incidence of pressure injuries for all patients and the confidence
interval is quite wide, therefore it is difficult to predict the true effect size. A larger
sample size is required to detect significance with a narrow confidence interval. No
patients in the possible delirium group developed a pressure injury during
admission.
4.3.5.2.6 Medical emergency team and code grey calls
There was no significant difference between the case and control group for the
number of patients that had a Medical Emergency Team (MET) call during
admission (OR 1.11, p = .876). This was also the case for the possible delirium group
and the control comparison (OR 1.72, p = .34). Additionally, a difference between
groups in the number of patients that prompted a code grey, called for aggressive
or threatening behaviour, was found. Patients with delirium had greater odds of
having a code grey called during admission compared to the control group (OR 50, p
< .000). However, due to the smaller sample size and only one patient in the control
group having a code grey call, the confidence interval is very wide and therefore it is
difficult to predict the true effect size. No patients with possible delirium had a code
grey called during admission.
4.3.5.2.7 Length of stay
An independent samples t test was conducted to compare the length of stay in days
for case and control groups. There was a significant difference in length of stay
(days) between the case (M = 12.42, SD = 7.1) and control groups ((M = 9.25, SD =
6.1), t (283.15) = 4.85, p < .00). The mean difference between the patients in the
groups was 3.16 days (95% CI: 3.73 to .70), indicating that on average the patients
122
that developed delirium were in hospital for 3 days longer than those who did not
develop delirium.
An independent samples t test was also conducted to compare the length of stay in
days for possible delirium and control groups. There was no significant difference in
length of stay (days) between the possible delirium (M = 11.09, SD = 6.1) and
control group ((M = 9.25, SD = 6.1), t (366) = 1.90, p = .06). The mean difference
between the patients in these groups was 1.8 days (95% CI: 3.73 to .70).
4.3.5.2.8 Logistic regression of outcomes for patients
Logistic regression was performed to determine the relationship between the
development of incident delirium and the outcomes experienced by patients.
Selection of outcomes for the logistic regression modelling was based upon the
bivariate analyses performed with outcome variables. Table 34 presents the results
of the initial logistic regression modelling of all variables for outcomes assessed.
Table 34. Initial logistic regression model for patient outcomes using cases and
controls
Outcome* B S.E. Wald df Sig.
Odds
Ratio
95% C.I. for
Odds Ratio
Lower Upper
Change in Continence 2.979 .421 50.046 1 .000 .051 .022 .116
Discharged to Rehabilitation .035 .317 .012 1 .913 .966 .519 1.798
Discharged to Care Facility .919 .320 8.237 1 .004 .399 .213 .747
Died after Admission .134 .350 .146 1 .702 .875 .440 1.738
Decline in Function 1.568 .348 20.336 1 .000 .208 .105 .412
Fall 1.200 .452 7.066 1 .008 3.321 1.371 8.046
Length of Stay .015 .022 .498 1 .480 .985 .944 1.027
Constant 3.999 .755 28.032 1 .000 54.565
Note. *Condition present (used as reference group for odds ratio), Significant p value.
Variables included in the analysis were fall during admission, decline in function,
episode of new incontinence, length of stay, discharged to rehabilitation,
123
discharged to care facility, and died after admission. The final logistic regression
model for patient outcomes is presented in Table 35.
Table 35. Final logistic regression model for patient outcomes using cases and
controls
Outcome * B S.E. Wald df Sig.
Odds
Ratio
95% C.I. for Odds
Ratio
Lower Upper
Decline in Function 1.528 .299 26.154 1 .000 4.61 2.56 8.26
Discharged to Care Facility .892 .297 9.028 1 .003 2.44 1.36 4.37
Fall 1.175 .446 6.944 1 .008 3.24 1.35 7.77
Incontinence 2.931 .411 50.808 1 .000 18.87 8.40 41.67
Constant 3.613 .529 46.577 1 .000 37.086
Note. *Condition present (used as reference group for odds ratio), Significant p value.
The final model contained four independent variables (fall during admission, change
in function, change in continence and discharge to a care facility). All independent
variables made a unique statistically significant contribution to the model (Table
35). Patients with delirium had greater odds of having a change in continence and
experiencing an episode of incontinence during their hospitalisation (OR 18.9, p <
.000), having a fall during admission (OR 3.2, p = .008), having a decline in their
functional status (OR 4.6, p = .000), and being discharged to a care facility (OR 2.4, p
= .003).
124
4.3.6 Delirium detection and management
The following section presents an analysis of data collected from the medical record
regarding the screening, identification, diagnosis and management of delirium,
including pharmacological and non pharmacological management during
hospitalisation.
4.3.6.1 Monitoring of cognition
Two cases (1.2%) and one control patient (0.3%) had a cognitive assessment on
admission using a formal cognitive assessment tool. No patients with possible
delirium had a cognitive assessment on admission. Despite no evidence of formal
cognitive assessment, data were also extracted if there was any documented
evidence that a family member had been asked about that patient’s previous state
of cognition. In the case group, 94 patients (58%) and in the control group, 46
patients (14.3%) had documented evidence that a family member was asked about
the patient’s previous level of cognition. In the possible delirium group, 13 patients
(27.1%) had documented evidence that a family member was asked about the
patient’s previous level of cognition.
For cases, after they had an initial acute change in cognition, 22 (13.7%) had their
level of cognition assessed using cognitive assessment tools (either the Mini Mental
State Exam (MMSE) or the Rowland Universal Dementia Assessment Scale (RUDAS).
After documentation of this initial acute change in cognition, on average 4.45 days
(Range 1 to 21 days; SD 4.9) elapsed before the cognitive assessment was
performed again.
4.3.6.2 Delirium risk assessment
Medical records were assessed for documentation of delirium risk factors. This may
have included use of a specific delirium risk factor assessment tool to identify the
patient’s level of risk on admission to hospital or stating the patient was at high risk
of delirium due to the presence of delirium risk factors. None of the patient records
125
for the case, control group, or possible delirium groups had any evidence that a risk
factor assessment was carried out.
4.3.6.3 Time of delirium development and documentation
Date of patient admission and the date on which the patient first displayed signs of
delirium were also extracted. A patient was considered to have developed signs of
delirium if they were described as being either: vague, confused, hallucinating,
agitated, aggressive, drowsier than usual, or having poor attention and had not
displayed these symptoms previously. For patients that developed delirium, it took
an average of 2.92 days (Range 1 to 15 days) from day of admission to the
documentation of delirium symptoms. For the possible delirium group it took an
average of 3.69 days (Range 2 to 10 days) from day of admission to the
documentation of possible delirium symptoms.
The dates on which delirium was diagnosed were also recorded. For the 48 patients
with possible delirium, a diagnosis was not documented. For cases, from the initial
development of symptoms, it took an average of 2.66 days (Range 0 to 17 days) for
a diagnosis of delirium to be documented in the medical history.
4.3.6.4 Words used to describe delirium
For cases, a variety of words were used to describe the signs/behaviours that the
patients first started displaying. Table 36 provides a list and the frequency of words
documented by health professionals to first describe the signs of delirium. The list
illustrates the signs of delirium that were first recognised by health professionals.
Patients that developed delirium were most likely to be described as confused, with
63.4% (n = 102) of patients identified as confused. A further 13.7% (n = 22) of
patients were first identified as vague and 9.3% (n = 15) of patients were described
as agitated. Patients in the possible delirium group were also likely to have
confusion (n = 35, 72.9%) or vague (n = 4, 8.4%) documented, as the first signs of
delirium.
126
Table 36. Description words used for the first symptoms of delirium and possible
delirium
Word Used to Describe Signs
Delirium
N = 161
n (%)
Possible Delirium
N = 48
n (%)
Agitated 15 (9.3) 3 (6.3)
Aggressive 5 (3.1) 0
Confused 102 (63.4) 35 (72.9)
Disorientated 6 (3.7) 3 (6.3)
Drowsy and confused 4 (2.5) 1 (2.1)
Hallucinating 5 (3.1) 1 (2.1)
Impulsive 1 (0.6) 0
Poor attention 1 (0.6) 0
Vague 22 (13.7) 4 (8.4)
Words used to describe the episode of delirium during admission were also
extracted during the medical record audit. Words may have been used more than
once in the same record and multiple words may have been used to describe the
patient. Patients with delirium and those with possible delirium were likely to be
described as confused during the admission. Ninety eight per cent of patients who
developed delirium and 91% of patients with possible delirium were described as
confused at some point during admission. Patients in both groups were also likely to
be described as vague and/or disorientated during the admission. Figure 14
illustrates the frequency of use of terms to describe patients in the delirium and
possible delirium groups.
127
Figure 14. Percentage of description words used during admission for patients with delirium and possible delirium
69.6
98.1
61.4
31.1
63.4
46.9
21.7
3.1
73
91.7
35.4
10.4
68.8
41.7
29.2
10.4
0
10
20
30
40
50
60
70
80
90
100
Vague Confused Agitated Hallucinating Disorientated Drowsy Short term memoryloss
Poor attention
Percen
tage
ofpatie
ntsw
ords
used
for
Words used to describe delirium symptom
Delirium Possible Delirium
128
4.3.6.5 Recognition of delirium signs
The discipline of health professionals who first documented signs of delirium in the
medical history is summarised in Figures 15 and 16. This represents the first time a
change in the patient’s cognition and signs of delirium were documented. Nurses
were most likely first to document changes in the patient’s cognition. For patients
with delirium evident in the documentation it was nurses who were the first to
document signs of delirium (80.1%, n = 130). For the possible delirium group, nurses
were also likely to be first to document the signs of delirium (85.4%, n = 41).
Figure 15. Health professional who first documented signs of delirium
Figure 16. Health professional who first documented signs of possible delirium
80.1%
13%
1.93.7 0.6 0.6
Nurse
Medical Doctor
Physio
Other Allied Health
Surgical Doctor
Geriatrician
Delirium group
85.4%
4.2 % 10.4 %
Nurse
Medical Doctor
Other Allied Health
Possible delirium group
129
4.3.6.6 Diagnosis of delirium
Medical records were examined for evidence that a diagnostic tool was used to
diagnose delirium. This may have included any of the tools recommended in the
Clinical Practice Guidelines for the Management of Delirium in Older People (Clinical
Epidemiology and Health Service Evaluation Unit and Delirium Clinical Guidelines
Expert Working Group 2006), including the CAM. There was no evidence to suggest
that a tool had been used to assist with the diagnosis of delirium in any of the
records for patients diagnosed. It was also unclear as to what signs had led the
medical team to diagnose delirium or if they had undertaken any assessments to
assist them to make the diagnosis in the medical history.
Documentation of possible causes of delirium, such as urinary tract infection or
electrolyte imbalance, was also extracted. For cases, 43.5% (n = 70) had a possible
cause of delirium documented. The remaining cases (n = 91, 56.5%) had no
documented evidence of the possible cause of the delirium.
4.3.6.7 Medication management
Data regarding prescription and use of medications were also extracted as part of
the medical records audit. For patients that developed delirium during admission,
65.2% (n = 105) were prescribed and administered medications for the
management of their symptoms. The most frequently prescribed medication was
haloperidol (n = 83, 51.6%). Patients were also prescribed olanzapine (n = 25,
15.5%), quetiapine (n = 7, 4.3%), diazepam (n = 4, 2.5%), and rispiridone (n = 12,
7.5%) to manage delirium signs or behaviours. However, indication for use of the
medication (for example agitation) was only documented 42.8% (n = 45) of the
time.
For patients that had possible delirium during admission, the most frequently
prescribed medication was haloperidol (n = 7, 14.6%) and one patient was
prescribed olanzapine (2.15%). The indication for use of the medication or
130
guidelines for when to administer the medication were documented for only one
patient (2.1%).
Of the patients in the case group that received antipsychotic medications, 91.5% (n
= 86) were receiving it for the first time. That is, the patient was not taking an
antipsychotic medication prior to admission. Only 8.5% (n = 8) of the patients that
received an antipsychotic medication had been previously receiving them. A further
eight patients were prescribed an antipsychotic medication, but it was not
administered, and 34.8% (n = 56) of patients in the delirium group were not
prescribed any antipsychotic medication.
The number of antipsychotic medications prescribed per patient was also
investigated. 78.4% (n = 80) of the cases were prescribed one antipsychotic
medication, 18.6% (n = 19) of the cases prescribed medications were prescribed two
antipsychotics, and 2.9% (n = 3) of the cases prescribed an antipsychotic were
prescribed three. Figure 17 illustrates the percentage of patients prescribed one or
more antipsychotics.
Figure 17. Number of antipsychotics prescribed for patients with delirium
4.3.6.7.1 Benzodiazepine medications
Patients that developed delirium during admission and patients in the control group
were prescribed benzodiazepine medications during admission. A number of those
patients were prescribed benzodiazepines prior to admission. Table 37 presents the
benzodiazepine medications patients were taking prior admission. Temazepam was
78.4
18.6
2.9
0 20 40 60 80 100
1 Antipsychotic
2 Antipsychotics
3 Antipsychotics
Percentage of patients prescribed antipsychotic medications
131
the most frequently used benzodiazepine with 8.1% of patients in the delirium
group, 11.2% of patients in the control group, and 14.6% of patients with possible
delirium taking the drug prior to admission.
Table 37. Benzodiazepine medications patients were taking prior to admission
Medication
CaseDeliriumN = 161n (%)
Control
N = 321n (%)
PossibleDeliriumN = 48n (%)
Alprazolam 5 (3.1) 5 (1.5) 0
Clonazepam 1 (0.6) 3 (0.9) 0
Diazepam 5 (3.1) 14 (4.4) 2 (4.2)
Lorazepam 0 4 (1.2) 0
Nitrazepam 2 (1.2) 4 (1.2) 0
Oxazepam 6 (3.7) 18 (5.6) 2 (4.2)
Temazepam 13 (8.1) 36 (11.2) 7 (14.6)
No benzodiazepine 131 (81.4) 245 (76.3) 37 (77.1)
Prescription and use of benzodiazepine medications were also recorded for each
patient during admission. Forty six per cent (n = 74) of patients that developed
delirium were administered a benzodiazepine during admission, compared to only
35.5% (n = 114) of patients in the control group. A number of patients that
developed delirium were newly prescribed benzodiazepines during admission. Of
the patients that received a benzodiazepine during admission, 55 in the delirium
group (34.2%), 49 (15.3%) in the control group, and 11 (22.9%) with possible
delirium, had newly prescribed benzodiazepines. Temazepam was the most
common newly prescribed benzodiazepine for both delirium (18%) and control
groups (9.3%). Patients with possible delirium were more likely to be prescribed
diazepam (12.5%) (Table 38).
132
Table 38. Newly prescribed benzodiazepines administered to patients during
admission
Medication DeliriumN = 161
ControlN = 321
Possible DeliriumN = 48
No. of times prescribed (% of patients prescribed for)Alprazolam 2 (1.2) 3 (0.9) 0
Clonazepam 2 (1.2) 0 0
Diazepam 11 (6.8) 11 (3.4) 6 (12.5)
Lorazepam 3 (1.9) 4 (1.2) 0
Midazolam 10 (6.2) 1 (0.3) 3 (6.25)
Nitrazepam 1 (0.6) 0 0
Oxazepam 7 (4.3) 3 (0.9) 0
Temazepam 29 (18.0) 30 (9.3) 3 (6.25)
4.3.6.8 Non pharmacological management strategies
Non pharmacological management strategies used for patients that developed
delirium were also extracted during the medical records audit. Medical histories of
patients that developed delirium during admission were scrutinised for evidence of
any management strategies outlined in the Clinical Practice Guidelines for the
Management of Delirium in Older People (Clinical Epidemiology and Health Service
Evaluation Unit and Delirium Clinical Guidelines Expert Working Group 2006). In
88.2% of cases, no documentation of any non pharmacological management
strategies for delirium existed in the medical record (Table 39). The most frequently
used strategy was encouragement of family members to stay with the patient
(8.7%). The use of a 1:1 support person was implemented, but it was rarely
documented that they were trained in delirium management specifically (n = 2,
1.2%). Sixteen cases (9.9%) had a 1:1 nurse at some stage of their admission. This
ranged from one shift (8 hours) to 4 days.
133
Table 39. Non pharmacological management strategies documented for the
management of patients with delirium
Intervention Frequency
used
Percentage
No strategies used 142 88.2
Allowing family members to stay with patient 14 8.7
Modification of environment to minimise risk of injury 2 1.2
Use of support person or 1:1 nurse who has been
trained in delirium
2 1.2
Providing relaxation strategies to assist with sleep 1 0.6
4.3.6.8.1 Physical restraints
During the episode of delirium, one strategy implemented by staff was physical
restraints. Medical doctors documented orders for use of physical restraints on a
separate physical restraint order and nursing staff undertook hourly observations.
For 140 cases (87%) no physical restraints were implemented. The remaining
patients (n = 21, 13%) were restrained with a variety of different restraint
combinations including: ankle, wrist and seat belt restraints. The time spent in
restraints varied between the patients. Duration ranged from one hour on one
occasion to one hour on multiple occasions, and a continuous seven hour period on
one occasion. The restraint used for the longest period was the seat belt restraint,
used on one patient for most of the day.
4.3.6.9 Delirium prevention strategies
Data for use of prevention strategies recommended in the Clinical Practice
Guidelines for the Management of Delirium in Older People (Clinical Epidemiology
and Health Service Evaluation Unit and Delirium Clinical Guidelines Expert Working
Group 2006) were also extracted during the medical record audit. The prevention
strategies were divided into environmental prevention strategies (including any
intervention that involved manipulation of the physical environment) and clinical
prevention strategies. Data were also collected for control patients and patients
134
with possible delirium in order to determine how often these prevention strategies
were documented.
For all groups, few environmental prevention strategies were documented in the
medical records. No documented evidence of any environmental prevention
strategies was found for 80.7% (n = 130) of patients that developed delirium, 97.2%
(n = 312) of control group patients and 87.5% (n = 42) of possible delirium patients.
Encouraging family involvement was documented for 11.8% (n = 19) of delirium
patients. Table 40 presents a summary of the environmental prevention strategies
documented in the medical record.
Table 40. Environmental prevention strategies documented for patients in case,
control and possible delirium group
Intervention* Delirium
N = 161
Control
N = 321
PossibleDelirium
N = 48Frequency used (% of delirium patients used for)
No environmental strategies 130 (80.7) 312 (97.2) 42 (87.5)
Encourage family involvement 19 (11.8) 3 (0.9) 1 (2.1)
Provision of a single room 13 (8.1) 6 (1.9) 4 (8.3)
Avoid room changes 1 (0.6) 0 0
Quiet environment 1 (0.6) 0 1 (2.1)
Provision of a clock and/or calendar 1 (0.6) 0 0
*Could have more than 1 intervention used per patient
All patients had multiple clinical prevention strategies documented. These directly
related to care and are therefore more likely to be documented by the nurses
involved in their care. Table 41 presents the clinical prevention strategies
documented in the medical records. Encouragement of food and fluids was
documented for 64% (n = 103) of cases, 45.8% (n = 22) of patients with possible
delirium and 36.4% (n = 117) of controls. Clinical intervention (prevention)
strategies were not documented for 37.1% (n = 119) of the control group, 23% (n =
37) of cases and 29.2% (n = 14) of the patients with possible delirium.
135
Table 41. Clinical prevention strategies documented for patients in case, control and
possible delirium group
Intervention*DeliriumN = 161
ControlN = 321
PossibleDeliriumN = 48
Frequency used (% of delirium patients intervention used for)
Encouragement of food and fluid intake 103 (64.0) 117 (36.4) 22 (45.8)
Encourage regular mobilisation 40 (24.8) 62 (19.3) 5 (10.4)
No evidence of clinical prevention
strategies
37 (23.0) 119 (37.1) 14 (29.2)
Pain management 28 (17.4) 51 (15.9) 6 (12.5)
Regulation of bowel function 20 (12.4) 14 (4.4) 2 (4.1)
Ensuring patient wears hearing aids 10 (6.2) 12 (3.7) 2 (4.1)
Encourage and establish a sleep routine 2 (1.2) 2 (0.6) 0
Encourage independence 1 (0.6) 3 (0.9) 1 (2.1)
Medication review 1 (0.6) 0 1 (2.1)
Ensuring patient wears glasses 0 0 1 (2.1)
*Could have more than 1 intervention used per patient
4.3.6.10 Follow up care after discharge
Of the 161 patients diagnosed with delirium, only six patients (3.7%) had
documented evidence of a referral or recommendations for follow up care for the
treatment of delirium with a specialist delirium clinic. The remaining 155 (96.3%)
patient records contained no documentation to indicate that any information
regarding specific follow up for their delirium had been given. Data in relation to
families receiving information regarding delirium were also extracted. This included
health professionals’ documentation about communicating with family members to
advise that the patient had delirium and to address any concerns the family had.
For six cases, the records contained documentation to indicate that delirium was
discussed with family members.
As none of the patients who had possible delirium during admission were diagnosed
with delirium, they did not have any follow up post discharge from hospital.
136
4.3.7 Case control study results summary
The results of the case control study build on findings of the systematic review by
providing further evidence for risk factors common in medical patients for incident
delirium. The results of the logistic regression analyses highlight that dementia,
cognitive impairment, functional impairment, fracture on admission and age had a
strong relationship with incident delirium. Interventions or physiological changes
while admitted to hospital such as insertion of an IDC, adding more than three
medications and abnormal sodium level were also strongly related to incident
delirium. A description of the characteristics of the patients, including their place of
residence on admission, their previous level of functioning, age and gender, is
reported. The outcomes for patients with delirium were more likely to be worse,
with these patients being more likely to be discharged to a care facility, have a
decline in level of functioning and incontinence, as well as increased incidence of
code grey calls, and falls. Results have also been presented for the medications that
patients received during hospitalisation. Patients with incident delirium were likely
to be given an antipsychotic for treatment of their delirium. The management and
prevention strategies documented in the patient’s medical record have also been
presented as well as the discipline of health professionals most likely to document
signs of delirium and time elapsed before diagnosis was made.
137
4.4 Phase 3 Delirium management survey results
In the following section, the delirium management survey recruitment and
participation rates will be presented as well as the detailed survey results.
4.4.1 Hospital characteristics and participation rates
Both public and private hospitals were involved in the survey. In terms of private
hospitals, twenty six were invited to participate in the survey. Respondents from
five of the 26 hospitals (19.2%) consented to participate by completing the survey.
The nominated respondent from 15 of the 26 private hospitals (57.7%) did not
respond to the invitation to participate, despite reminder emails. Respondents from
two of the 26 hospitals (7.7%) responded to the initial email expressing interest but
despite reminders to complete the survey did not participate. A further four
respondents of the 26 hospitals (15.4%) responded to initial contact but declined to
participate; two declined stating this was because the hospital did not have a
specific policy, and one declined stating they did not have any patients that
developed delirium.
The public hospitals in Melbourne are grouped into regions and are managed as a
single network within each region. Several networks exist in the Melbourne
metropolitan area. This is relevant because one of the survey questions asked if
policies were used across all campuses of the health network. All respondents from
public health networks stated that this was the case and therefore it was only
necessary to contact one respondent from each of the public hospital networks.
Nine of the hospital networks within the Melbourne region were contacted and
invited to participate in the research. Respondents from seven of the nine hospital
networks (77.8%) consented to participate by completing the survey. Respondents
from two of the nine (22.2%) networks did not respond to the invitation to
participate. The overall response rate for the survey from both public hospital
networks and private hospitals was 34% (N = 12).
138
Respondents included senior staff members in various roles, including: Directors of
Nursing (DON), Quality Systems Coordinators, Nurse Unit Managers, Staff
Development Coordinators, Quality Services Managers, Cognition Clinical Nurse
Consultant, Health Educators and Associate Managers.
Publically available data about the approximate number of overnight admissions
per hospital network were collected through themyhospitals.gov.au website. Data
regarding the number of overnight admissions were not publically available for the
private hospitals via this website. Respondents were requested to report an
estimate or known figure for the number of patients per year that developed
delirium in their organisation. Only five of the public hospital respondents were able
to provide this information. Private hospital respondents did not provide these
data. Table 42 presents the approximate number of overnight only admissions per
public hospital and the estimated percentage of delirium incidence reported by the
respondents.
Table 42. Hospital network data for the approximate number of patient admissions
per year and estimates of delirium incidence
Public hospital
network
Approx. number of overnight
admissions for year 2011 –
2012*
% of patients that develop
delirium (estimate)
Network 1 54,188 2%
Network 2 66,765 Unsure
Network 3 47,895 Unsure
Network 4 30,832 2%
Network 5 43,567 9.85%
Network 6 30,753 40%
Network 7 33,661 50 – 60% in general medical
Note. *As reported on the myhospitals.gov.au website
139
4.4.2 Delirium management policies
A summary of the responses to each survey question is presented in Figure 18 (see
Appendix 9 for the survey tool). The graph provides a summary of the proportion of
public and private hospital respondents that stated, ‘yes’ for questions regarding a
delirium policy and the content of the policy. A detailed overview of the results for
each of the survey questions will be presented in the following sections.
Figure 18. Delirium management policies and guidelines in both public networks (n =
7) and private hospitals (n = 5). (*Clinical Practice Guidelines for the Management of
Delirium in Older People)
60 60 60
40 40
60
20
57.1
42.9
100
57.1
14.3
57.1
71.4
Percen
tage
YesR
espo
nse
Private Public
140
4.4.2.1 Delirium management policy
Survey respondents were asked if the hospital had developed a specific delirium
management policy. Respondents from three (60%) of the private hospitals and
four (57.1%) of the public networks stated they had a policy relating specifically to
the management of patients with delirium (see Figure 18). Respondents from some
of the hospitals (n = 2, 1 public and 1 private) that did not have a delirium
management policy expressed their desire to have one, or stated that while they
did not currently have a policy, one was in the process of being developed (n = 1
public).
4.4.2.2 Awareness of the Clinical Practice Guidelines for the Management of
Delirium in Older People (Clinical Epidemiology and Health Service Evaluation Unit
and Delirium Clinical Guidelines Expert Working Group 2006).
As shown in Figure 18 (‘Aware of Guidelines’) all respondents from public hospital
networks reported being familiar with the Clinical Practice Guidelines for the
Management of Delirium in Older People. Three out of the five (60%) respondents
from private hospitals were aware of the guidelines.
4.4.2.3 Delirium policy developed using the Clinical Practice Guidelines for the
Management of Delirium in Older People as a guide
Respondents from two (40%) of the private hospitals stated the delirium
management policy was developed using the guidelines. All respondents from the
four (57.1%) public networks that stated they had a policy for delirium management
indicated the policy had been developed using the guidelines (Figure 18 ‘Policy
Developed Using Guidelines’).
4.4.2.4 Screening and diagnosing delirium
As shown in Figure 18 (‘Screening Policy’), respondents from three of the private
hospitals (60%) and three of the public networks (42.9%) stated (as part of the
delirium management policy or a separate policy) there were guidelines available
for screening and for diagnosing of delirium. The tool that was most often
recommended was the Confusion Assessment Method (CAM). One public network
141
adapted the CAM into a flow chart. A respondent from one public network reported
they had an initial delirium assessment tool, which included screening questions
and a confusion assessment. At one public network, staff had developed a screening
tool (based on the Clinical Practice Guidelines for the Management of Delirium in
Older People) for use within their own hospital. Other respondents stated that a
tool was not currently used, but a policy was being developed.
4.4.2.5 Documentation of the diagnosis
Respondents reported that the diagnosis of delirium was mostly documented in the
patient medical record. Respondents from one private hospital and one public
hospital network reported they used the CAM and documented the diagnosis on a
specific CAM assessment form. In three of the public networks (42.9%) respondents
reported there was nowhere specific to document a diagnosis of delirium. One
respondent noted that at their organisation the delirium diagnosis was also
documented as an alert on the patient’s electronic medical record.
4.4.2.6 Cognitive assessment on admission
Respondents from one public network (14.3%) and two private hospitals (40%)
stated there was a specific policy, or a section of the existing delirium management
policy, that recommended a formal cognitive assessment for adults over 65 years
on admission (refer to Figure 18 for ‘Cognitive Assessment Policy’). The remaining
hospitals did not have a policy for cognitive assessment on admission to hospital.
Despite not having a policy for cognitive assessment on admission a number of the
respondents from the public networks and private hospitals stated a tool was used
to undertake cognitive assessments. It was not clear when these tools should be
used and respondents reported it being at the discretion of the health professional.
The tools most often used were Abbreviated Mental Test (AMT) and the Mini
Mental State Exam (MMSE). The AMT was used by two (40%) of the private
hospitals, and the MMSE was used by one private hospital (20%). The remaining
two private hospitals did not use a tool for cognitive assessment (40%). Two of the
public networks used a series of cognitive questions on the admission form to
assess cognition (28.6%), two used both the AMT and the MMSE (28.6%), and one
142
used just the MMSE as a cognitive assessment tool (14.3%). Two of the public
networks did not use a cognitive assessment tool (28.6%). Only two of the public
networks (28.6%) and one of the private hospitals (20%) reported that they
provided specific training for the use of these tools.
4.4.2.7 Risk factor assessment
Respondents from four of the public networks (57.1%) and three of the private
hospitals (60%) stated there was a section of the policy that recommended delirium
risk assessment (refer Figure 18 ‘Risk Factor Assessment Policy’). One of the
respondents from a public network stated that a positive result on the risk factor
assessment would lead to a further comprehensive assessment. Respondents were
also asked about the training that was provided for staff regarding the risk factors
for delirium development. Much of the training provided by the public networks
included in service training sessions, delirium study days and online education
packages. Respondents from two of the public networks stated that no in service
delirium education was provided. It was reported that three private hospitals (60%)
did not provide any delirium risk factor training. Respondents from the remaining
two hospitals (40%) stated that ward based training was provided and information
sheets were given to staff.
4.4.2.8 Pharmacological management policy
There was a pharmacological policy used in five of the public hospital networks and
one of the private hospitals. The respondent for one of the public networks that did
not have a policy indicated that for guidance on medication for patients with
delirium, they often referred to the Clinical Practice Guidelines for the Management
of Delirium in Older People (Clinical Epidemiology and Health Service Evaluation Unit
and Delirium Clinical Guidelines Expert Working Group 2006). Another respondent
in a private hospital stated that due to the nature of the private system they could
not provide a specific hospital policy for medication management. The reason given
was that doctors prescribing the medications are not specifically employed by the
hospital and the hospital cannot implement guidelines on specific use of
medication. Respondents reported that a range of medications was recommended
143
for the management of severely agitated patients. Four hospital respondents stated
that haloperidol was recommended in their policy. Doses of haloperidol differed,
with three policies recommending doses between 0.25mg to 0.5mg, and one
hospital policy recommending 0.5mg to 1mg of haloperidol. Table 43 summarises
the medications and doses that were recommended in the hospital policies.
Table 43. Medications and doses recommended for patients with agitation and
aggression in a pharmacological management policy
Medication
Number ofHospitals thatrecommendedmedication Dose recommended (max 24 hours)
Haloperidol 4 3 hospitals recommended 0.25 mg – 0.5 mg
1 hospital recommended 0.5 mg – 1 mg
Rispiridone 5 3 hospitals recommended 0.25 mg BD (max 2 mg)
1 hospital recommended 0.25 mg – 0.5mg (max 2 mg)
1 hospital recommended 0.5 mg – 1 mg
Midazolam 1 Patients with no dementia 5 – 10 mg (elderly)
Olanzapine 5 3 hospitals recommended 2.5mg (max dose 10 mg)
1 hospital recommended 2.5 mg – 5 mg (max dose 5
mg)
1 hospital recommended 5 – 10 mg
Quetiapine 3 1 hospital recommended 12.5 mg (max 37.5 mg)
2 hospitals recommended 12.5 mg – 25 mg (max 50
mg)
4.4.2.9 Medical review of patients with delirium
Respondents from four of the private hospitals (80%) stated there were no
recommended guidelines for the frequency of a medical review. The remaining
respondent stated such patients needed to be monitored hourly; however, in the
private hospital a doctor is not always available to review the patient. Respondents
from two of the public networks (28.6%) stated that recommendations for medical
review did not exist. Three of the public hospital respondents (42.9%) stated that
there were no guideline recommendations for medical review but the patient
should be closely monitored. One public hospital network respondent stated that
144
post the administration of a medication such as Haloperidol, medical staff should
review the patient within one hour. Another public network respondent stated that
according to their policy a patient would be reviewed daily and on request.
When asked what training is provided to medical staff regarding medication
management and review needs of a patient with delirium, most respondents of
both public networks and private hospitals stated that they did not know. Four of
the public hospital respondents (57.1%) stated that new medical staff are educated
during orientation. One public hospital service provided access to delirium study
days for medical staff.
4.4.2.10 Barriers to implementation or development of policies
Respondents were asked if they had experienced any barriers or difficulties
implementing a delirium management policy. There was a range of issues identified
as barriers. Lack of knowledge of delirium among clinical and managerial staff was a
common concern (n = 3, 2 public and 1 private). This lack of knowledge included
continued use of contraindicated medications such as benzodiazepines, lack of
understanding of how to recognise delirium, or knowledge regarding delirium
causes.
One private hospital respondent stated that due to some staff’s personal
experiences and their differences in opinions regarding the best delirium
management there was a lack of consensus and an inability to develop guidelines
for their hospital. Two private hospital responses included resistance from clinical
staff members to the implementation of new documentation required. One public
network respondent also stated that there was disagreement regarding
management of delirium between treating teams. Another public network
respondent stated that they were in the process of developing a delirium
management policy but were experiencing difficulty with the actual content of the
policy. There was inconsistency regarding what assessment tools should be used,
who should conduct assessments, timing of assessments, and clarification of
medication recommendations. Finally, one public network respondent expressed
145
concern regarding complacency about delirium. They stated that a barrier to
developing and/or implementing a delirium management policy was that delirium
was not a focus of the organisation. The respondent stated that the organisation
was more clearly focused on passing accreditation and meeting the National Safety
and Quality Health Service (NSQHS) Standards (which do not currently include
delirium management).
4.4.3 Survey results summary
The results of this survey indicate that a number of private hospitals and public
hospital networks do not currently have a delirium management policy. There is a
lack of guidance in most hospitals on how to screen for and diagnose delirium.
Cognitive assessments are seldom recommended in the delirium management
policies of the organisations. A number of barriers were identified by respondents
to have impacted on the implementation of a policy on delirium management in
their organisation.
4.5 Conclusion
This chapter has presented the results of all three phases of the study. The
following chapter will provide a discussion of these results, followed by a concluding
chapter that will provide the concluding statements of the research.
146
Chapter 5 Discussion
5.1 Introduction
This chapter provides a discussion of the results of all three phases of the research
as well as how these results relate to existing literature. Firstly, risk factors for
delirium identified in the systematic review and the case control study are
discussed and findings compared to previous research. Health professionals’
assessment of risk factors and management of patients, according to the results of
the case control study, will be discussed in relation to existing research and policies.
The strengths and limitations of the research will also be discussed in this chapter.
5.2 Risk factors
Risk factors identified in the systematic review and the case control medical record
audit reinforce existing evidence that a variety of factors can contribute to incident
delirium. Predisposing risk factors identified in the research as having the strongest
association with incident delirium include dementia, cognitive impairment, previous
delirium, functional impairment, a fall prior to admission, and a fracture prior to
admission. Factors for which some association with incident delirium was evident,
but which need more investigation to confirm a relationship, include: age, visual
and hearing impairment, depression, severe illness, and years of education.
Importantly, this study also identified factors not associated with incident delirium
that had previously been shown to have an association with development of
delirium. These include male gender and pneumonia. One of the greatest
challenges when considering delirium risk factors is that, for two patients with
similar risk factors, one may develop delirium while the other may not (Inouye
2006). Thus it is vital that factors that contribute most significantly to the likelihood
of developing incident delirium are highlighted. In the case control study, logistic
regression was therefore used to identify factors with independent associations
with delirium, compared to other factors. The following sections will discuss the
results in relation to individual risk factors identified in both the systematic review
and the case control study and existing research literature.
147
5.2.1 Predisposing risk factors
5.2.1.1 Dementia
In both the systematic review and the case control study, dementia was found to be
a significant risk factor for incident delirium in the medical patient population.
Patients that were identified in the case control study as having a possible delirium
were also likely to have dementia. These findings support those of an earlier review
conducted by Elie et al. (1998). Although, Elie and colleagues included studies from
all areas of the hospital (including medical, surgical and psychiatric services), their
review also showed that evidence of dementia was associated with incident
delirium occurring in patients in all settings. These findings support the view that
dementia is one of the leading risk factors for delirium across a range of hospital
settings (Inouye 2006). The results of the systematic review and the case control
study with respect to dementia are not surprising given that an extensive amount of
research has identified dementia as a risk factor for delirium (Ajilore & Kumar 2004;
Boettger, Passik & Breitbart 2009; Dasgupta & Hillier 2010; Fick, Agostini & Inouye
2002; Inouye 1999; Inouye et al. 1993; Margiotta et al. 2006; Rabins & Folstein
1982).
5.2.1.2 Cognitive impairment
In the systematic review and the case control study cognitive impairment was
found to be strongly associated with a patient’s risk of developing incident delirium.
Most studies in the systematic review used the Mini Mental State Exam (MMSE) to
measure cognitive impairment. A result on the MMSE of less than 24 indicates a
degree of cognitive impairment and is significantly related to incident delirium.
Although retrospective analysis of patients’ cognition using the MMSE could not be
undertaken for patients in the case control study, descriptors of the patients’
cognition were used. Patients in the case control study were considered to have
cognitive impairment if they were described as having significant memory
problems, or were described by a family member as having some cognitive issues.
Although the MMSE was not used to assess level of cognition of patients in the
case control study, a patient whose family stated the patient had some cognitive
148
impairment (which was not previously diagnosed) had greater odds of developing
incident delirium than a patient who had not been described as having some
cognitive impairment. This highlights the need for health professionals to take into
account and consider the input of family members in regards to reporting the
patient’s cognition status. This can also be a cue for health professionals to conduct
a formal cognitive assessment on the patient. The relationship between cognitive
impairment and delirium has frequently been examined in the literature and
evidence shows an increased risk for delirium if cognitive impairment is present
(Harwood, Hope & Jacoby 1997; Jackson et al. 2004; Korevaar, van Munster & de
Rooij 2005; MacLullich et al. 2009). As a result, assessment of cognition on
admission is important in order to determine a patient’s possible risk for delirium.
Voyer et al. (2007) identified that when determining risk for delirium, severity of
prior cognitive impairment was less important than the patient’s cognitive status on
admission. That is, recent deterioration in level of cognition appeared to be more
important in determining delirium risk. This is important when determining the
appropriate time to undertake an assessment of the patient’s cognition. A more
detailed discussion regarding the importance of undertaking cognitive assessments
for patients on admission will be presented in section 5.3.
5.2.1.3 Advanced age
Mixed results regarding age as a risk factor for incident delirium were found in the
systematic review and case control study. In a number of previous studies,
advanced age has been identified as a risk factor for delirium (Edlund et al. 2006;
Elie et al. 1998; Inouye 2006; Inouye et al. 1993; Levkoff et al. 1992; Mattar, Chan &
Childs 2012). In the systematic review, meta analysis of results from two studies
that reported data on patients over 80 years did not produce a significant result (OR
1.42, p = .33). That is, advanced age was not associated with the likelihood of
developing incident delirium. Similarly, when the average age of patients with and
without delirium from four of the studies were compared, similar non significant
results were found (p = .57). This was an unexpected finding of the systematic
review. It could be that in the medical patient population, age is not a strong risk
factor for incident delirium. Factors that are generally associated with advanced
149
age, such as cognitive impairment, dementia, functional impairment and visual
impairment could impact on the risk of developing incident delirium as opposed to
age alone. That is, an older person with cognitive or functional impairment may be
more at risk of incident delirium than an older person with no functional or
cognitive impairment. The review findings suggest that more evidence is needed to
determine if age is an independent risk factor for delirium.
Other research supports the findings of the systematic review that risk factors such
as functional impairment could be more important than age alone in determining
delirium risk. For example, Korevaar, van Munster and de Rooij (2005) found a
strong relationship between age and impaired functional capacity. If the authors
had excluded functional impairment from their analysis, advanced age would have
been independently associated with delirium development. Margiotta et al. (2006)
found similar results and concluded that in patients who developed delirium, illness
severity, impaired cognitive function and decreased functional capacity played a
greater role in delirium development than age alone. Margiotta et al. (2006) also
found that in patients with no dementia, age was not a risk factor for delirium,
although MMSE scores (p = .012) and functional loss (p = .000) were significantly
related to delirium development. Findings of these studies also suggest that factors
that may be due to advanced age, such as cognitive and functional impairment,
have a more significant impact than age alone.
When comparing average ages between the delirium and control groups included in
the case control study, patients with delirium were significantly older, compared to
the control group. Logistic regression was performed using average age as a
variable and this was independently associated with incident delirium in this
population. Patients with a possible delirium were also significantly older than
patients in the control group. This finding supports much of the previous research
into delirium risk factors that suggests age is an independent risk factor for delirium
(Inouye, Westendorp & Saczynski 2014; Levkoff et al. 1992; Schor et al. 1992).
However, logistic regression was also performed with the variable ‘age greater than
80 years’. Multivariate analysis showed that advanced age of greater than 80 years,
when added to the model, was not independently associated with incident delirium
150
in this population when compared with other factors in the model. This finding
confirms the results of the systematic review. The varying results of some research
studies may be due to a number of reasons. Delirium can be caused by multiple
factors and has the potential to be experienced by different patient populations
that have variable natural histories (Dasgupta & Hillier 2010). This variation may
account for some of the differences in the findings regarding the relationship
between age and delirium. The presence of co morbid illness and the severity of
dementia, cognitive impairment and/or functional impairment, as well as how
delirium is measured, may be impacting upon how age influences a patient’s risk for
delirium. Importantly, when it came to advanced age, other factors such as
cognitive impairment, dementia, and functional impairment were more likely to
increase a patient’s risk of incident delirium. This reinforces the need for health
professionals to assess cognitive impairment on admission to hospital.
5.2.1.4 Functional impairment
Functional impairment, as measured by the Katz Index of Independence in Activities
of Daily Living (Katz et al. 1970), was significantly related to incident delirium in
both the systematic review and the case control study. In the systematic review,
meta analysis indicated that patients with functional impairment have a 75%
increase in likelihood of developing incident delirium. Multivariate analysis showed
that having functional impairment significantly increased the likelihood of
developing incident delirium in this population (OR 2.0, p = .006). This indicates that
patients with functional dependence have greater odds of developing incident
delirium. Functional impairment was also found to be a risk factor for patients with
possible delirium (p < .000).
These findings are consistent with findings of other research and provide further
evidence for the impact of functional impairment on delirium. Previous research
has shown that functional dependence, immobility, low levels of activity and a
history of falls are risk factors for delirium (Inouye 1999, 2006; Margiotta et al.
2006; Murray et al. 1993; Voyer et al. 2007). The systematic review of delirium risk
factors conducted by Elie et al. (1998) also found that diminished activities of daily
151
living were significantly associated with an increased risk of delirium (OR 2.5) (Elie
et al. 1998). These findings highlight the importance of providing a safe
environment in the clinical setting where patients can be encouraged to undertake
activities independently. Patients with pre existing functional impairment,
therefore, need to be well supported in the hospital setting in order to reduce the
risk of developing delirium or becoming more functionally dependent.
5.2.1.5 Gender
A significant relationship between male gender and incident delirium was not found
in either the systematic review or the case control study. In previous studies, male
gender was identified as a risk factor for the development of delirium (Elie et al.
1998; Inouye 2006; Schor et al. 1992). The review conducted by Elie and colleagues
(1998), included studies conducted in surgical, medical and psychiatric patient
settings. Two of the studies found the strongest relationship between male gender
and delirium occurred among patients in the surgical setting. Studies conducted in
medical and psychiatric settings did not show an association between male gender
and incident delirium.
Similarly, mixed results were found in studies included in the systematic review
conducted as part of this research. Six of the studies included data regarding
delirium risk according to male gender but only one found that male gender was
significantly related to development of delirium (OR 3.06) (Campbell et al. 2011).
The authors acknowledged that this was an unexpected finding and were unclear as
to why this may have occurred (Campbell et al. 2011). Considering the variability in
results, more research is needed to conclusively determine the risk of male gender
associated with development of incident delirium in the medical population.
5.2.1.6 Sensory impairment
Visual impairment was not found to have a relationship with delirium in the
systematic review. The confidence interval crosses the line of no effect and
consequently, the results must be interpreted with caution. In the case control
152
study, multivariate analysis showed that visual impairment was not independently
associated with incident delirium.
Hearing impairment was reported in only one of the studies included in the
systematic review and thus could not be used for meta analysis. In the case control
study, neither the case/control nor possible delirium/control comparisons indicated
hearing impairment was significantly related to incident delirium, suggesting that
those with hearing impairment are unlikely to have an increased risk of incident
delirium. However, in the case control study because evidence of the presence of
hearing impairment relied on documentation, it is possible that the influence of
hearing impairment is underestimated.
Both visual and hearing impairment were found to be moderate risk factors for
delirium in the systematic review conducted by Elie et al. (1998) (OR 1.9). For
studies included in the present systematic review, there were slight differences in
the way that each of the studies defined vision impairment, however all eye tests
indicated a significant problem with the patients’ eyesight. The studies included in
the present systematic review defined visual impairment as vision worse than 20/70
(Franco et al. 2010), vision requiring the use of aids and continuing to interfere with
activities of daily living (O'Keeffe & Lavan 1996), and abnormal vision using a
standard Jaeger test (McAvay et al. 2007). The risk prediction model developed by
Inouye et al. (1993) included visual impairment as one of the greatest predictors of
delirium development. Participants were considered to have vision impairment if
their corrected vision was worse than 20/70. The authors found that patients with
visual impairment had a relative risk of 3.51 for the development of delirium.
However, it was acknowledged that a low prevalence of visual impairment (only 6
patients with visual impairment) existed and this may have skewed the results. The
impact of visual impairment on development of delirium varies among studies and
needs further investigation in the medical patient population to determine the
strength of its association with delirium.
153
5.2.1.7 Level of education
Studies included in the systematic review also investigated whether level of
education influenced development of delirium. However, in terms of the case
control study this information could not be retrospectively collected and thus was
not examined. Three of the studies in the systematic review reported on the
relationship between the number of years of education and incident delirium.
Combining the results of these studies and calculating the mean years of education
for those who developed delirium and those who did not, revealed no significant
difference between the two groups (p = .64). However, one of the studies was
conducted in Colombia and the other two were conducted in the United States of
America (USA). Educational facilities and practices may vary between the two
countries. In the study conducted in Colombia, the patients who had no delirium
had significantly less years of education compared to those in the studies conducted
in the USA who did develop delirium. Taking this into consideration and comparing
only the two studies that were conducted in the USA, a significant difference
existed in the years of education attained between delirium and non delirium
groups (p = 0.04), with those with more years of education less likely to develop
incident delirium.
Previous research has suggested there may be a strong association between level of
education attainment and the development of a dementia (Mortimer & Graves
1993). A study by Jones et al. (2006) showed that the greater the level of
educational attainment, the stronger the cognitive reserve, and it has been
postulated that this may act as a protective mechanism against the development of
dementia. Considering the strong relationship between dementia and delirium,
some researchers have considered that educational attainment may also impact on
development of delirium. However, further research is needed to confirm the
relationship between delirium and years of education.
5.2.1.8 Illness severity and co morbidity
In the systematic review severity of illness and illness co morbidity were
investigated for their impact on risk for delirium. Due to difficulties with
154
retrospectively determining without documented evidence if an illness was ‘severe’,
‘illness severity’ could not be assessed in the case control study. In the systematic
review, data could not be pooled for meta analysis due to the heterogeneity in
definitions and measures of severe illness. Use of tools, such as the APACHE scoring
system (to predict outcomes in ICU), produced different results between studies,
with McAvay et al. (2007) finding a strong association of illness severity and
delirium, whilst Wilson et al. (2005) failed to find any association. The utility of
illness severity prediction tools needs further exploration because sample size may
have contributed to the lack of conclusive findings in these studies. Furthermore,
when investigating the co morbidity of illness in the studies by Campbell et al.
(2011) and McAvay et al. (2007), when illness co morbidity was included in the
analysis, there were no significant differences found in scores on the Charlson co
morbidity index between those who developed delirium and those that did not.
These findings suggest a significant relationship between the burden of co morbid
illness and incident delirium does not exist.
Elie et al. (1998) found illness severity to be the second most important risk factor
for delirium with a combined OR of 3.8. A large quantity of research has been
conducted to examine the role that severe illness plays in the development of
delirium. Inouye et al. (1993) included severe illness in the delirium risk prediction
model they developed. Furthermore, a study conducted by Voyer et al. (2007) also
found that patients who developed a moderate to severe delirium were more likely
to have a severe illness status. This suggests that the severity of the illness can
impact on severity of delirium. More research is needed to determine the impact of
severe illness on delirium risk and use of valid and reliable tools will help to improve
comparability of results among studies.
5.2.1.9 Depression
Mixed findings regarding the relationship between depression and delirium were
found in the systematic review and case control study. All studies included in the
systematic review used the Geriatric Depression Scale (GDS) to score depression
and this limited the heterogeneity between studies. However, due to differences in
155
the way results of the GDP were reported in the included studies, meta analysis was
unable to be undertaken to compare the findings. Of the four studies included in
the systematic review that measured depression, two studies failed to find a
significant relationship between depression and incident delirium (O'Keeffe & Lavan
1996; Wakefield 1996 & 2002). However, two studies did find some relationship
between delirium and depression, indicating that patients who were depressed at
the time of admission were more at risk of developing incident delirium (McAvay et
al. 2007; Wilson et al. 2005).
In the case control study, a diagnosis of depression was only recorded if the patient
had a documented past history of depression. This was not an indication of the
current depressive state of the patient, only that they had been previously
diagnosed with depression. Analysis did not reveal a relationship between a history
of depression and incident delirium. A similar finding emerged for patients with
possible delirium. Thus, findings from the case control study suggest that a history
of depression may not contribute to delirium risk. As a result, future research
should be undertaken to determine if depression at the time of admission
contributes to patients’ risk of delirium, as this was identified as a possible risk
factor in two of the studies included in the systematic review.
In the review conducted by Elie et al. (1998) depression was associated with
delirium in only two out of five studies, but was still found to be significantly
associated with delirium (OR 1.9). A more recent study found that depression was
not related to delirium (Edlund et al. 2006). Considering the similarity between the
symptoms of delirium, particularly hypoactive delirium, and depression the mixed
findings are not a surprising result. A number of researchers have noted that
delirium is often misdiagnosed as depression (Cerejeira & Taylor 2011; Eliopoulos
2010; Mauk 2010; Saxena & Lawley 2009) and so it is difficult to clearly identify or
understand the relationship between delirium and depression. It may be that
patients with depression and displaying symptoms of dysphonic mood and
hopelessness may be more likely to develop hypoactive delirium with symptoms of
lethargy, difficulty in focusing attention and decreased motivation. Further research
156
on incident delirium including the subtype presentation and depression is needed
to more fully understand this association.
5.2.1.10 Previous delirium
Having a previous episode of delirium was identified as a risk factor for incident
delirium in the case control study. Existence of previous delirium was not
investigated in the systematic review but multivariate analysis of data for patients
in the case control study showed that previous delirium was independently
associated with an increased risk of incident delirium during admission. This result is
not surprising, as the patients would likely have had the same predisposing risk
factors that previously increased their risk of delirium. However, as the case control
medical records review was conducted retrospectively, documentation of the
previous episode was necessary. It is possible that patients that developed delirium
were more likely to have a previous delirium documented because medical staff
were more inclined to investigate for evidence of previous delirium when the
patient was currently experiencing delirium. This could therefore have resulted in
more patients with delirium having an episode of previous delirium documented
and therefore influenced the results. Results of the case control study indicate that
patients who experience delirium during hospitalisation should be informed of the
diagnosis and the risk of subsequent episodes of delirium during future admissions.
Previous research has also identified history of delirium as a risk factor for current
delirium. Litaker et al. (2001) assessed 500 patients pre operatively for predisposing
risk factors and post operatively for the presence of delirium. They found that,
when compared to other factors, a history of delirium was a predictor of post
operative delirium. It is important to note that the patients included in the study
were undergoing elective surgery and were not from the medical population.
Despite this, a history of delirium should be considered for patients in the medical
population since there is evidence that previous delirium is a risk factor for incident
delirium (Inouye 2006; Inouye, Westendorp & Saczynski 2014). Edlund et al. (2006),
using bivariate analysis, found previous delirium was a risk factor for prevalent
delirium among patients admitted to a medical setting (Edlund et al. 2006).
157
However, when combined with other factors in logistic regression, it was not found
to be an independent predictor of delirium. Rather the results indicate that other
factors such as age and severity of illness were greater predictors of prevalent
delirium. To establish robust evidence for an association between history of
delirium and development of incident delirium during hospitalisation in the medical
setting, further research is required.
5.2.1.11 Fall and fracture prior to admission
In the case control study, having a fracture on admission was independently
associated with increased risk of incident delirium. The most common fractures
experienced by patients in the case control study were hip fractures. However, the
type and location of fracture was not independently investigated in this study. It is
also important to note that patients in the case control study were treated
conservatively (such as bed rest) for their hip fractures; patients who underwent
surgery to treat hip fractures were excluded because surgical intervention was an
exclusion criteria for this study. Factors such as pain and narcotic use were not
included in the logistic regression model for the case control study. Future
prospective studies will need to include these factors in relation to fractures and
risk for delirium in medical patients.
Patients in the case control study who had sustained a fracture prior to admission
were likely to have a decrease in mobility or a resulting functional impairment.
Fracture on admission was not investigated in any of the studies included in the
systematic review. Previous research supports the findings of the case control study
that having a fracture on admission increased the risk of delirium. For example,
Schor et al. (1992) also found that fracture on admission was an independent risk
factor for delirium (OR 6.57, p < .01). However, the authors do not identify the type
of fracture present on admission and there were low numbers of patients that had a
fracture. The study findings also included patients from both the medical and
surgical populations, although it was not clear how patients with fractures were
managed or if they had undergone surgical intervention. Much research has been
undertaken to assess the risk factors for delirium among hip fracture patients
following surgery, yet little research has investigated the risk of fracture among
158
patients receiving conservative management. The findings of the case control study
suggest that patients admitted with a fracture and managed conservatively should
be carefully monitored for development of delirium.
Patients with delirium in the case control study were more likely to be admitted
after experiencing a fall. In bivariate analyses, having a fall prior to admission was
associated with a risk for delirium, however when combined in multivariate logistic
regression a fall was not independently associated with incident delirium. This may
be because those who had a fall and sustained a fracture were more likely to
develop incident delirium, compared to having a fall but not sustaining a fracture.
Additionally, patients who experienced a fall but did not sustain a fracture may have
had the fall because of functional impairment. As both functional impairment and
fracture were independently associated with delirium, this may explain why a fall
prior to admission was not independently associated with delirium in the
multivariate analysis.
5.2.2 Precipitating risk factors
Possible precipitating risk factors for delirium were investigated in the systematic
review and case control study. Of the studies included in the systematic review,
only one investigated precipitating risk factors for delirium (Inouye & Charpentier
1996). Factors identified were use of physical restraints, malnutrition, more than
three medications added, use of a bladder catheter, and any iatrogenic event. These
factors were examined in the case control study. However, due to the nature of
retrospective data collection it was difficult to assess some of the risk factors. For
example, Inouye and Charpentier (1996) defined malnutrition as an albumin level
less than 30g/L. The medical team rarely measured albumin levels for patients in
the case control study and therefore this factor could not be assessed.
Other potential precipitating risk factors examined in the case control study
included use of an indwelling catheter, benzodiazepine use, use of physical
restraints, adding more than three medications, abnormal sodium, and one
environmental factor: moving wards. Multivariate logistic regression showed that
159
use of an indwelling catheter, adding more than three medications during
admission, and having an abnormal sodium level were independently associated
with delirium.
In the bivariate analyses of precipitating risk factors, use of physical restraints was
also associated with delirium (OR 3.25, p < .000). As none of the patients in the
control group were physically restrained, this variable could not be included in the
logistic regression model. There is evidence to suggest a relationship between
restraint use and delirium in this study and also previous research (Inouye &
Charpentier 1996), but it is unclear if the relationship is because of the use of
physical restraints in the management of delirium signs or if delirium develops due
to the use of physical restraints. Evidence from the case control study indicated that
patients that developed delirium were often restrained in order to control
behaviours the patients exhibited after delirium developed. None of the patients in
the control group were physically restrained; this reinforces that physical restraint
may not be a risk factor for delirium, only that they are used in response to patients
displaying behaviours common to delirium, such as aggression and agitation.
However, the use of restraints could possibly contribute to worsening of delirium.
5.2.3 Risk factor prediction assessment
One of the aims of this research was to identify which of the risk factors for delirium
are more common in the medical population and if health professionals had
identified them. The case control study included examination of patient records for
evidence that patients had been assessed for these risk factors using risk prediction
assessment tools, or documentation of a patient’s increased risk due to certain risk
factors. The Clinical Practice Guidelines for the Management of Delirium in Older
People (Clinical Epidemiology and Health Service Evaluation Unit and Delirium
Clinical Guidelines Expert Working Group 2006) recommend undertaking a risk
assessment for all older patients admitted to a health care setting. Despite these
recommendations, none of the patient records in the case or control group had
documented evidence of a risk factor assessment for delirium on admission to
hospital.
160
There is strong evidence for the need to undertake delirium risk assessments for
patients admitted to acute hospital settings (Elie et al. 1998; Inouye 1999; Inouye et
al. 2000; Litaker et al. 2001; Marcantonio et al. 1994; O'Keeffe & Lavan 1996). The
survey conducted in Phase 3 of this research also sought to determine if hospitals in
Melbourne had a policy that reflected the need to undertake a delirium risk
assessment on admission. Results indicated that three of the five private hospitals
and four of the seven public hospital networks surveyed had a policy that
recommended a delirium risk factor assessment.
Identification of patients as moderate to high risk of developing delirium is
important for their management in hospital. Predictive models, such as those
developed by Inouye and colleagues (Inouye & Charpentier 1996; Inouye et al.
1993), help identify risk factors to target using preventative interventions.
Modifying identified risk factors could prevent a possible delirium. However,
applicability of risk prediction models needs to be investigated in the Australian
medical setting.
5.3 Assessing patient cognition on admission
Previous research (Harwood, Hope & Jacoby 1997; Jackson et al. 2004; Korevaar,
van Munster & de Rooij 2005; MacLullich et al. 2009; Voyer et al. 2007), as well as
risk factors identified in the systematic review and in the case control study, have
highlighted that cognitive impairment can increase a patient’s risk for delirium.
Identification of cognitive impairment during a patient’s admission to hospital
should prompt screening for possible delirium and implementation of delirium
prevention strategies. As discussed previously, the assessment of patients’
cognition on admission is an important screening intervention that should be
undertaken for all adults over the age of 65 in order to identify cognitive
impairment (such as signs of dementia and mild cognitive impairment) in patients
not previously diagnosed with cognitive impairment (Clinical Epidemiology and
Health Service Evaluation Unit and Delirium Clinical Guidelines Expert Working
Group 2006).
161
In the case control study, two cases (N = 161) and one control (N = 321) patient had
a cognitive assessment using a formal cognitive assessment tool on admission. This
suggests a large number of patients with cognitive impairment may not have been
identified on admission and indicates a gap in health professionals’ clinical
understanding of the importance of undertaking cognitive assessments for older
patients.
According to the findings of the case control study, family members were often
relied upon to provide an account of the patient’s prior cognitive status. For 58% of
patients that developed delirium, family members were asked about the patient’s
prior cognitive status. This information, while helpful, is not sufficient to provide a
comprehensive assessment of the patient’s level of cognition. Research has
indicated that an accurate way to assess cognition is to use existing validated tools
(Burleigh et al. 2002). If the patient’s cognition is not assessed using such tools,
cognitive impairment may not be recognised. This is highlighted in a study
conducted by Harwood, Hope and Jacoby (1997) who found that up to 46% of
patients with cognitive impairment had no record of cognitive impairment in the
medical notes, suggesting it had been missed by their physicians. Burleigh et al.
(2002) also found that physicians were poor at predicting abbreviated mental test
scores, which highlights the potential for cognitive impairment to be overlooked
and the importance of physicians not relying on their own observations.
If a patient experiences a sudden change in behaviour or cognition, cognitive
assessments should be repeated because a decline in cognition can indicate
delirium (Clinical Epidemiology and Health Services Evaluation Unit and Delirium
Clinical Guidelines Expert Working Group 2006). If delirium is suspected, further
testing using a diagnostic instrument is required to make a delirium diagnosis. In
this study, only 22 patients who developed delirium were assessed using cognitive
assessment tools (such as the MMSE) when a change in their cognition had been
identified. However, this assessment was not carried out soon after the identified
change in cognition and it took an average of 4.45 days for the assessment to be
undertaken. By this time the patient had been showing signs of delirium for a
162
significant period of time without further investigation of their cognition to assist in
diagnosis. Delay in undertaking the assessments reflects a lack of understanding
regarding when and why the test should be undertaken. The diagnosis and
detection of delirium identified in the case control study will be further discussed in
subsequent sections.
According to findings of the Phase 3 survey, only three hospitals (one public and
two private) had a policy that recommended cognitive assessment for older
patients on admission to hospital. This is concerning because as the population
ages, the incidence of cognitive impairment and dementia is increasing with
numbers of people with dementia in Australia predicted to triple by 2050
(Australian Institute of Health and Welfare 2012). As a result, the ability to detect
and respond immediately to changes in a patient’s cognition is extremely
important. It is important to note however that absence of a policy does not
necessarily mean that cognitive assessments are not being undertaken in clinical
practice. Nevertheless, a systematic and widespread approach to implementing
policies and procedures for cognitive assessments for older patients is necessary.
In responses to the survey, three of the private and five of the public hospital
network respondents indicated that despite not having a policy for screening for
cognition, the Mini Mental State Exam (MMSE) and the Abbreviated Mental Test
(AMT) were sometimes used. It was not clear in their responses when these tools
were recommended for use or if they were used regularly to detect changes in
patients’ cognition. Policies guiding the use of these tools would assist to increase
the numbers of patients that have their cognitive impairment correctly identified.
This would not only improve the potential to detect delirium, but also help prevent
deterioration in a patients’ cognitive state.
5.4 Delirium recognition and diagnosis
Delirium recognition and diagnosis is important for the treatment and management
of delirium. However, in older people in the medical setting, delirium is frequently
overlooked, misdiagnosed or goes undocumented (Inouye 1994; Inouye et al. 2001).
163
In most cases, usual nursing assessments during routine bedside nurse patient
interactions often fail to detect delirium (Gillis & MacDonald 2006; Mistarz et al.
2011). It is important to accurately identify delirium in order to treat the cause and
implement other treatment interventions. As previously described, a range of
assessment and detection tools have been developed to assist clinicians to screen
the patients’ cognition as well as making the delirium diagnosis. The Clinical Practice
Guidelines for the Management of Delirium in Older People state, “a structured
process for screening and diagnosis of delirium should be established in all health
care settings” (Clinical Epidemiology and Health Service Evaluation Unit and
Delirium Clinical Guidelines Expert Working Group 2006, p. 12).
For patients in the case control study who developed delirium, there was no
documentation to suggest that a delirium identification and diagnostic tool (such as
the Confusion Assessment Method) was used to diagnose delirium. The diagnosis of
delirium was left to the discretion of the medical doctor or the physician caring for
the patient. This highlights the possibility that patients with delirium are being
missed. Additionally, this may have been the reason why a number of patients were
identified, during the audit of medical records, as having possible delirium. This is
an important finding of this research and is supported by previous research into
delirium identification, which suggests that the recognition of delirium is poor in the
acute setting (Inouye et al. 2001; McCrow, Sullivan & Beattie 2014). Authors of one
point prevalence study conducted in Finland identified 77 patients with a diagnosis
of delirium yet only 31 (40.3%) had an actual diagnosis recorded in the medical
record (Laurila et al. 2004). Recognition of delirium by nurses varies, with studies
suggesting that rates of recognition can range between 26% and 83% (Steis & Fick
2008). In the case control study, documentation of a delirium diagnosis was not
consistent or structured, and lack of documented communication between treating
teams suggesting a clear plan for treatment of the delirium did not exist.
In the case control study, in order to describe the patients’ behaviour, nursing staff
used a number of different terms, including: confused, vague, agitated and
disorientated. These terms were often left undefined and ‘confused’ was often the
only term used to describe the patient’s behaviour. It was unclear what triggered
164
the nurse to state the patient was confused and little detail was documented
regarding the patient’s behaviour. A detailed description of the actual behaviour
that prompted the nurse to describe the patient as ‘confused’ would be more
helpful, especially since the term ‘confused’ can be used to describe a variety of
behaviours. A retrospective investigation of patients in a post acute care setting
conducted by Morandi et al. (2009) found similar results. The term confusion was
most frequently noted in the records of patients with delirium (95%), with the term
delirium less frequently documented (7%) for these patients. Morandi et al. (2009)
noted that patients with a more visible symptom profile for delirium had a greater
number of keywords associaited with delirium symptoms documented.
There are a number of problems with describing a patient as confused. Nurses tend
to use the term ‘confused’ to report signs of cognitive impairment and this may not
clearly indicate a delirium (Voyer et al. 2008). The term confusion can also be used
to describe a variety of psychiatric disturbances and may not be used appropriately
to describe the behaviours of a patient with delirium (Voyer et al. 2008). When
diagnosing delirium, the description of patients as ‘confused’ has been reported to
have 23.9% sensitivity (Voyer et al. 2008) and does not provide adequate
information to make a delirium diagnosis. Milisen et al. (2002) state that nurses and
physicians often use vague and inconsistent terminology to describe a patient’s
mental state. This reinforces the conclusions of Inouye (1994), that detection of
delirium could be improved if the use of terminology was consistent and there were
clear instructions on how and where to document the signs of delirium. Vague and
inconsistent approaches to describing and documenting the behaviour of patients
with delirium may have detrimental effects for the patient in terms of appropriate
diagnosis and treatment of delirium.
An important role of the nurse is to provide an accurate account of the patient’s
clinical situation. As identified in the case control study, in medical records of
patients with delirium, nurses were the group of health professionals most likely to
document possible signs of delirium. Because nurses have frequent or continuous
contact with patients they should be educated in recognising the signs of delirium.
In this study, from the time of documentation of delirium signs (such as a change in
165
cognition), it took an average of 2.66 days (with a range from same day to 17 days
later) for a delirium diagnosis to be documented. In the medical record nurses may
have documented the patient was vague or confused for a number of days before
delirium was diagnosed. On a number of occasions there was no documented
evidence to suggest that such signs of delirium were followed up or adequately
escalated. The delirium diagnosis may have only been made because the patient’s
symptoms worsened and the medical team was called to review the patient.
From the survey, a number of the hospital networks (three private and three public
networks) had a policy that recommended use of a tool to diagnose delirium. One
respondent stated the hospital was currently developing this part of the policy.
Almost half the hospitals surveyed did not have a recommendation on how to
screen and diagnose delirium. This suggests that delirium could go undiagnosed or
be misdiagnosed in a large number of hospitals in Melbourne. Use of a systematic
approach to screening for delirium and to diagnosing the syndrome is important to
providing targeted, safe and consistent care for these patients.
5.5 Delirium medication management
The findings of the case control study contribute further evidence that prescribing
and administering antipsychotic medications are prevalent in the medical setting.
Patients that developed delirium in the case control study were often managed
pharmacologically with antipsychotic medication. Over half of cases were
administered antipsychotic medication during their admission. Of these, over 90%
had not been prescribed an antipsychotic medication previously. The medications
given to patients in the case control study included haloperidol, olanzapine and
risperidone. While antipsychotic medications are recommended in some instances
(for example if the patient with delirium is experiencing severe behavioural
disturbance) it is recommended that medication intervention occurs only after the
implementation of environmental and non pharmacological strategies (Clinical
Epidemiology and Health Service Evaluation Unit and Delirium Clinical Guidelines
Expert Working Group 2006). However, the results of the case control study
revealed that this is not usual practice for patients who develop delirium in this
166
setting. There is little evidence to support the use of any particular pharmacological
therapy for the treatment of delirium (Tropea et al. 2009) and treatment is mainly
based on experiences and observations of clinicians (Tabet & Howard 2009).
Consistent with the findings of the case control study, an audit of medical records
conducted by Tropea et al. (2009) found that patients were frequently prescribed
antipsychotics in the treatment of delirium, with 66% of patients newly prescribed
antipsychotics. The authors also reported a lack of medication review for patients
on antipsychotic medications (Tropea et al. 2009). This finding was also consistent
with the case control study findings.
In the case control study, a large number of patients in both the case and control
groups also received a benzodiazepine during admission. Benzodiazepines are
contraindicated in patients with a high risk for delirium and are not recommended
for use in older patients due to risk of side effects including worsening of the
delirium (Clegg & Young 2011). A review conducted by Elliott (2006) of the
Australian literature on problems associated with medication use in the elderly
examined the quality of prescribing for older patients in hospital. Studies were
identified that demonstrated the overuse of benzodiazepines in hospital settings. In
one study, authors reported that up to 80% of prescriptions for benzodiazepine
medications in hospital were classified as inappropriate and there was over
prescription of benzodiazepines and antipsychotics in hospital settings (Woodward,
Elliott & Oborne 2003). Although these figures for over prescription do not relate
specifically to the use of benzodiazepines in delirium management, the case control
study findings contribute further evidence that older people in hospital are
receiving benzodiazepines despite recommendations to avoid them.
In the survey, respondents were asked if a policy existed for medication
management of patients with delirium or patients that were agitated. Respondents
reported a variety of medication management recommendations. The Clinical
Practice Guidelines for the Management of Delirium in Older People state that
antipsychotic doses ‘must commence from a low dose, typically commencing with
the equivalent of 0.25mg – 0.5mg of haloperidol; olanzapine 2.5mg orally; or
risperidone 0.25mg orally’ (Clinical Epidemiology and Health Service Evaluation Unit
167
and Delirium Clinical Guidelines Expert Working Group 2006, p. 61). Three hospital
respondents reported recommended doses consistent with these guidelines.
However, other respondents reported that their hospital policy included
recommendations for doses above this level. For example, two respondents
reported doses of olanzapine over 2.5mg and one respondent reported doses of
haloperidol over 0.5mg for patients with delirium. One respondent also reported
that midazolam was recommended for elderly patients with no dementia. These
results highlight an inconsistent approach to medication management across the
hospital networks surveyed.
Respondents were also asked about the frequency of medical review for a patient
who receives an antipsychotic. For a number of the public hospital networks and
private hospitals, there were no recommendations for medical review of a patient
who had been administered an antipsychotic. This may indicate that physicians may
not be appropriately reviewing patients receiving antipsychotic medications for the
first time. Further study is required to determine the current practice of
administration of antipsychotics and the frequency of medical review following
prescription of antipsychotics.
5.6 Delirium management strategies
There is evidence to suggest that some episodes of delirium are preventable
(Inouye 2006; Martinez, Tobar & Hill 2014). Implementing effective strategies in
order to prevent delirium is therefore important. Not only is the patient at risk of
worse outcomes when they experience delirium, delirium care can also impact on
the demands for nursing resources. Patients with delirium tend to display unsafe
behaviours, require increased supervision, are more prone to falls, and require
greater support during their activities of daily living (Milisen et al. 2004).
Consequently, it is especially important to recognise and manage patients with
delirium effectively in order to reduce the likelihood of adverse events.
The Clinical Practice Guidelines for the Management of Delirium in Older People
(Clinical Epidemiology and Health Service Evaluation Unit and Delirium Clinical
168
Guidelines Expert Working Group 2006) outline a number of strategies, both clinical
and environmental, for the prevention and treatment of delirium. Prevention
strategies are based on interventions and recommendations from previous
research, specifically the Hospital Elder Life Program (HELP) (Inouye et al. 2000).
This program recommends developing a plan of care to prevent cognitive decline in
older people during hospitalisation. The program involves implementing strategies
targeting risk factors for delirium including cognitive impairment, sleep deprivation,
immobilisation, vision impairment, hearing impairment, and dehydration.
In the case control study, care was taken to review records for documented
evidence of interventions to prevent delirium occurring. Delirium prevention
strategies are classified into environmental strategies: interventions to manipulate
the clinical environment, and clinical strategies: interventions undertaken by clinical
staff caring for patients at risk of delirium. Environmental prevention strategies
documented frequently in the medical records included encouraging family
involvement and providing a single room. These strategies were often used for
patients with prior cognitive impairment or dementia. Strategies were not
consistently documented and were often only documented by one nurse during the
patient’s hospitalisation. It was also not clear that these strategies were specifically
implemented to reduce the risk of the patient developing delirium. This may
indicate an inconsistent approach to preventing delirium in the hospital or
inadequate documentation of the care implemented by nursing staff in the
prevention of delirium.
In the case control study, documentation of clinical prevention strategies was also
examined. Strategies that were frequently documented included encouragement of
food and fluids, encouraging mobilisation and pain management. These strategies
are nursing responsibilities, which include: assessing and monitoring the patient’s
nutrition, hydration, level of comfort and mobility. There was some difference
between the case and control groups in terms of the strategies that were
documented. This may be because a higher percentage of patients in the delirium
group needed assistance with activities of daily living, and were more likely to need
support with food and fluid intake as well as mobilisation.
169
Lack of documentation regarding prevention strategies could indicate either a lack
of use of prevention strategies in clinical practice or failure to document. That is, it
is not necessary to document these strategies in the medical history of the patient.
These strategies were monitored in this study in order to determine the current
documentation practices regarding implementation of delirium prevention
strategies. However, of the strategies documented in the medical history, it was not
clear if they were used to assist in prevention of delirium. The documentation, or
lack of, suggested that strategies were inconsistently implemented during the
patient’s hospital admission. Implementation of policies regarding appropriate
prevention strategies may be useful in creating a more consistent approach to
delirium prevention. One such policy may include the implementation of nursing
decision tools to assess patient risk as well as prompt prevention strategies. This
type of decision tool is common when implementing strategies for reduction in falls
risk as well as pressure injury prevention (such as the Braden Scale, Braden &
Bergstrom 1988). Further research is required to assess the actual documentation
of intervention strategies for delirium prevention and if nursing assessment
prompts would be helpful in this setting.
In the case control study, medical records were also examined closely to determine
the documented strategies for the management of signs and behaviours of patients
that developed delirium. When a patient developed delirium, the most frequently
used strategy was enabling family members to stay with the patient. This was
documented for 14 patients, though this may have occurred more frequently than
was documented. It was not clear if this was a nursing initiative or if family
members requested to stay with the patient. At times, the patient was monitored
with the assistance of a one to one nurse. This nurse was often present for one or
two shifts of the patient’s admission but some patients had a one to one nurse
patient ratio for up to four days.
Use of physical restraint is not recommended for the management of patients with
delirium (Inouye & Charpentier 1996). Such restraint can increase the risk of
delirium and increase the severity of existing delirium (Flaherty & Little 2011).
170
However, the use of physical restraints is common in the management of older
people with delirium and steps should be implemented to reduce their use (Flaherty
& Little 2011). Twenty one patients with delirium (13%) in the case control study
were physically restrained during admission. As previously discussed, physical
restraints were frequently used to control aggressive and agitated behaviour after
the patient had developed delirium. Use of physical restraints can be distressing for
patients and contribute to adverse events and poor outcomes such as fracture,
increased length of stay and worsening of the delirium (Michaud, Thomas &
McAllen 2014).
5.7 Outcomes for patients
Patients that developed delirium in this study experienced a number of poor
outcomes. This is consistent with literature, which suggests that delirium in older
people is associated with poor outcomes, independent of confounders such as age,
illness severity and dementia (Witlox et al. 2010). Furthermore, some research has
shown that the effects of delirium can continue to have adverse effects on patient
outcomes for up to one year after the initial delirium episode (McAvay et al. 2006).
Many researchers argue that delirium should be treated as a medical emergency
(Flaherty et al. 2007; Young & Inouye 2007). In the case control study, symptoms of
delirium were not considered medical emergencies as demonstrated by the lack of
Medical Emergency Team (MET) calls. A health professional can initiate a MET call if
they are worried about a patient as a result of their deteriorating condition, or if a
patient experiences an acute change in their vital signs. In the case control study,
patients that developed delirium were likely to have a ‘code grey’ called for
aggressive behaviour, resulting in security attendance and possible physical
restraint and administration of pharmacological therapy. This finding suggests that
health professionals were more likely to be concerned regarding patients’
aggressive behaviour than calling the medical team to address the patient’s acute
change in cognition and worsening condition. In the case control study, outcomes
experienced by patients who developed delirium included likely discharge to a care
facility, decline in function, incontinence, falls and increased length of stay.
171
5.7.1 Discharge to a care facility
Patients with delirium in the case control study were more likely to be discharged
to a care facility. Specifically, logistic regression modelling showed that patients
with delirium had greater odds of being discharged to low level care, high level
care, transitional care or a rehabilitation facility. This is similar to findings of other
research where delirium was found to have a strong association with nursing home
placement at discharge (McAvay et al. 2006). It is likely that patients that developed
delirium during admission were unable to return home because of a decline in
cognition or physical function. Prevention of delirium in hospital may see a decrease
in the number of patients being discharged to care facilities. Thus, not only is
prevention of delirium important for the health outcomes of the patient during
hospitalisation, it can also have implications for the patient beyond discharge.
5.7.2 Falls
Patients with delirium in the case control study had greater odds of experiencing a
fall than patients with no delirium. Additionally, some patients in the case control
study that developed delirium had more than one fall during hospitalisation.
Specifically, six patients in the delirium group had two falls, one patient had three
falls and another had four falls. This finding highlights that patients with delirium in
this acute setting may not have been supported with appropriate interventions to
prevent them from experiencing a fall. However, patients that have delirium are
often agitated and confused and are therefore difficult to engage in order to give
them an understanding of the interventions used to prevent them from falling.
Consequently, delirium prevention strategies need to be widely implemented in
order to reduce the incidence of falls in hospital. Multi component non
pharmacological interventions for delirium prevention have been found to be highly
effective in decreasing the incidence of falls related to delirium. A recent systematic
review and meta analysis found that these interventions helped to reduce the
incidence of delirium and had a greater than 60% odds reduction in the incidence of
falls, which represents the equivalent of 4.26 falls prevented per 1000 patient days
(Hshieh et al. 2015).
172
Preventing patients from falling in hospital is important because falls lead to serious
complications for patients; this is the reason for falls initiative in Australia and
elsewhere. Patients who fall can have greater rates of mortality, and around 4 8 %
of people that fall will sustain injuries, including hip, knee and back fractures that
may require major operative procedures (Bates et al. 1995; Doherty et al. 2014;
Nadkarni et al. 2005). Not only do falls have physiological implications for patients,
including declining physical function, loss of independence and restriction in
activities of daily living, they can also impact on their psychological wellbeing by
increasing their fear of experiencing another fall (Decrane, Culp & Wakefield 2012;
Lakatos et al. 2009). For a patient recovering from an episode of delirium, these
outcomes as a result of falling could potentially increase the risk for worsening
delirium. These complications can impact severely on the health of patients and
strategies to prevent delirium could help to improve the rates of falls in hospital.
Previous research (including research conducted in medical and surgical settings)
has found that delirium is one of the most important risk factors for falling
(Decrane, Culp & Wakefield 2012; Lakatos et al. 2009). Presence of confusion is also
an indicator of patients at greatest risk of a fall (Vassallo et al. 2004). One recent
study on the outcomes of patients who develop delirium following cardiac surgery
found that post operative delirium was associated with an increased risk for in
hospital falls (p < .001, OR 3.95) (Mangusan et al. 2015). Results of the case control
study support the findings of previous research in both medical and surgical
settings, which indicates that having a fall is independently associated with the
presence of delirium (Decrane, Culp & Wakefield 2012; Lakatos et al. 2009).
Consideration of the time of the day that patients experience falls is important. In
the case control study the majority of falls in both groups occurred overnight. In
Australia, nursing staff numbers in the general medical setting are significantly
reduced during nightshift and often nursing staff can be allocated up to 10 patients.
Measures are required to enable nursing staff overnight to manage patients with
delirium in order to prevent falls from occurring.
173
A recent case control study conducted by Doherty et al. (2014) examined patients
with delirium that had experienced a fall. The authors found that alterations in the
patient’s level of consciousness and inattention prior to the fall were associated
with falls. This suggests that brief assessments of the patient’s conscious state and
level of attention are needed since they may help to identify individuals at greatest
risk of falling and enable implementation of strategies to reduce falls risk, thereby
helping to decrease the potential for fall related morbidity and mortality.
5.7.3 Decline in functioning and incontinence
Previous research has indicated that development of delirium can have detrimental
implications for the functional status of the patient. A cohort study conducted by
Inouye et al. (1998) aimed to determine the independent contribution of delirium
to outcomes for patients in hospital. The authors found that delirium was a
predictor of a significant decline in ADLs even after controlling for potential
confounding factors (adjusted OR 3.0). The results of the case control study
contribute further evidence to this association as it was found that during
hospitalisation, patients who developed delirium had greater odds of a change in
their functional status. Change in functional status included increased need for
assistance with ADLs or an inability to walk without use of an aid having been
previously independent. This finding may also explain why patients with delirium
were more likely to be discharged to a care facility. Care needs for these patients
increased during hospitalisation, and could have been a result of the development
of delirium.
Furthermore, previous research has indicated that delirium during hospitalisation is
an independent predictor of reduced functional status during the year after a
medical admission to hospital (McCusker et al. 2001b). Additionally, research
conducted by Rudolph et al. (2010) investigated the impact of delirium following
cardiac surgery on functional decline for the patient. The presence of functional
decline was investigated in 190 patients with results indicating that delirium was
associated with a decline in function at one month following discharge. Not only can
the development of delirium have implications for the physical decline of a patient
during hospitalisation, it can have long term effects for the patient post discharge.
174
Patients that developed delirium in the case control study were also likely to have
an alteration to their continence during their admission, leading to increased
episodes of incontinence. This finding was based on reports that a patient was
continent on admission and had an episode of incontinence during their admission.
Patients often experienced incontinence while displaying symptoms of delirium.
Change in continence has not previously been described as an outcome of delirium
in the literature. Further research is required to investigate how this impacted on a
patient’s continence following discharge.
5.7.4 Increased length of stay
Patients that developed delirium in the case control study had a longer average
length of stay than control patients. Patients with delirium were on average likely to
stay in hospital 3.16 days longer than patients that did not develop delirium.
However, logistic regression did not find length of stay to be independently
associated with delirium. Findings of other research indicate that incident delirium
is independently associated with increased length of stay even after adjusting for
co morbidity, severity of illness, and other confounding variables (McCusker et al.
2003). McCusker et al. (2003) investigated if incident delirium increased length of
stay for hospitalised patients. The mean length of stay for 36 cases of incident
delirium and 118 matched controls were 20.27 and 10.7, respectively. The authors
speculated that increased length of stay could have occurred for a number of
reasons, including delirium as a consequence of inter current illness or
complication, or deterioration in physical functioning of the patient. Furthermore,
the diagnosis of delirium may have prompted further investigation and tests that
resulted in a longer hospital stay (McCusker et al 2003). The patients in the present
case control study had, on average, shorter lengths of stay than indicated in
previous research. Differences in severity of illness on admission, the identification
and management of patients with delirium, varying patient discharge practices
between countries (Canada and Australia), and discharge destination, may have
contributed to this difference.
175
Patients who had delirium also had longer length of stay when compared with
patients with possible delirium. Identification of delirium may have resulted in
patients staying in hospital longer, until a resolution of delirium. As patients with
possible delirium did not have a diagnosis of delirium, they may have been
discharged from the acute facility with delirium. Previous research has indicated
that delirium often remains present on the day of discharge from hospital (Cole
2010). This persistent delirium is often associated with poorer outcomes for these
patients compared to patients that had the delirium resolved (Inouye et al. 2007).
This information highlights the importance of recognising delirium in patients
during admission to hospital and to be continually monitoring for the presence of
delirium during admission and on discharge from hospital. Further research on the
outcomes of patients suspected to have delirium, but not diagnosed during
admission, is required to investigate the impact of missed diagnosis.
5.8 Delirium follow up
Long term, delirium can be detrimental to the patient and follow up to monitor
progress is important. The case control study investigated the number of patients
that were referred for, or given information regarding, follow up of their delirium.
Six patients that developed delirium were specifically referred to a delirium clinic
post discharge from hospital. Previous research conducted by McAvay et al. (2006)
on 433 patients who developed incident delirium during hospitalisation suggested
that delirium experienced by patients in hospital could have implications for their
outcomes up to one year following the episode. One of the potential residual
effects of delirium is cognitive decline. Research has indicated that patients who
experience a delirium during hospitalisation have greater decline, according to
cognition measures, compared to control patients (Jackson et al. 2004). This
evidence emphasises the importance for patients of continuity of care in order to
address potential residual effects of incident delirium.
Follow up for delirium should include professional monitoring of cognition and
treatment of potentially lingering symptoms of delirium. Follow up care may also
give the patient an opportunity to discuss events experienced during their
176
hospitalisation, as the psychological implications of experiencing an episode of
delirium should also be considered. Researchers have investigated the potential for
patients to develop long term psychological issues such as post traumatic stress
disorder following an episode of delirium (Davydow 2009). Often these patients
recall frightening episodes and experience feelings of guilt associated with their
behaviour (Grover & Shah 2011). It was unclear if patients in the case control study
were given an opportunity to talk about their experiences, indicating that these
patients may be at risk of experiencing long term psychological implications as a
result of delirium.
For patients in the case control study there was no documentation to indicate that
they had been informed of their delirium episode during hospitalisation. Because
having a previous delirium was found to be a risk factor for development of delirium
in the case control study, patients should be provided with this information and
educated, where appropriate, so they are aware they may be at risk of developing
the syndrome during subsequent admissions to hospital. There was documented
evidence to suggest that some families were given information regarding delirium.
These family members may have been able to pass on this information to the
patient when they were well and able to understand. In order to prevent possible
long term complications of delirium, patients need to be actively involved and
educated about delirium.
5.9 Difficulties with implementing a delirium management policy
The findings of the survey identified a number of barriers to the development and
subsequent implementation of policies regarding delirium management.
Understanding these barriers can inform and assist implementation of delirium
management policies. One of the most common barriers identified was lack of
education about delirium identification and management. Education of health
professionals is critical to promoting skill level and confidence to identify patients
with delirium. A study conducted by Tabet et al. (2005) examined whether an
educational package for medical and nursing staff would reduce the number of
incident cases of delirium and increase recognition of cases of delirium within an
177
acute medical ward. The authors found that incidence of delirium was significantly
reduced and staff on the intervention ward recognised more delirium cases. An
Australian study used a pre test post test time series randomised control trial to
investigate the impact of a web based learning website on nurses’ ability to
recognise delirium (McCrow, Sullivan & Beattie 2014). In total, 175 registered
nurses participated in the web based delirium intervention program developed by
the authors and aimed to improve delirium knowledge. Recognition and knowledge
of delirium was assessed using questionnaires and vignette scenarios. The authors
found that web based delirium education was effective in improving nurses’
knowledge and ability to recognise delirium (McCrow, Sullivan & Beattie 2014).
However, further research is required to determine the effect of educational
interventions on clinical practice. Other barriers reported by respondents included
personal experience and differences in opinion about ways to deal with delirium,
highlighting the need to implement evidence based delirium management.
Implementing strategies based on best available evidence in order to overcome
barriers should be a priority for health services.
5.10 Strengths and limitations of the research
The research presented in this thesis has made an important contribution to the
evidence base regarding incident delirium in the acute medical setting. Not only
have risk factors for incident delirium in medical in patients been identified through
a systematic review of existing studies, these have also been compared with
findings from an Australian setting using case control methods.
The use of systematic review methods is an effective way to isolate, critically
evaluate, and synthesise previous research (Joanna Briggs Institute 2011). As such,
systematic review methods were appropriate for identifying the risk factors for
delirium in medical in patients. The use of systematic review methods provided a
comprehensive review of the current research available on risk factors for incident
delirium in medical patients and identified risk factors that were most commonly
investigated across the studies. The systematic review also identified gaps in the
evidence base and risk factors that need to be further investigated in this setting.
178
The systematic review has also provided more reliable and accurate data on
delirium risk factors than individual studies alone; as such these results can be
generalised and extrapolated to the general medical population more broadly. The
review findings also provided a solid evidence base against which the findings of the
case control study could be compared. The systematic review has therefore filled a
gap and contributed to the evidence base regarding risk factors for delirium in an
acute medical setting.
Despite the strengths of the systematic review, there are limitations that should be
acknowledged. Publication bias (only reporting significant results) in the included
studies is possible. However to counteract this, when necessary some authors were
contacted to obtain original data that were not reported in the published article.
This resulted in some authors providing original data sets and allowed for the
investigation of variables that were not reported in the published article.
Furthermore, not all included publications reported all characteristics of study
participants and, in order to provide a greater understanding of delirium risk
factors, results indicating the relationship of each potential risk factor with delirium
need to be reported, even if the relationship was not significant. Contact with some
authors was unsuccessful and resulted in some variables being excluded. In
addition, differences between included studies in the way some variables were
measured, prevented pooling of data and may have impacted on the overall result,
as meta analysis could not be undertaken.
Another limitation of the systematic review was that not all authors compared
characteristics of study participants who developed delirium against those who did
not. Data presented in this way could not be used for analysis. Moreover, most
studies considered likely for inclusion in the initial phase of searching were
subsequently excluded because they did not clearly indicate that patients with
prevalent delirium were excluded on admission. This may have resulted in exclusion
of studies that may have been appropriate to include.
Case control methods were used for Phase 2 of the research. Case control methods
can contribute important findings in a relatively short time and for low cost
179
(Rothman et al. 2008). The case control method was chosen for the present study
as it was conducted for the purpose of doctoral dissertation and there was a short
time frame allocated for conducting this research. Due to the nature of delirium,
and possible low incidence rates, the case control study was the most effective way
of obtaining the required sample size for the study in a short time frame, compared
to larger prospective cohort studies (during the early development of this research
study, a larger prospective cohort study was considered but due to limited
resources and timeframe decided not to proceed). Limited funds also required the
researcher to undertake all the data collection; thus, prospective screening and
recruiting of patients with delirium would not have been possible. The use of case
control methods is also effective for the investigation of diseases and determining
possible exposure risks. Case control designs, while they cannot yield incidence
rates, can provide odds ratios, which are derived from the number of individuals
exposed to each variable in each of the case and control groups. Because an aim of
the present research was to determine risk factors for delirium, case control design
was considered an effective method for addressing this aim. Because case control
study designs are undertaken retrospectively this allowed the researcher to
determine how patients with delirium were diagnosed, managed and the
medications they received. This study therefore provides a baseline for further
delirium research in this setting.
While case control study design has many strengths, some limitations to this
approach should be acknowledged. One of the limitations of the case control
records review was that the researcher relied upon documentation in the medical
record. The medical records of patients are known to be a potentially unreliable
source of information due to poor documentation. Retrospective diagnosis of
delirium from medical records can also be unreliable. In order to deal with this only
the medical records of patients diagnosed with delirium, based on discharge
summary documentation, were included in the case group. It was likely a number of
patients that had delirium were missed using this method. The small number of
patients identified from the control group that had been diagnosed with delirium
during admission, but for whom the coding for delirium in their records at discharge
did not occur highlighted this. Furthermore, a possible delirium group was also
180
identified which highlights potential cases that could have been identified if the
research was conducted prospectively. Because it was an aim of the research to
determine what and how much information is documented when patients have
delirium, it was considered that retrospective analysis was necessary to obtain this
information.
Another limitation of the case control study is the potential for differential
reporting of information between cases and controls. For example, for patients
exhibiting signs of delirium, doctors may have been more likely to investigate for
evidence of a past history of delirium. Conversely, patients who did not develop
delirium during admission may not have had a previous episode of delirium
documented, even if they had previously had delirium. External validity and
generalisability of the case control study may be limited to populations in the
Australian setting. However, this method has provided a preliminary understanding
of the management of delirium in this setting and further research can be
conducted to compare findings between Australian hospitals.
The survey method was chosen for Phase 3 of this research as it provides the ability
to include a large sample, and in the present research this enabled the collection of
data representing all public and private hospital networks in Melbourne. The data
gathered provided comprehensive information regarding the policies and
procedures used throughout hospitals in the metropolitan area. Another strength
and benefit is the low cost attributed to conducting the survey and the convenience
of being able to undertake the survey through telephone calls and emails. The
survey method also ensured that the researchers own biases were eliminated.
For the survey, there were also some limitations. The response rate was poor,
especially in the private setting. This may have provided a biased sample.
Specifically, due to the nature of the survey, respondents may have only been
willing to participate if their organisation had policies in place for delirium
management. In fact, potential respondents from two private hospitals declined to
participate because they stated their organisation did not have a policy for
management of delirium. However, this is an important finding suggesting that a
181
number of hospitals and health care settings do not have a policy or
recommendations about care for patients with delirium.
5.11 Summary of discussion
This chapter has presented a discussion of the research findings in relation to the
aims of the research. Risk factors for incident delirium identified in the systematic
review and case control study have been discussed by comparing the findings to
previous research. Outcomes experienced by patients who developed delirium in
the case control study have also been identified and compared to the findings of
other relevant research. Management of delirium identified in the case control
study as well as polices for delirium management identified in the survey have also
been examined. The limitations of the study have been outlined. The following
chapter will provide a concluding statement in relation to each of the study aims
and will present implications for practice including education and policy and
recommendations for further research.
182
Chapter 6 Conclusion and
Recommendations
6.1 Introduction
The purpose of this research was to add to the evidence regarding delirium risk
factors, clinical characteristics and management of patients with incident delirium
in the acute general medical setting. In order to address the aims of the research
this study was conducted in three phases. The first aim of the study was to
synthesise the evidence for risk factors related to the development of incident
delirium in general medical patients. A systematic review of the literature was
undertaken to address the first research aim. To address the second aim, which
involved describing the characteristics of general medical patients who developed
incident delirium in Australia, a retrospective case control study was undertaken.
Results of the case control study were then compared to findings of the systematic
review. Case control methods were also used to examine the clinical management
of patients who developed delirium in an Australian hospital, particularly
considering the diagnosis of delirium, assessment of risk, medication management
as well as environmental and clinical prevention and management strategies.
Finally, to address the third aim of the research, which was to examine and describe
current delirium management practices in acute hospital settings in Melbourne,
Australia, a survey was undertaken to determine what policies and guidelines are
used in hospitals to detect and manage patients with delirium.
Incident delirium is a significant problem in the acute health care setting. This
research has addressed a number of gaps regarding risk factors, characteristics,
outcomes and management of general medical patients with incident delirium in
the Australian setting. When identified early by health professionals and active
management is implemented, duration and severity of incident delirium can be
reduced. Identification of delirium is important to the safety and quality of patient
care and health professionals should be alert for signs of delirium in older patients
admitted to the hospital setting. As identified in this study, nurses are often the first
183
to recognise and respond to changes in patients’ behaviour. This indicates the need
for nursing staff to be educated on how to appropriately recognise delirium.
Patients with delirium that are promptly diagnosed and appropriately managed
have reduced potential for complications and poor outcomes. The care for patients
with delirium included in the case control study was somewhat haphazard and
highlights the need to implement consistent management in hospital settings. The
data collected during this study adds to the growing knowledge base in Australia
regarding detection, diagnosis and management of incident delirium in older
patients admitted to a general medical setting.
In the final chapter of this thesis, the aims are restated and a summary of the key
findings specific to each aim is presented, together with a list of the
recommendations informed by these findings.
6.2 Aims of the research
The overall purpose of this research was to add to current evidence regarding the
clinical risk factors for incident delirium, characteristics and management of
hospitalised medical in patients who develop incident delirium in Australia. The
specific aims of this research were to:
1. Systematically review the evidence for risk factors related to the
development of incident delirium in general medical patients.
2. Describe the characteristics of medical patients who develop incident
delirium during hospitalisation in Australia. These include the demographic
characteristics (age, gender, residency prior to admission, functional and
cognitive status prior to admission), potential risk factors (predisposing and
precipitating), and outcomes for patients including discharge destination,
length of stay in hospital and medication treatment.
3. Examine and describe the current policy for delirium management in acute
care hospital settings in Victoria.
184
6.2.1 Aim one
The first aim of this study was to:
Systematically review the evidence for risk factors related to development of
incident delirium in general medical patients.
This study has added to the evidence regarding risk factors that contribute to
development of incident delirium in acute general medical patients. The review has
provided further evidence of the impact that dementia and cognitive impairment
have on development of delirium. Both dementia and cognitive impairment were
identified as significant risk factors for incident delirium in hospitalised medical
patients. Assessment for the presence and severity of dementia and cognitive
impairment is essential to the care of patients at risk of delirium and care needs to
be taken when patients with these conditions are admitted to hospital, including
the implementation of preventative measures.
Age is a well documented risk factor for delirium but this review found that factors
related to advanced age, such as a greater severity of illness, functional impairment,
dementia, cognitive impairment and visual impairment appear to have a greater
impact on delirium in this medical population than age alone. This is an important
finding as treatment strategies can be targeted towards patients with these factors
and will help to encourage nurses and medical staff to carefully monitor for delirium
in vulnerable patients, especially those with dementia, as delirium is often difficult
to detect in these patients.
Patients with functional impairment were also identified in the systematic review as
at risk of delirium. This highlights the importance of health professionals in medical
settings undertaking assessments to identify patients with functional impairment
and acting on their findings. Factors that were not associated with incident delirium
in the medical population according to the systematic review findings included male
gender and depression. Assessment of precipitating risk factors was not undertaken
in the systematic review, as only one study reported evidence regarding
precipitating factors. The review findings highlight factors that should be assessed
when a patient is admitted to a medical setting. This review has added further
185
evidence to the growing body of knowledge regarding risk factors for incident
delirium in the acute general medical setting.
6.2.2 Aim two
The second aim of the study was to:
Describe the characteristics of medical patients who develop incident delirium
during hospitalisation in Australia. These include the demographic characteristics
(age, gender, residency prior to admission, functional and cognitive status prior to
admission), potential risk factors (predisposing and precipitating), and outcomes for
patients including discharge destination, length of stay in hospital and medication
treatment.
This aim was addressed by undertaking a retrospective case control study that
examined medical records of patients with delirium and a control group with no
delirium. Predisposing factors specific to patients in the medical setting identified in
the systematic review informed factors for which data were collected in the case
control study. As the systematic review did not include any studies conducted in the
Australian setting, the case control study allowed the comparison of factors
identified outside the Australian setting with those identified in the Australian
context. The case control study confirmed the systematic review findings that both
dementia and cognitive impairment are strongly related to the development of
incident delirium in acute general medical patients. According to previous research,
this finding appears to be consistent across different hospital settings as well as
other developed countries with an ageing population.
In the systematic review and in the case control study, advanced age of greater
than 80 years was not found to be a risk factor for delirium. Specifically, advanced
age of greater than 80 years was not independently associated with incident
delirium when compared to other factors such as cognitive impairment, dementia
and functional impairment. However, age was associated with incident delirium, as
patients with delirium were significantly older than patients in the control group.
This result highlights that for older patients (aged 80 years or over), other factors
186
such as cognitive impairment, have a greater impact on the incidence of delirium
compared to advanced age alone.
The case control study also added evidence for other predisposing risk factors for
incident delirium identified in the systematic review, such as functional impairment
and visual impairment. Additionally, the case control study provided evidence
regarding predisposing risk factors for incident delirium that were not identified in
the systematic review. In particular, previous delirium and a fracture present on
admission were also identified as predisposing risk factors for incident delirium in
the case control study.
The case control study also identified precipitating factors for development of
incident delirium among patients in the medical setting. This information is
important, as no research studies were identified that have investigated
precipitating risk factors for delirium in the Australian medical setting. Factors
found to be independently associated with incident delirium in this population
were: use of an indwelling catheter, adding more than three new medications
during hospitalisation, and abnormal serum sodium level during admission. These
findings are consistent with previous international research into precipitating
factors and further add to the evidence regarding precipitating factors in the
Australian medical setting.
Little research has been undertaken to test a risk prediction model in the Australian
setting. The current research has identified a set of risk factors that are
independently associated with incident delirium and could potentially be used in
the Australian medical health care setting to test a risk prediction model for the
identification of delirium risk.
In the case control study the outcomes for patients that developed delirium were
also assessed. This study has provided further evidence that incident delirium
contributes to worse outcomes for patients in hospital. The patients with incident
delirium had more falls, were more likely to be discharged to a care facility and
were also more likely to have a decline in their functional status. Thus, patients are
187
experiencing outcomes that may not have otherwise occurred if the delirium was
prevented. Patients with delirium in the case control study were also more likely to
have a longer length of stay than those in the control group. Longer length of stay
increases the likelihood of complications (such as hospital acquired infections and
pressure injuries) associated with hospitalisation and can potentially further impact
on the outcomes for patients with incident delirium.
These findings add to the growing body of evidence in the Australian medical acute
health care context regarding the risk factors that are independently associated
with incident delirium and the outcomes that these patients experience.
6.2.3 Aim three
The final aim of the study was to:
Examine and describe the current state of delirium management in acute hospital
settings in Victoria.
This final aim of the research was addressed by the case control study and a survey
of public and private hospitals in Melbourne. In the case control study, medical
records of patients with delirium were examined to determine how they were
managed in the hospital setting. This included examining documentation of how
delirium was diagnosed, if patients had a cognitive assessment on admission, the
medications given to patients and the management strategies undertaken by staff.
The Clinical Practice Guidelines for the Management of Delirium in Older People
(Clinical Epidemiology and Health Service Evaluation Unit and Delirium Clinical
Guidelines Expert Working Group 2006) were used as the guidelines against which
to compare policy informing practice within hospitals.
The survey was undertaken to determine the policies in existence for the
management of patients with delirium in hospitals in Melbourne. The survey was
used to describe the policies in relation to identification, assessment and
management practices for patients with delirium. It is the first study to document
the availability of current delirium management policies and procedures relevant to
188
cognitive assessment, delirium risk factor assessment, delirium diagnosis, and
pharmacological management in Melbourne’s public and private hospitals.
Although some hospitals have implemented policies for delirium management it is
clear that more work is needed to ensure that all hospitals have a delirium
management policy. Variability in policy availability and content shows that a clear
systematic approach to the implementation of best practice guidelines is required
to help guide the practice of clinicians.
Practice identified in the case control study was characterised by inconsistent
application of existing knowledge, including detection of delirium and treatment
practices. Assessment of patients’ risk for delirium is important to identifying high
risk patients that require delirium prevention and/or management strategies. In the
case control study, assessment of patients’ level of risk for delirium was not
documented for any patient during their admission. There were also indications in
the survey that not all hospitals had a policy recommendation for assessment of
patients’ risk for delirium. Assessment of cognitive function was also rarely
documented according to the findings of the case control study.
Patients in the case control study often showed signs of delirium days before health
professional’s actually diagnosed delirium. Further, when delirium was identified,
there was no documented evidence to suggest that a diagnostic tool was used to
aid the diagnosis. Nurses were often the first to recognise and document a change
in cognition, yet there was poor documentation indicating that it was further
escalated to medical or other health care staff. Delay in delirium diagnosis and lack
of use of reliable tools to diagnose it represents potentially avoidable delays in
providing effective treatment and also reflects lack of consistently implemented
procedures for detecting and monitoring of delirium. In order to reduce the many
consequences of delirium, the application of standard procedures in relation to
behavioural and educational intervention is necessary. Poor recognition and
diagnosis of delirium was also reflected in the number of patients that were
identified as having a possible delirium during hospitalisation compared to the
control group. These patients also experienced worse outcomes, which may have
been avoided if the possible delirium was recognised. The rate of failure to identify
189
potential patients with delirium highlights that the approach to screening for
delirium is inadequate. Use of a tool to screen for delirium could increase the
recognition of these patients.
Case control study patients that developed incident delirium were frequently
managed with antipsychotic agents. The frequency of use of pharmacological
management highlighted that clinicians often rely upon medication to manage
behaviours associated with delirium. This practice does not align with the current
recommendations to limit the use of pharmacological treatments and undertake
non pharmacological strategies as a first response. Survey respondents also noted
that there were varying recommendations in their policy guidelines in regards to
medication management. Hospital respondents reported different medication and
dose recommendations for patients that develop delirium. This suggests that
clearer guidelines are required for when pharmacological treatment is necessary
and when it should be avoided. Education of nursing and medical staff is required to
minimise the overuse of medications for delirium management and to use doses
appropriate for older patients.
The overall management of the patients with incident delirium was haphazard and
did not follow established guidelines. Documentation of non pharmacological
strategies to manage and prevent delirium was poor, reflecting either poor
knowledge of the strategies to prevent delirium or poor documentation practices.
Health professionals also did not provide adequate follow up care for patients with
incident delirium. Lack of follow up care could have detrimental effects on the long
term outcomes of patients who develop delirium.
190
6.3 Recommendations and implications for practice and policy
Findings from this study have informed a number of recommendations for
management of delirium in the acute care setting. The systematic review identified
a number of risk factors that should be identified and documented on admission to
hospital. The case control study and survey identified that despite the existence of
the Clinical Practice Guidelines for the Management of Delirium in Older People,
implementation of the guidelines has been poor and not all patients are receiving
adequate delirium management.
Recommendations for practice and policy:
Ensure consistent practice and policies regarding the identification and
management of delirium in all health care settings that have patients at risk
of developing delirium.
Delirium Clinical Care Standard (ACSCHC 2015) needs to widely implemented
and rigorously tested in the Australian population in order to promote
awareness of delirium as a fundamental quality and safety issue.
Policies relating specifically to management of patients with delirium be
implemented to provide a consistent evidence based approach.
Due to the high risk for delirium, all older people (aged 65 years and over) be
screened using valid and reliable tools on admission, for cognitive
impairment and undiagnosed dementia.
Appropriate screening and diagnostic tools for delirium be implemented into
acute health care settings to improve the recognition of delirium.
Delirium risk factor assessments be undertaken for all older adults admitted
to the medical setting in order to identify patients at high risk for delirium
and implement appropriate prevention strategies.
Interventions for delirium management (for example monitoring for
delirium, providing adequate diet and fluids, encouraging mobility, reducing
191
noise and sleep interruptions) be put in place for patients identified as high
risk for delirium.
Family members/significant others or carers of patients with a cognitive
impairment or dementia be made aware of their family members’ condition
and increased risk of delirium.
Patients and family members/carers/significant others be informed if a
patient develops delirium and support (including education) be provided for
family members and carers.
Targeted education (such as online learning packages) regarding delirium
identification, risk factors, and outcomes of delirium, and prevention and
management strategies be provided to all health professionals that come
into contact with older people who have the potential to develop delirium.
Follow up for the patient in a delirium clinic be implemented in order to
provide continuity of care in delirium management and potentially prevent
long term complications of delirium.
192
6.4 Recommendations for future research
Findings from this study have informed a number of recommendations for research
regarding incident delirium in the acute care setting. These recommendations
reflect a gap in delirium research in the Australian setting.
Recommendations for research:
Further studies investigating delirium risk factors according to specific
health care settings are required e.g. medical or surgical or ICU.
Researchers separate incident delirium from prevalent delirium in order to
define risk factors that relate only to the development of incident delirium
during hospitalisation.
Measurement of risk factors be carried out using valid and reliable
standardised tools so that the results may be more easily combined and
analysed.
Results of studies be published even if they are not significant or positive
results, including evidence of factors that are not related to delirium.
Investigate and assess the usability of delirium screening tools and
diagnostic instruments in clinical practice in the Australian setting.
Assess usability of different cognitive assessment tools that allow for quick
identification of cognitive impairment in the Australian setting such as the
4AT (Bellelli et al. 2014).
Assess the usability and investigate the implementation of the
recommendations outlined in the Clinical Practice Guidelines for the
Management of Delirium in Older People.
Investigate the implementation and usability of delirium risk prediction tools
in clinical practice in the Australian setting.
Investigate the implementation of educational strategies for nursing and
other health care staff in the recognition and management of delirium.
193
6.5 Conclusion
The purpose of this research was to add to the evidence base regarding clinical risk
factors, characteristics and management of hospitalised general medical patients
who develop incident delirium in Australia. It is clear from all three phases of this
research that delirium is a serious problem that needs to be addressed in the acute
health care setting. The ageing population in Australia and other developed
countries around the world has seen an increase in the number of patients with
dementia and cognitive impairment. Because dementia and cognitive impairment
are leading risk factors for delirium it is important to identify these and implement
preventative strategies. Knowledge and awareness of risk factors for delirium
enable health care professionals to be proactive in implementing prevention
strategies. Thus, the identification of risk factors during a patient’s admission to
hospital is an essential step in implementing strategies to reduce the incidence of
delirium.
This research highlights gaps in practice in the Australian acute clinical setting
regarding how delirium is identified and managed. These gaps in practice were
especially evident with a group of patients with possible, but undetected, delirium
who were identified during the case control study. These issues need to be
addressed if the incidence of delirium is to be reduced, and delirium is to be
managed appropriately. Patients who develop delirium may be experiencing poor
outcomes unnecessarily. As a result, the incidence of delirium should be a key
target area for the improvement of practice in hospital settings. It appears from this
study that the management of delirium in acute medical settings in the Australian
context needs to be substantially improved. A number of clinical practice
recommendations have been outlined as well as recommendations for future
research on incident delirium in the Australian health care setting. Delirium remains
a common, poorly recognised problem in the Australian health care setting. All
members of the health care team need to be involved in implementing strategies to
improve the safety and quality of care provided to these patients.
194
ReferencesAbbott, D 2014, Applied Predictive Analytics: Principles and Techniques for the
Professional Data Analyst, Wiley, San Diego.
Adunsky, A, Levy, R, Heim, M, Mizrahi, E & Arad, M 2003, 'The unfavorable nature
of preoperative delirium in elderly hip fractured patients', Archives of Gerontology
and Geriatrics, vol. 36, no. 1, pp. 67 74.
Ajilore, O & Kumar, A 2004, 'Delirium and Dementia', The Journal of Lifelong
Learning in Psychiatry, vol. 11, no. 2, pp. 210 9.
Alagiakrishnan, K, Marrie, T, Rolfson, D, Coke, W, Camicioli, R, Duggan, D,
Launhardt, B, Fisher, B, Gordon, D, Hervas Malo, M, Magee, B & Wiens, C 2009,
'Gaps in patient care practices to prevent hospital acquired delirium', Canadian
Family Physician, vol. 55, no. 10, pp. e41 6.
Albert, MS, Levkoff, SE, Reilly, C, Liptzin, B, Pilgrim, D, Cleary, PD, Evans, D & Rowe,
JW 1992, 'The Delirium Symptom Interview: An Interview for the Detection of
Delirium Symptoms in Hospitalized Patients', Journal of Geriatric Psychiatry and
Neurology, vol. 5, no. 1, pp. 14 21.
American Psychiatric Association 1999, 'Practice Guideline for the Treatment of
patients with Delirium', American Journal of Psychiatry, vol. 156, no. 5, pp. 1 20.
American Psychiatric Association 2000, Diagnostic and Statistical Manual of Mental
Disorders, 4th edn, American Psychiatric Publishing, Washington DC.
American Psychiatric Association 2013, Diagnostic and Statistical Manual of Mental
disorders, 5th edn, American Psychiatric Publishing, Washington DC.
195
Anderson, D 2005, 'Preventing delirium in older people', British Medical Bulletin,
vol. 73 74, no. 1, pp. 25 34.
Arnold, E 2005, 'Sorting out the 3 D's: delirium, dementia, depression: learn how to
sift through overlapping signs and symptoms so you can help improve an older
patient's quality of life', Holistic Nursing Practice, vol. 19, no. 3, pp. 99 105.
Australian Commission on Safety and Quality in Health Care 2011, National
Standards on Safety and Quality Health, ACSQHC, Sydney.
Australian Commission on Safety and Quality in Health Care 2014, A better way to
care: Safe and high quality care for patients with cognitive impairment (dementia
and delirium) in hospital, ACSQHC, retrieved 5th December 2014,
<http://www.safetyandquality.gov.au/our work/cognitive impairment/better way
to care/>
Australian Commission on Safety and Quality in Health Care 2015, Delirium Clinical
Care Standard, ACSQHC, retrieved 3rd February 2015,
<http://www.safetyandquality.gov.au/our work/clinical care standards/delirium
clinical care standard/>
Australian Institute of Health and Welfare 2012, Dementia in Australia, Australian
Institute of Health and Welfare, retrieved 3rd July 2013, <http://www.aihw.gov.au>.
Ayto, J 2005,Word origins: the hidden histories of English words from A to Z, A. & C.
Black, London.
Balas, MC 2005, 'The course of delirium in older surgical intensive care unit
patients', Dissertation thesis, PhD thesis, University of Pennsylvania. Retrieved July
2012 from ProQuest Dissertation and Theses.
196
Bates, DW, Pruess, K, Souney, P & Platt, R 1995, 'Serious falls in hospitalized
patients: correlates and resource utilization', The American Journal Of Medicine, vol.
99, no. 2, pp. 137 43.
Bellelli, G, Morandi, A, Davis, DHJ, Mazzola, P, Turco, R, Gentile, S, Ryan, T, Cash, H,
Guerini, F, Torpilliesi, T, Del Santo, F, Trabucchi, M, Annoni, G & MacLullich, AMJ
2014, 'Validation of the 4AT, a new instrument for rapid delirium screening: a study
in 234 hospitalised older people', Age and Ageing, vol. 44, no. 4, pp. 496 502
Bjoro, K 2008, 'Pain treatment: A risk factor for delirium in older adults with hip
fracture', 3407045 thesis, PhD thesis, The University of Iowa. Retrieved April 2012
from ProQuest Dissertation and Theses.
Boettger, S & Breitbart, W 2011, 'Phenomenology of the subtypes of delirium:
phenomenological differences between hyperactive and hypoactive delirium',
Palliative & Supportive Care, vol. 9, no. 2, pp. 129 35.
Boettger, S, Friedlander, M, Breitbart, W & Passik, S 2011, 'Aripiprazole and
haloperidol in the treatment of delirium', Australian & New Zealand Journal of
Psychiatry, vol. 45, no. 6, pp. 477 82.
Boettger, S, Passik, S & Breitbart, W 2009, 'Delirium superimposed on dementia
versus delirium in the absence of dementia: Phenomenological differences',
Palliative & Supportive Care, vol. 7, pp. 495 500.
Bogardus, S, Desai, M, Williams, C, Leo Summers, L, Acampora, D & Inouye, S 2003,
'The effects of a targeted multicomponent delirium intervention on post discharge
outcomes for hospitalized older adults', The American Journal Of Medicine, vol. 114,
no. 5, pp. 383 90.
197
Bouch, DC & Thompson, JP 2008, 'Severity scoring systems in the critically ill',
Continuing Education in Anaesthesia, Critical Care & Pain, vol. 8, no. 5, pp. 181 5.
Bergstrom, N, Braden, BJ, Laguzza, A & Holman, V 1987, 'The Braden Scale for
Predicting Pressure Sore Risk', Nurs Res, vol. 36, no. 4, pp. 205 10.
Breitbart, W, Gibson, C & Tremblay, A 2002, 'The delirium experience: delirium
recall and delirium related distress in hospitalized patients with cancer, their
spouses/caregivers, and their nurses', Psychosomatics, vol. 43, no. 3, pp. 183 94.
British Geriatric Society 2006, Guidelines for the prevention, diagnosis and
management of delirium in older people in hospital, retrieved 8th April 2013,
<http://www.bgs.org.uk >.
Bruce, N, Pope, D & Stanistreet, D 2008b, 'Surveys', in Quantitative Methods for
Health Research, John Wiley & Sons, Ltd, pp. 129 91.
Burleigh, E, Reeves, I, McAlpine, C & Davie, J 2002, 'Can doctors predict patients'
abbreviated mental test scores', Age and Ageing, vol. 31, no. 4, pp. 303 6.
Burns, A, Gallagley, A & Byrne, J 2004, 'Delirium', Journal of Neurology,
Neurosurgery & Psychiatry, vol. 75, no. 3, pp. 362 7.
Caeiro, L, Ferro, JM, Albuquerque, R & Figueira, ML 2004, 'Delirium in the first days
of acute stroke', Journal Of Neurology, vol. 251, no. 2, pp. 171 8.
Campbell, N, Perkins, A, Hui, S, Khan, B & Boustani, M 2011a, 'Association between
prescribing of anticholinergic medications and incident delirium: A cohort study',
Journal of the American Geriatrics Society, vol. 59, no. Suppl 2, pp. S277 S81.
198
Caraceni, A & Grassi, L 2004, Delirium: Acute confusional states in palliative
medicine, Oxford University Press, New York.
Casey, DA, DeFazio, JV, Vansickle, K & Lippmann, SB 1996, 'Delirium. Quick
recognition, careful evaluation, and appropriate treatment', Postgraduate medicine,
vol. 100, no. 1, pp. 121 4, 8, 33 4.
Cerejeira, J & Taylor, J P 2011, 'Diagnosing, treating and preventing delirium', GM:
Midlife & Beyond, vol. 41, no. 2, pp. 081 85.
Cheah, E, Rajaram, S, Chua, HC, Ng, HL, Tim, HM, Cinnappan, S & Lim, ST 2011,
'Managing the cognitive impairment of elderly patients using the Mini Mental State
Examination (MMSE)', BMJ Supportive & Palliative Care, vol. 1, no. 1, p. 95.
Choi, S H, Lee, H, Chung, TS, Park, KM, Jung, YC, Kim, SI & Kim, JJ 2012, 'Neural
network functional connectivity during and after an episode of delirium', The
American Journal Of Psychiatry, vol. 169, no. 5, pp. 498 507.
Ciampi, A, Dyachenko, A, Cole, M & McCusker, J 2011, 'Delirium superimposed on
dementia: defining disease states and course from longitudinal measurements of a
multivariate index using latent class analysis and hidden Markov chains',
International Psychogeriatrics, vol. 23, no. 10, pp. 1659 70.
Clegg, A & Young, JB 2011, 'Which medications to avoid in people at risk of delirium:
a systematic review', Age and Ageing, vol. 40, no. 1, pp. 23 9.
Clinical Epidemiology and Health Service Evaluation Unit and Delirium Clinical
Guidelines Expert Working Group 2006. Clinical Practice Guidelines for the
199
Management of Delirium in Older People, Victorian Government Department of
Human Services, retrieved 1st May 2012, <http://www.health.vic.gov.au>.
Cole, MG 2010, 'Persistent delirium in older hospital patients', Current Opinion
Psychiatry, vol. 23, no. 3, pp. 250 4.
Cole, M, McCusker, J, Bellavance, F, Primeau, F, Bailey, R, Bonnycastle, M &
Laplante, J 2002, 'Systematic detection and multidisciplinary care of delirium in
older medical inpatients: a randomized trial', CMAJ: Canadian Medical Association
Journal = Journal De L'association Medicale Canadienne, vol. 167, no. 7, pp. 753 9.
Cole, M, Primeau, F, Bailey, R, Bonnycastle, M, Masciarelli, F, Engelsmann, F, Pepin,
M & Ducic, D 1994, 'Systematic intervention for elderly inpatients with delirium: a
randomized trial', CMAJ: Canadian Medical Association Journal = Journal De
L'association Medicale Canadienne, vol. 151, no. 7, pp. 965 70.
Cole, MG, Primeau, F & McCusker, J 1996, 'Effectiveness of interventions to prevent
delirium in hospitalized patients: a systematic review', CMAJ: Canadian Medical
Association Journal = Journal De L'association Medicale Canadienne, vol. 155, no. 9,
pp. 1263 8.
Cull, EJ, Kent, B, Phillips, NM and Mistarz, R 2012, Protocol: Risk factors for incident
delirium in acute medical in patients: a systematic review, JBI library of systematic
reviews, vol. 10, no. 56, Supplement, pp. 1 12.
Dahl, MH, Rønning, OM & Thommessen, B 2010, 'Delirium in acute stroke—
Prevalence and risk factors', Acta Neurologica Scandinavica, vol. 122, no. Suppl 190,
pp. 39 43.
200
Dasgupta, M & Hillier, LM 2010, 'Factors associated with prolonged delirium: a
systematic review', International Psychogeriatrics, vol. 22, no. 3, pp. 373 94.
Davis, DHJ, Muniz Terrera, G, Keage, H, Rahkonen, T, Oinas, M, Matthews, FE,
Cunningham, C, Polvikoski, T, Sulkava, R, MacLullich, AMJ & Brayne, C 2012,
'Delirium is a strong risk factor for dementia in the oldest old: a population based
cohort study', Brain, vol. 135, no. 9, pp. 2809 16.
Davydow, DS 2009, 'Symptoms of Depression and Anxiety after Delirium',
Psychosomatics, vol. 50, no. 4, pp. 309 16.
Day, J, Higgins, I & Keatinge, D 2011, 'Orientation strategies during delirium: are
they helpful?', Journal of Clinical Nursing, vol. 20, no. 23 24, pp. 3285 94.
de Castro, SMM, Ünlü, Ç, Tuynman, JB, Honig, A, van Wagensveld, BA, Steller, EP &
Vrouenraets, BC 2014, 'Incidence and risk factors of delirium in the elderly general
surgical patient', The American Journal of Surgery, vol. 208, no. 1, pp. 26 32.
Decrane, SK, Culp, KR & Wakefield, B 2012, 'Twelve Month Fall Outcomes among
Delirium Subtypes', Journal for Health care Quality: Promoting Excellence in Health
care, vol. 34, no. 6, pp. 13 20.
Department of Health 2012, Best care for older people. The toolkit 2012,
Department of Health and Ageing, retrieved 1st May 2013,
<http://www.health.vic.gov.au/older/toolkit/ >.
Divatia, J 2006, 'Delirium in the ICU', Indian Journal of Critical Care Medicine, vol. 10,
no. 4, p. 215.
201
Doherty, K, Archambault, E, Kelly, B & Rudolph, JL 2014, 'Delirium markers in older
fallers: a case control study', Clinical Interventions In Aging, vol. 9, pp. 2013 8.
Duppils, GS & Wikblad, K 2007, 'Patients’ experiences of being delirious', Journal of
Clinical Nursing, vol. 16, no. 5, pp. 810 8.
Edlund, A, Lundström, M, Karlsson, S, Brännström, B, Bucht, G & Gustafson, Y 2006,
'Delirium in Older Patients Admitted to General Internal Medicine', Journal of
Geriatric Psychiatry and Neurology, vol. 19, no. 2, pp. 83 90.
Elie, M, Cole, MG, Primeau, FJ & Bellavance, F 1998, 'Delirium risk factors in elderly
hospitalized patients', Journal Of General Internal Medicine, vol. 13, no. 3, pp. 204
12.
Eliopoulos, C 2010, Gerontological Nursing, Seventh edn, Lippincott William &
Wilkins, Philadelphia.
Elliott, R 2006, 'Problems with medication use in the elderly: an Australian
perspective', Journal of Pharmacy Practice and Research, vol. 36, no. 1.
Elmore, DM 2002, 'Delirium in the hospitalized elderly: Clinical indicators found in
medical records', Master’s thesis, University of Nevada, retrieved 20th February
2012, ProQuest Dissertations and Theses database.
Engel, GL & Romano, J 1959, 'Delirium, a syndrome of cerebral insufficiency',
Journal Of Chronic Diseases, vol. 9, no. 3, pp. 260 77.
Faezah, SK, Zhang, D & Yin, LF 2008, 'The prevalence and risk factors of delirium
amongst the elderly in acute hospital', Singapore Nursing Journal, vol. 35, no. 1, pp.
11 4.
202
Fick, D, Agostini, J & Inouye, S 2002, 'Delirium Superimposed on Dementia: A
Systematic Review', Journal of the American Geriatrics Society, vol. 50, no. 10, pp.
1723 32.
Flacker, JM & Lipsitz, LA 1999, 'Neural Mechanisms of Delirium: Current Hypotheses
and Evolving Concepts', Journal of Gerontology, vol. 54A, no. 6, pp. 239 46.
Flaherty, JH 2006, 'Delirium', in Principles and Practice of Geriatric Medicine, John
Wiley & Sons, Ltd, pp. 1047 60.
Flaherty, JH & Little, MO 2011, 'Matching the Environment to Patients with
Delirium: Lessons Learned from the Delirium Room, a Restraint Free Environment
for Older Hospitalized Adults with Delirium', Journal of the American Geriatrics
Society, vol. 59, pp. S295 S300.
Flaherty, JH, Rudolph, J, Shay, K, Kamholz, B, Boockvar, K, Shaughnessy, M, Shapiro,
R, Stein, J, Weir, C & Edes, T 2007, 'Delirium is a Serious and Under recognized
Problem: Why Assessment of Mental Status Should be the Sixth Vital Sign', Journal
of the American Medical Directors Association, vol. 8, no. 5, pp. 273 5.
Flaherty, JH, Shay, K, Weir, C, Kamholz, B, Boockvar, K, Shaughnessy, M, Shapiro, R,
Gordon, S, Stein, J & Rudolph, J 2009, 'The development of a mental status vital sign
for use across the spectrum of care', Journal of American Medical Directors
Association, vol. 10, no. 6, pp. 379 80.
Fong, TG, Tulebaev, SR & Inouye, SK 2009, 'Delirium in elderly adults: diagnosis,
prevention and treatment', Nature Reviews Neurology, vol. 5, no. 4, pp. 210 20.
203
Foreman, L, Cavalieri, TA, Galski, T, Dinsmore, S, Kay, PAJ & Pomerantz, S 1995,
'Occurrence and impact of suspected delirium in hospitalized elderly patients',
JAOA, vol. 95, no. 10.
Foreman, M, Wakefield, B, Culp, K & Milisen, K 2001, 'Delirium in elderly patients:
An overview of the state of the science', Journal of Gerontological Nursing, vol. 27,
no. 4, pp. 12 20.
Franco, JG, Trzepacz, PT, Meagher, DJ, Kean, J, Lee, Y, Kim, J L, Kishi, Y, Furlanetto,
LM, Negreiros, D, Huang, M C, Chen, C H, Leonard, M & de Pablo, J 2013, 'Three
Core Domains of Delirium Validated Using Exploratory and Confirmatory Factor
Analyses', Psychosomatics, vol. 54, no. 3, pp. 227 38.
Franco, JG, Valencia, C, Bernal, C, Ocampo, MV, Trzepacz, PT, de Pablo, J & Mejia,
MA 2010, 'Relationship between cognitive status at admission and incident delirium
in older medical patients', The Journal Of Neuropsychiatry And Clinical
Neurosciences, vol. 22, no. 3, pp. 329 37.
Galanakis, P, Bickel, H, Gradinger, R, Von Gumppenberg, S & Förstl, H 2001, 'Acute
confusional state in the elderly following hip surgery: incidence, risk factors and
complications', International Journal of Geriatric Psychiatry, vol. 16, no. 4, pp. 349
55.
Gillis, AJ & MacDonald, B 2006, 'Unmasking delirium', Can Nurse, vol. 102, no. 9, pp.
18 24.
Gofton, TE 2011, 'Delirium: a review', The Canadian Journal of Neurological
Sciences. Le Journal Canadien Des Sciences Neurologiques, vol. 38, no. 5, pp. 673 80.
204
Grover, S & Shah, R 2011, 'Distress due to delirium experience', General Hospital
Psychiatry, vol. 33, no. 6, pp. 637 9.
Harrington, CJ & Vardi, K 2014, 'Delirium: presentation, epidemiology, and
diagnostic evaluation (part 1)', Rhode Island Medical Journal (2013), vol. 97, no. 6,
pp. 18 23.
Harwood, DM, Hope, T & Jacoby, R 1997, 'Cognitive impairment in medical
inpatients. II: Do physicians miss cognitive impairment?', Age and Ageing, vol. 26,
no. 1, pp. 37 9.
Higgins, J & Green, S 2011, Cochrane Handbook for Systematic Reviews of
Interventions Version 5.1.0 [updated March 2011]. ed. TC Collaboration,
<http://www.cochrane handbook.org.>.
Holden, J, Jayathissa, S & Young, G 2008, 'Delirium among elderly general medical
patients in a New Zealand hospital', Internal Medicine Journal, vol. 38, no. 8, pp.
629 34.
Hshieh, TT, Yue, J, Oh, E & et al. 2015, 'Effectiveness of multicomponent non
pharmacological delirium interventions: A meta analysis', JAMA Internal Medicine,
vol. 175, no. 4, pp. 512 20.
Inouye, S 1994, 'The dilemma of delirium: Clinical and research controversies
regarding diagnosis and evaluation of delirium in hospitalized elderly medical
patients', The American Journal Of Medicine, vol. 97, no. 3, pp. 278 88.
Inouye, S 1998a, 'Delirium in hospitalized older patients', Clinics in Geriatric
Medicine, vol. 14, no. 4, pp. 745 64.
205
Inouye, S 1998b, 'Delirium in hospitalized older patients: recognition and risk
factors', Journal of Geriatric Psychiatry and Neurology, vol. 11, no. 3, pp. 118 25.
Inouye, S 1999, 'Predisposing and precipitating factors for delirium in hospitalized
older patients', Dementia and Geriatric Cognitive Disorders, vol. 10, no. 5, pp. 393
400.
Inouye, S 2000, 'Prevention of delirium in hospitalized older patients: risk factors
and targeted intervention strategies', Annals of Medicine, vol. 32, no. 4, pp. 257 63.
Inouye, S 2006, 'Delirium in Older Persons', The New England Journal of Medicine,
vol. 354, no. 11, pp. 1157 65.
Inouye, S, Baker, DI, Fugal, P & Bradley, EH 2006, 'Dissemination of the Hospital
Elder Life Program: implementation, adaptation, and success', Journal of the
American Geriatrics Society, vol. 54, no. 10, pp. 1492 9.
Inouye, S, Bogardus, ST, Baker, DI, Leo Summers, L & Cooney, LM, Jr. 2000, 'Hospital
Elder Life Program: a model of care to prevent cognitive and functional decline in
older hospitalized patients', Journal of the American Geriatrics Society, vol. 48, no.
12, pp. 1697 706.
Inouye, S, Bogardus, ST, Jr., Charpentier, PA, Leo Summers, L, Acampora, D, Holford,
TR & Cooney, LM, Jr. 1999, 'A multicomponent intervention to prevent delirium in
hospitalized older patients', The New England Journal of Medicine, vol. 340, no. 9,
pp. 669 76.
Inouye, S & Charpentier, PA 1996, 'Precipitating factors for delirium in Hospitalized
elderly persons: Predictive model and Interrelationship with baseline Vulnerability',
JAMA: The Journal of the American Medical Association, vol. 275, no. 11, pp. 852 7.
206
Inouye, S & Ferrucci, L 2006, 'Elucidating the Pathophysiology of Delirium and the
Interrelationship of Delirium and Dementia', The Journals of Gerontology Series A:
Biological Sciences and Medical Sciences, vol. 61, no. 12, pp. 1277 80.
Inouye, S, Foreman, M, Mion, LC, Katz, KH & Cooney, LC 2001, 'Nurses' recognition
of Delirium and its symptoms', Arch Intern Med, vol. 161, pp. 2467 73.
Inouye, S, Rushing, J, Foreman, M, Palmer, R & Pompei, P 1998, 'Does delirium
contribute to poor hospital outcomes?', Journal of General Internal Medicine, vol.
13, no. 4, pp. 234 42.
Inouye, S, Schlesinger, MJ & Lydon, TJ 1999, 'Delirium: a symptom of how hospital
care is failing older persons and a window to improve quality of hospital care', The
American Journal Of Medicine, vol. 106, no. 5, pp. 565 73.
Inouye, S, van Dyck, CH, Alessi, CA, Balkin, S, Siegal, AP & Horwitz, RI 1990,
'Clarifying confusion: the confusion assessment method. A new method for
detection of delirium', Annals of Internal Medicine, vol. 113, no. 12, pp. 941 8.
Inouye, S, Viscoli, CM, Horwitz, RI, Hurst, LD & Tinetti, ME 1993, 'A Predictive Model
for Delirium in Hospitalized Elderly Medical Patients Based on Admission
Characteristics', Annals of Internal Medicine, vol. 119, no. 6, pp. 474 81.
Inouye, S, Westendorp, RGJ & Saczynski, JS 2014, 'Delirium in elderly people', The
Lancet, vol. 383, no. 9920, pp. 911 22.
Inouye, SK, Zhang, Y, Jones, RN, Kiely, DK, Yang, F & Marcantonio, ER 2007, 'Risk
factors for delirium at discharge: development and validation of a predictive model',
Archives Of Internal Medicine, vol. 167, no. 13, pp. 1406 13.
207
Iseli, RK, Brand, C, Telford, M & LoGiudice, D 2007, 'Delirium in elderly general
medical inpatients: a prospective study', Internal Medicine Journal, vol. 37, no. 12,
pp. 806 11.
Jackson, JC, Gordon, SM, Hart, RP, Hopkins, RO & Ely, EW 2004, 'The association
between delirium and cognitive decline: a review of the empirical literature',
Neuropsychology Review, vol. 14, no. 2, pp. 87 98.
Joanna Briggs Institute 2011, Joanna Briggs Institute Reviewers’ Manual: 2011
edition, The Joanna Briggs Institute, Adelaide.
Jones, RN, Fong, TG, Metzger, E, Tulebaev, S, Yang, FM, Alsop, DC, Marcantonio, ER,
L. Adrienne, C, Gottlieb, G & Inouye, SK 2010, 'Aging, Brain Disease, and Reserve:
Implications for Delirium', American Journal Geriatric Psychiatry, vol. 18, no. 2, pp.
117 27.
Jones, RN, Yang, FM, Zhang, Y, Kiely, DK, Marcantonio, ER & Inouye, SK 2006, 'Does
educational attainment contribute to risk for delirium? A Potential Role for
Cognitive Reserve', The Journals of Gerontology Series A: Biological Sciences and
Medical Sciences, vol. 61, no. 12, pp. 1307 11.
Kalisvaart, KJ, Vreeswijk, R, de Jonghe, JFM, van Gool, WA & Eikelenboom, P 2006,
'Risk factors and prediction of postoperative delirium in elderly hip surgery patients:
implementation and validation of a medical risk factor model', Journal of the
American Geriatrics Society, vol. 54, no. 5, pp. 817 22.
Kamholz, B 2010, 'Update on delirium: Diagnosis, management, and
pathophysiology', Psychiatric Annals, vol. 40, no. 1, pp. 52 62.
208
Katz, S, Downs, TD, Cash, HR & Grotz, RC 1970, 'Progress in Development of the
Index of ADL', The Gerontologist, vol. 10, no. 1 Part 1, pp. 20 30.
Kazmierski, J, Kowman, M, Banach, M, Fendler, W, Okonski, P, Banys, A, Jaszewski,
R, Rysz, J, Sobow, T & Kloszewska, I 2010, 'The use of DSM IV and ICD 10 criteria
and diagnostic scales for delirium among cardiac surgery patients: results from the
IPDACS study', Journal Neuropsychiatry Clinical Neuroscience, vol. 22, no. 4, pp. 426
32.
Khurana, PS, Sharma, PSVN & Avasthi, A 2002, 'Risk factors in delirious geriatric
general medical inpatients', Indian Journal Of Psychiatry, vol. 44, no. 3, pp. 266 72.
Korevaar, JC, van Munster, BC & de Rooij, SE 2005, 'Risk factors for delirium in
acutely admitted elderly patients: a prospective cohort study', BMC Geriatrics, vol.
5, pp. 6.
Lacasse, H, Perreault, MM &Williamson, DR 2006, 'Systematic review of
antipsychotics for the treatment of hospital associated delirium in medically or
surgically ill patients', The Annals of pharmacotherapy, vol. 40, no. 11, pp. 1966 73.
Lakatos, B, Capasso, V, Mitchell, M, Kilroy, S, Lussier Cushing, M, Sumner, L, Repper
Delisi, J, Kelleher, E, Delisle, L, Cruz, C & Stern, T 2009, 'Falls in the General Hospital:
Association With Delirium, Advanced Age, and Specific Surgical Procedures',
Psychosomatics, vol. 50, no. 3, pp. 218 26.
Laurila, JV, Pitkala, KH, Strandberg, TE & Tilvis, RS 2004, 'Detection and
documentation of dementia and delirium in acute geriatric wards', General Hospital
Psychiatry, vol. 26, no. 1, pp. 31 5.
209
Leslie, DL, Marcantonio, ER, Zhang, Y, Leo Summers, L & Inouye, SK 2008, 'One Year
Health Care Costs Associated With Delirium in the Elderly Population', Archives
Internal Medicine, vol. 168, no. 1, pp. 27 32.
Levkoff, SE, Evans, DA, Liptzin, B, Cleary, PD, Lipsitz, LA, Wetle, TT, Reilly, CH,
Pilgrim, DM, Schor, J & Rowe, J 1992, 'Delirium: the occurrence and persistence of
symptoms among elderly hospitalized patients', Archives of Internal Medicine, vol.
152, no. 2, p. 334.
Lipowski, ZJ 1987, 'Delirium (Acute Confusional States)', JAMA: The Journal of the
American Medical Association, vol. 258, no. 13, pp. 1789 92.
Lipowski, ZJ 1990, Delirium: Acute Confusional States, Oxford University Press, New
York.
Litaker, D, Locala, J, Franco, K, Bronson, DL & Tannous, Z 2001, 'Preoperative risk
factors for postoperative delirium', General Hospital Psychiatry, vol. 23, no. 2, pp.
84 9.
Lonergan, E, Britton, AM, Luxenberg, J & Wyller, T 2007, 'Antipsychotics for
delirium', Cochrane Database Systematic Reviews, vol. 2, no. 2.
Lonergan, E, Luxenberg, J, Areosa Sastre, A & Wyller, TB 2009, 'Benzodiazepines for
delirium', Cochrane Database Systematic Reviews, vol. 1.
Lundstrom, M, Edlund, A, Karlsson, S, Brannstrom, B, Bucht, G & Gustafson, Y 2005,
'Multifactorial intervention program reduces the duration of delirium, length of
hospitalization, and mortality in delirious patients', Journal of the American
Geriatrics Society, vol. 53, no. 4, pp. 622 8.
210
MacLullich, AMJ, Beaglehole, A, Hall, RJ & Meagher, DJ 2009, 'Delirium and long
term cognitive impairment', International Review of Psychiatry, vol. 21, no. 1, pp.
30 42.
MacLullich, AMJ, Ferguson, KJ, Miller, T, de Rooij, SEJA & Cunningham, C 2008,
'Unravelling the pathophysiology of delirium: A focus on the role of aberrant stress
responses', Journal of Psychosomatic Research, vol. 65, no. 3, pp. 229 38.
Maldonado, JR 2008, 'Pathoetiological model of delirium: a comprehensive
understanding of the neurobiology of delirium and an evidence based approach to
prevention and treatment', Critical Care Clinics, vol. 24, no. 4, pp. 789 856.
Mangusan, RF, Hooper, V, Denslow, SA & Travis, L 2015, 'Outcomes Associated With
Postoperative Delirium After Cardiac Surgery', American Journal of Critical Care, vol.
24, no. 2, pp. 156 63.
Marcantonio, E 2007, 'Clinical Management and Prevention of Delirium', Psychiatry,
vol. 7, no. 1, pp. 42 8.
Marcantonio, E, Flacker, JM, Wright, RJ & Resnick, NM 2001, 'Reducing delirium
after hip fracture: a randomized trial', Journal of the American Geriatrics Society,
vol. 49, no. 5, pp. 516 22.
Marcantonio, ER, Goldman, L, Mangione, CM, Ludwig, LE, Muraca, B, Haslauer, CM,
Donaldson, MC, Whittemore, AD, Sugarbaker, DJ, Poss, R, Haas, S, Cook, EF, Orav, EJ
& Lee, TH 1994, 'A Clinical Prediction Rule for Delirium After Elective Non cardiac
Surgery', JAMA: The Journal of the American Medical Association, vol. 271, no. 2, pp.
134 9.
211
Margiotta, A, Bianchetti, A, Ranieri, P & Trabucchi, M 2006, 'Clinical characteristics
and risk factors of delirium in demented and not demented elderly medical
inpatients', The Journal Of Nutrition, Health & Aging, vol. 10, no. 6, pp. 535 9.
Martinez, F, Tobar, C & Hill, N 2014, 'Preventing delirium: should non
pharmacological, multicomponent interventions be used? A systematic review and
meta analysis of the literature', Age and Ageing, Epub 25 November.
Mattar, I, Chan, MF & Childs, C 2012, 'Factors causing acute delirium in critically ill
adult patients: a systematic review', JBI Library of Systematic Review, vol. 10, no. 3,
pp. 187 231.
Mattoo, SK, Grover, S & Gupta, N 2010, 'Delirium in general practice', The Indian
Journal Of Medical Research, vol. 131, pp. 387 98.
Mauk, KL 2010, Gerontological Nursing: Competencies for Care, 2nd edn, Jones and
Bartlett Publishers LLC, Sudbury.
McAvay, GJ, Van Ness, PH, Bogardus Jr, ST, Zhang, Y, Leslie, DL, Leo Summers, LS &
Inouye, SK 2007, 'Depressive symptoms and the risk of incident delirium in older
hospitalized adults', Journal Of The American Geriatrics Society, vol. 55, no. 5, pp.
684 91.
McAvay, GJ, Van Ness, PH, Bogardus Jr, ST, Zhang, Y, Leslie, DL, Leo Summers, LS &
Inouye, SK 2006, 'Older adults discharged from the hospital with delirium: 1 year
outcomes', Journal of the American Geriatrics Society, vol. 54, no. 8, pp. 1245 50.
McCrow, J, Sullivan, KA & Beattie, ER 2014, 'Delirium knowledge and recognition: A
randomized controlled trial of a web based educational intervention for acute care
nurses', Nurse Education Today, vol. 34, no. 6, pp. 912 7.
212
McCusker, J, Cole, M, Abrahamowicz, M, Han, L, Podoba, JE & Ramman Haddad, L
2001a, 'Environmental risk factors for delirium in hospitalized older people', Journal
of the American Geriatrics Society, vol. 49, no. 10, pp. 1327 34.
McCusker, J, Cole, M, Dendukuri, N, Belzile, E & Primeau, F 2001b, 'Delirium in older
medical inpatients and subsequent cognitive and functional status: a prospective
study', CMAJ: Canadian Medical Association Journal = Journal De L'association
Medicale Canadienne, vol. 165, no. 5, pp. 575 83.
McCusker, J, Cole, MG, Dendukuri, N & Belzile, E 2003, 'Does Delirium Increase
Hospital Stay?', American Geriatrics Society, vol. 51, no. 11.
McNicoll, L, Pisani, MA, Ely, E, Gifford, D & Inouye, S 2005, 'Detection of delirium in
the intensive care unit: comparison of confusion assessment method for the
intensive care unit with confusion assessment method ratings', Journal of the
American Geriatrics Society, vol. 53, pp. 495 500.
McNicoll, L, Pisani, MA, Zhang, Y, Ely, E, Siegel, MD & Inouye, SK 2003, 'Delirium in
the intensive care unit: occurrence and clinical course in older patients', Journal of
the American Geriatrics Society, vol. 51, no. 5, pp. 591 8.
Mentes, J, Culp, K, Maas, M & Rantz, M 1999, 'Acute confusion indicators: Risk
factors and prevalence using MDS data', Research in Nursing & Health, vol. 22, no.
2, pp. 95 105.
Michaud, CJ, Thomas, WL & McAllen, KJ 2014, 'Early Pharmacological Treatment of
Delirium May Reduce Physical Restraint Use: A Retrospective Study', Annals of
Pharmacotherapy, vol. 48, no. 3, pp. 328 34.
213
Michaud, L, Büla, C, Berney, A, Camus, V, Voellinger, R, Stiefel, F & Burnand, B 2007,
'Delirium: Guidelines for general hospitals', Journal of Psychosomatic Research, vol.
62, no. 3, pp. 371 83.
Milisen, K, Cremers, S, Foreman, M, Vandevelde, E, Haspeslagh, M, Geest, S &
Abraham, I 2004, 'The strain of care for delirium index: a new instrument to assess
nurses’ strain in caring for patients with delirium', International Journal of Nursing
Studies, vol. 41, no. 7, pp. 775 83.
Milisen, K, Foreman, M, Abraham, I, De Geest, S, Godderis, J, Vandermeulen, E,
Fischler, B, Delooz, H, Spiessens, B & Broos, P 2001, 'Nurse led interdisciplinary
intervention program for delirium in elderly hip fracture patients', Journal of the
American Geriatrics Society, vol. 49, no. 5, pp. 523 32.
Milisen, K, Foreman, MD, Wouters, B, Driesen, R, Godderis, J, Abraham, IL & Broos,
PL 2002, 'Documentation of delirium in elderly patients with hip fracture', J
Gerontol Nurs, vol. 28, no. 11, pp. 23 9.
Milisen, K, Lemiengre, J, Braes, T & Foreman, MD 2005, 'Multicomponent
intervention strategies for managing delirium in hospitalized older people:
systematic review', Journal of Advanced Nursing, vol. 52, no. 1, pp. 79 90.
Miller, MO 2008, 'Evaluation and management of delirium in hospitalized older
patients', American Family Physician, vol. 78, no. 11, pp. 1265 70.
Mistarz, R, Eliott, S, Whitfield, A & Ernest, D 2011, 'Bedside nurse–patient
interactions do not reliably detect delirium: An observational study', Australian
Critical Care, vol. 24, no. 2, pp. 126 32.
214
Moerman, S, Tuinebreijer, WE, de Boo, M, Pilot, P, Nelissen, RGHH & Vochteloo,
AJH 2012, 'Validation of the Risk Model for Delirium in hip fracture patients',
General Hospital Psychiatry, vol. 34, no. 2, pp. 153 9.
Morandi, A, Solberg, LM, Habermann, R, Cleeton, P, Peterson, E, Ely, EW & Schnelle,
J 2009, 'Documentation and Management of Words Associated With Delirium
Among Elderly Patients in Postacute Care: A Pilot Investigation', Journal of the
American Medical Directors Association, vol. 10, no. 5, pp. 330 4.
Mortimer, JA & Graves, AB 1993, 'Education and other socioeconomic determinants
of dementia and Alzheimer's disease', Neurology, vol. 4, no. 8, pp. S39 S44.
Mudge, AM, Maussen, C, Duncan, J & Denaro, CP 2012, 'Improving quality of
delirium care in a general medical service with established interdisciplinary care: a
controlled trial', Internal Medicine Journal, vol. 43, no. 3, pp. 270 7.
Murray, AM, Levkoff, SE, Wetle, TT, Beckette, L, Cleary, PD, Schor, J, Lipsitz, LA,
Rowe, J & Evans, DA 1993, 'Acute delirium and functional decline in the hospitalized
elderly patient', Journal of Gerontology, vol. 48, no. 5, pp. 181 6.
Nadkarni, JB, Iyengar, KP, Dussa, C, Watwe, S & Vishwanath, K 2005, 'Orthopaedic
injuries following falls by hospital in patients', Gerontology, vol. 51, no. 5, pp. 329
33.
Neelon, VJ, Champagne, MT, Carlson, JR & Funk, SG 1996, 'The NEECHAM Confusion
Scale: construction, validation, and clinical testing', Nursing Research, vol. 45, no. 6,
pp. 324 30.
O'Hanlon, S, O'Regan, N, Maclullich, AMJ, Cullen, W, Dunne, C, Exton, C & Meagher,
D 2014, 'Improving delirium care through early intervention: from bench to bedside
215
to boardroom', Journal Of Neurology, Neurosurgery, And Psychiatry, vol. 85, no. 2,
pp. 207 13.
O'Keeffe, ST 1999, 'Clinical subtypes of delirium in the elderly', Dementia and
Geriatric Cognitive Disorders, vol. 10, no. 5, pp. 380 5.
O'Keeffe, ST & Lavan, JN 1996, 'Predicting delirium in elderly patients: development
and validation of a risk stratification model', Age And Ageing, vol. 25, no. 4, pp. 317
21.
Ouimet, S, Kavanagh, BP, Gottfried, SB & Skrobik, Y 2007, 'Incidence, risk factors and
consequences of ICU delirium', Intensive Care Medicine, vol. 33, no. 1, pp. 66 73.
Pallant, J 2013, SPSS survival manual: A step by step guide to data analysis using
IBM SPSS, 5th edn, McGraw Hill, New York.
Pandharipande, P, Cotton, BA, Shintani, A, Thompson, J, Pun, BT, Morris, JA, Jr.,
Dittus, R & Ely, EW 2008, 'Prevalence and risk factors for development of delirium in
surgical and trauma intensive care unit patients', The Journal Of Trauma, vol. 65, no.
1, pp. 34 41.
Pompei, P, Foreman, M, Rudberg, MA, Inouye, SK, Braund, V & Cassel, CK 1994,
'Delirium in hospitalized older persons: outcomes and predictors', Journal of the
American Geriatrics Society, vol. 42, no. 8, pp. 809 15.
Potter, J & George, J 2006, 'The prevention, diagnosis and management of delirium
in older people: concise guidelines', Clinical Medicine, vol. 6, no. 3, pp. 303 8.
216
Quinlan, N, Marcantonio, ER, Inouye, SK, Gill, TM, Kamholz, B & Rudolph, JL 2011,
'Vulnerability: the crossroads of frailty and delirium', Journal of the American
Geriatrics Society, vol. 59 Suppl 2, pp. S262 S8.
Rabins, PV & Folstein, MF 1982, 'Delirium and dementia: diagnostic criteria and
fatality rates', The British Journal of Psychiatry, vol. 140, no. 2, pp. 149 53.
Rigney, T 2010, 'Allostatic load and delirium in the hospitalized older adult', Nursing
Research, vol. 59, no. 5, pp. 322 30.
Rockwood, K, Cosway, S, Stolee, P, Kydd, D, Carver, D, Jarrett, P & O'Brien, B 1994,
'Increasing the rocognition of delirium in elderly patients', Journal of American
Geriatrics Society, vol. 26, no. 3, pp. 252 6.
Romano, J & Engel, GL 1944, 'Delirium I: Electroencephalographic data.', Archives of
Neurology and Psychiatry, vol. 51, pp. 356 77.
Rothman, KJ, Greenland, S & Lash, TL 2008, 'Case–Control Studies', in Encyclopedia
of Quantitative Risk Analysis and Assessment, John Wiley & Sons, Ltd.
Rudolph, JL, Harrington, MB, Lucatorto, MA, Chester, JG, Francis, J & Shay, KJ 2011,
'Validation of a Medical Record Based Delirium Risk Assessment', Journal of the
American Geriatrics Society, pp. S289 94.
Rudolph, JL, Inouye, SK, Jones, RN, Yang, FM, Fong, TG, Levkoff, SE & Marcantonio,
ER 2010, 'Delirium: An Independent Predictor of Functional Decline After Cardiac
Surgery', Journal of the American Geriatrics Society, vol. 58, no. 4, pp. 643 9.
217
Sanders, RD 2011, 'Hypothesis for the pathophysiology of delirium: role of baseline
brain network connectivity and changes in inhibitory tone',Medical Hypotheses, vol.
77, no. 1, pp. 140 3.
Santos, FS, Wahlund, LO, Varli, F, Velasco, IT & M.E, J 2005, 'Incidence, Clinical
Features and Subtypes of Delirium in Elderly Patients Treated for Hip Fractures',
Dementia and Geriatric Cognitive Disorders, vol. 20, pp. 231 7.
Saxena, S & Lawley, D 2009, 'Delirium in the elderly: a clinical review', Postgraduate
Medical Journal, vol. 85, no. 1006, pp. 405 13.
Schofield, I 2008, 'Delirium: challenges for clinical governance', Journal of Nursing
Management, vol. 16, no. 2, pp. 127 33.
Schofield, I & Hasemann, W 2011, 'Delirium', in D Tolson, J Booth & I Schofield (eds),
Evidence Informed Nursing with Older People, Blackwell Publishing Ltd, Oxford.
Schor, JD, Levkoff, SE, Lipsitz, LA, Reilly, CH, Cleary, PD, Rowe, JW & Evans, DA 1992,
'Risk factors for delirium in hospitalized elderly', JAMA: The Journal of the American
Medical Association, vol. 267, no. 6, pp. 827 31.
Schreier, AM 2010, 'Nursing care, delirium, and pain management for the
hospitalized older adult', Pain Management Nursing, vol. 11, no. 3, pp. 177 85.
Schuurmans, MJ, Shortridge Baggett, LM & Duursma, SA 2003, 'The Delirium
Observation Screening Scale: a screening instrument for delirium', Research &
Theory for Nursing Practice, vol. 17, no. 1, pp. 31 50.
Schulz, KF & Grimes, DA 2002, 'Case control studies: research in reverse', The
Lancet, vol. 359, no. 9304, pp. 431 4.
218
Seitz, DP, Gill, SS & Zyl, LTv 2007, 'Antipsychotics in the treatment of delirium: a
systematic review', Journal of Clinical Psychiatry, vol. 68, no. 1, pp. 11 21.
Sendelbach, S & Guthrie, PF 2009, 'Acute Confusion/Delirium: Identification,
Assessment, Treatment, and Prevention', Journal of Gerontological Nursing, vol. 35,
no. 11, pp. 11 8.
Sheng, AZ, Shen, Q, Cordato, D, Zhang, YY & Chan, DKY 2006, 'Delirium within Three
Days of Stroke in a Cohort of Elderly Patients', Journal of the American Geriatrics
Society, vol. 54, no. 8, pp. 1192 8.
Siddiqi, N, House, AO & Holmes, JD 2006, 'Occurrence and outcome of delirium in
medical in patients: a systematic literature review', Age and Ageing, vol. 35, no. 4,
pp. 350 64.
Simone, MJ & Tan, Z 2011, 'The role of inflammation in the pathogenesis of delirium
and dementia in older adults: a review', CNS Neuroscience & Therapeutics, vol. 17,
no. 5, pp. 506 13.
Ski, C & O'Connell, B 2006, 'Mismanagement of delirium places patients at risk', The
Australian Journal Of Advanced Nursing, vol. 23, no. 3, pp. 42 6.
Speed, G, Wynaden, D, McGowan, S, Hare, M & Landsborough, I 2007, 'Prevalence
rate of delirium at two hospitals in Western Australia', Australian Journal of
Advanced Nursing, vol. 25, no. 1, pp. 38 43.
Steis, M & Fick, D 2008, 'Are nurses recognizing delirium? A systematic review',
Journal of Gerontological Nursing, vol. 34, no. 9, pp. 40 8.
219
Tabachnick, BG, & Fidell, L. S. 2012, Using Multivariate Statistics (6th Edition).
Boston: Pearson Education.
Tabet, N & Howard, R 2009, 'Pharmacological treatment for the prevention of
delirium: review of current evidence', International Journal of Geriatric Psychiatry,
vol. 24, no. 10, pp. 1037 44.
Tabet, N, Hudson, S, Sweeney, V, Sauer, J, Bryant, C, Macdonald, A & Howard, R
2005, 'An educational intervention can prevent delirium on acute medical wards',
Age and Ageing, vol. 34, no. 2, pp. 152 6.
Travers, C, Byrne, G, Pachana, N, Klein, K & Gray, L 2013, 'Prospective observational
study of dementia and delirium in the acute hospital setting', Internal Medicine
Journal, vol. 43, no. 3, pp. 262 9.
Treloar, A 1998, 'Delirium: prevalence, prognosis and management', Reviews in
Clinical Gerontology, vol. 8, no. 3, pp. 241 9.
Tropea, J, Slee, J, Brand, C, A, Gray, L & Snell, T 2008, 'Policy and Practice update:
Clinical practice guidelines for the management of Delirium in older people in
Australia', Australasian Journal on Ageing, vol. 27, no. 3, pp. 150 6.
Tropea, J, Slee, J, Holmes, ACN, Gorelik, A & Brand, CA 2009, 'Use of antipsychotic
medications for the management of delirium: an audit of current practice in the
acute care setting', International Psychogeriatrics, vol. 21, no. 01, pp. 172 9.
Tune, L 2000, 'Serum anticholinergic activity levels and delirium in the elderly',
Seminars In Clinical Neuropsychiatry, vol. 5, no. 2, pp. 149 53.
220
Tune, L & DeWitt, M 2011, 'Delirium', in CE Coffey & JL Cummings (eds), The
American Psychiatric Publishing Textbook of Geriatric Neuropsychiatry, 3rd edn, The
American Psychiatric Publishing, Inc., Arlington.
Twedell, D 2005, 'Delirium', The Journal of Continuing Education in Nursing, vol. 36,
no. 3, pp. 102 3.
van Munster, BC 2009, 'Pathophysiological studies in delirium: a focus on genetics'.
PhD thesis, University of Amsterdam, retrieved 10th February 2012, ProQuest
Dissertation and Theses database.
Vardi, K & Harrington, CJ 2014, 'Delirium: treatment and prevention (part 2)', Rhode
Island Medical Journal (2013), vol. 97, no. 6, pp. 24 8.
Vassallo, M, Vignaraja, R, Sharma, JC, Briggs, R & Allen, SC 2004, 'Predictors for falls
among hospital inpatients with impaired mobility', Journal of the Royal Society of
Medicine, vol. 97, no. 6, pp. 266 9.
Victorian Department of Health 2013 ‘Melbourne Metropolitan Hospitals & Health
Services Locations’, Department of Health, retrieved 15th July 2013,
<https://www.healthcollect.vic.gov.au/HospitalLists/MetroHospitalList.aspx>
Vidán, MT, Sánchez, E, Alonso, M, Montero, B, Ortiz, J & Serra, JA 2009, 'An
intervention integrated into daily clinical practice reduces the incidence of delirium
during hospitalization in elderly patients', Journal of the American Geriatrics Society,
vol. 57, no. 11, pp. 2029 36.
Villalpando Berumen, JM, Pineda Colorado, AM, Palacios, P, Reyes Guerrero, J,
Villa, AR & Gutiérrez Robledo, LM 2003, 'Incidence of Delirium, Risk Factors, and
221
Long Term Survival of Elderly Patients Hospitalized in a Medical Specialty Teaching
Hospital in Mexico City', International Psychogeriatrics, vol. 15, no. 04, pp. 325 36.
Voyer, P, Cole, MG, McCusker, J & Belzile, E 2006, 'Prevalence and Symptoms of
Delirium Superimposed on Dementia', Clinical Nursing Research, vol. 15, no. 46, pp.
46 65.
Voyer, P, Cole, MG, McCusker, J, St Jacques, S & Laplante, J 2008, 'Accuracy of nurse
documentation of delirium symptoms in medical charts', International Journal of
Nursing Practice, vol. 14, no. 2, pp. 165 77.
Voyer, P, McCusker, J, Cole, MG, St Jacques, S & Lioudmila, K 2007, 'Factors
associated with delirium severity among older patients ', Journal of Clinical Nursing,
vol. 16, pp. 819 31.
Wakefield, BJ 1996, ‘Prevalence, incidence, risk factors and short term outcomes for
hospitalized elderly patients experiencing acute confusion’, PhD thesis, University of
Iowa, retrieved 17th May 2012, ProQuest Dissertations and Theses database.
Wakefield, BJ 2002, 'Risk for acute confusion on hospital admission', Clinical Nursing
Research, vol. 11, no. 2, pp. 153 72.
Wang, E, Mabasa, V, Loh, G & Ensom, M 2012, 'Haloperidol Dosing Strategies in the
Treatment of Delirium in the Critically Ill', Neurocritical Care, vol. 16, no. 1, pp. 170
83.
Wanich, CK, Sullivan Marx, EM, Gottlieb, GL & Johnson, JC 1992, 'Functional status
outcomes of a nursing intervention in hospitalized elderly', Image: Journal of
Nursing Scholarship, vol. 24, no. 3, pp. 201 7.
222
Watson, R, Brand, C, Watson, R & LoGiudice, D 2009, 'Delirium in the elderly. A
survey of environmental policies and procedures in Melbourne hospitals', Australian
Health Review, vol. 33, no. 4, pp. 656 62.
Wei, LA, Fearing, MA, Sternberg, EJ & Inouye, SK 2008, 'The Confusion Assessment
Method: A Systematic Review of current usage', Journal of the American Geriatrics
Society, vol. 56, no. 5, pp. 823 30.
White, S 2002, 'The neuropathogenesis of delirium', Reviews in Clinical Gerontology,
vol. 12, no. 01, pp. 62 7.
Wilson, K, Broadhurst, C, Diver, M, Jackson, M & Mottram, P 2005, 'Plasma insulin
growth factor 1 and incident delirium in older people', International Journal Of
Geriatric Psychiatry, vol. 20, no. 2, pp. 154 9.
Witlox, J, Eurelings, L, de Jonghe, J, Kalisvaart, K, Eikelenboom, P & van Gool, W
2010, 'Delirium in Elderly Patients and the Risk of Post discharge Mortality,
Institutionalization, and Dementia', JAMA: The Journal of the American Medical
Association, vol. 304, no. 4, pp. 443 51.
Wong, CL, Holroyd Leduc, J, Simel, DL & Straus, SE 2010, 'Does this Patient have
Delirium? Value of bedside Instruments ', JAMA: The Journal of the American
Medical Association, vol. 304, no. 7, pp. 779 86.
Woodward, M, Elliott, R & Oborne, CA 2003, 'Quality of prescribing for elderly
inpatients at nine hospitals in Victoria, Australia'. Journal of Pharmacy Practice and
Research, vol. 33, no. 2, pp. 101 5.
223
World Health Organisation 1992, The ICD 10 classification of mental and
behavioural disorders clinical descriptions and diagnostic guidelines, WHO,
Geneva.
Young, J & Inouye, SK 2007, 'Delirium in older people', BMJ, vol. 334, no. 7598, pp.
842 6.
Young, J, Leentjens, AF, George, J, Olofsson, B & Gustafson, Y 2008, 'Systematic
approaches to the prevention and management of patients with delirium', Journal
of Psychosomatic Research, vol. 65, no. 3, pp. 267 72.
Young, LJ & George, J 2003, 'Do guidelines improve the process and outcomes of
care in delirium?', Age and Ageing, vol. 32, no. 5, pp. 525 8.
224
Appendices
225
Appendix 1 – Systematic review protocol
226
227
228
229
230
231
Appendix 2 – Search strategy for systematic review
Cinahl
Search Search Parameters Results
S1 AB ("medical ward*" OR "hospital inpatient*" OR "medical
inpatient*" OR "general medical inpatient*" OR "medical
admission*" OR hospitali?ation OR medical ) OR TI (
"medical ward*" OR "hospital inpatient*" OR "medical
inpatient*" OR "general medical inpatient*" OR "medical
admission*" OR hospitali?ation OR medical )
131139
S2 MH Delirium 2151
S3 MH Confusion 1036
S4 AB ("acute confus*" OR confu*) OR TI ("acute confus*" OR
confu* )
7391
S5 S2 or S3 or S4 9658
S6 MH risk factors 53046
S7 MH Fever OR "urinary tract infection" OR dementia OR
depression OR dehydration OR infection OR stroke OR "hip
fractures"
89851
S8 TI (Fever OR "urinary tract infection" OR dementia OR
depression OR dehydration OR infection OR stroke OR "hip
fractures" OR "cognitive impairment" OR "hearing impair*"
OR "visual impair*" OR "precipitat* factor*" OR "predispos*
factor*" ) OR AB ( Fever OR "urinary tract infection" OR
dementia OR depression OR dehydration OR infection OR
stroke OR "hip fractures" OR "cognitive impairment" OR
"hearing impair*" OR "visual impair*" OR "precipitat*
factor*" OR "predispos* factor*" )
139791
S9 S6 or S7 or S8 217533
232
S10 S1 and S5 and S9 368
S11 S1 and S5 and S9 Limiters Published Date from: 19960101
20121231
336
Medline
Search Search Parameters Results
S1 AB ( "medical ward*" OR "hospital inpatient*" OR "medical
inpatient*" OR "general medical inpatient*" OR "medical
admission*" OR hospitali?ation OR medical ) OR TI ( "medical
ward*" OR "hospital inpatient*" OR "medical inpatient*" OR
"general medical inpatient*" OR "medical admission*" OR
hospitali?ation OR medical )
714583
S2 MH Delirium 4875
S3 MH Confusion 3398
S4 AB ( "acute confus*" OR confu* ) OR TI ( "acute confus*" OR
confu* )
37291
S5 S2 or S3 or S4 43026
S6 MH risk factors 482109
S7 MH Fever OR "urinary tract infection" OR dementia OR “mild
cognitive impairment” OR depression OR dehydration OR
infection OR stroke OR "hip fractures"
216522
S8 TI ( Fever OR "urinary tract infection" OR dementia OR
depression OR dehydration OR infection OR stroke OR "hip
fractures" OR "cognitive impairment" OR "hearing impair*"
OR "visual impair*" ) OR AB ( Fever OR "urinary tract
infection" OR dementia OR depression OR dehydration OR
infection OR stroke OR "hip fractures" OR "cognitive
impairment" OR "hearing impair*" OR "visual impair*" )
1136263
233
S9 S6 or S7 or S8 1617737
S10 S1 and S5 and S9 1291
S11 S1 and S5 and S9 Limiters Published Date from: 19960101
20121231
967
PsycInfo
Search Search Parameters Results
S1 MJ medical patients 3197
S2 TI ( "medical ward*" OR "hospital inpatient*" OR "medical
inpatient*" OR "general medical inpatient*" OR "medical
admission*" OR hospitali?ation OR medical ) OR AB ( "medical
ward*" OR "hospital inpatient*" OR "medical inpatient*" OR
"general medical inpatient*" OR "medical admission*" OR
hospitali?ation OR medical )
130956
S3 S1 or S2 132678
S4 MJ Delirium 1946
S5 TI ( "acute confus*" OR confu* ) OR AB ( "acute confus*" OR
confu* )
22747
S6 S4 or S5 24327
S7 MJ Risk factors 23836
S8 MJ Dementia 21619
S9 MJ pneumonia 214
S10 TI (Fever OR "urinary tract infection" OR dementia OR
depression OR dehydration OR infection OR stroke OR "hip
fractures" OR "cognitive impairment" OR "hearing impair*"
OR "visual impair*" OR "precipitat* factor*" OR "predispos*
factor*") OR AB (Fever OR "urinary tract infection" OR
231475
234
dementia OR depression OR dehydration OR infection OR
stroke OR "hip fractures" OR "cognitive impairment" OR
"hearing impair*" OR "visual impair*" OR "precipitat* factor*"
OR "predispos* factor*" )
S11 S7 or S8 or S9 or S10 250202
S12 S3 and S6 and S11 447
S13 S3 and S6 and S11 Limiters Published Date from: 19960101
20121231
350
Proquest Health and Medical
Search Search Parameters Results
S1 ab("medical ward*" OR "hospital inpatient*" OR "medical
inpatient*" OR "general medical inpatient*" OR "medical
admission*" OR hospitali?ation OR medical ) OR ti("medical
ward*" OR "hospital inpatient*" OR "medical inpatient*" OR
"general medical inpatient*" OR "medical admission*" OR
hospitali?ation OR medical )
729540
S2 mesh(delirium) 1001
S3 ab("acute confus*" OR confu* ) OR ti("acute confus*" OR
confu* )
13818
S4 S2 OR S3 14668
S5 mesh(risk factors) 72469
S6 mesh(dementia) 5333
S7 ab(Fever OR "urinary tract infection" OR dementia OR
depression OR dehydration OR infection OR stroke OR "hip
fractures" OR "cognitive impairment" OR "hearing impair*"
OR "visual impair*" OR "precipitat* factor*" OR "predispos*
556423
235
factor*") OR ti(Fever OR "urinary tract infection" OR
dementia OR depression OR dehydration OR infection OR
stroke OR "hip fractures" OR "cognitive impairment" OR
"hearing impair*" OR "visual impair*" OR "precipitat*
factor*" OR "predispos* factor*")
S8 S5 OR S6 OR S7 614729
S9 S1 AND S4 AND S8 417
S10 S1 AND S4 AND S8 Limiters Published Date from:
19960101 20121231
400
Informit Health Collection
Search Search Parameters Results
#1 AB = ( "medical ward*" OR "hospital inpatient*" OR
"medical inpatient*" OR "general medical inpatient*" OR
"medical admission*" OR hospitali?ation OR medical ) OR TI
= ( "medical ward*" OR "hospital inpatient*" OR "medical
inpatient*" OR "general medical inpatient*" OR "medical
admission*" OR hospitali?ation OR medical )
3721
#2 SUBJECT=(Delirium) 29
#3 SUBJECT=(confusion) 0
#4 TI=("acute confus*" OR confu*) OR AB=("acute confus*" OR
confu*)
229
#5 #2 OR #3 OR #4 252
#6 SUBJECT=(risk factors) 480
#7 TI = (Fever OR "urinary tract infection" OR dementia OR
depression OR dehydration OR infection OR stroke OR "hip
fractures" OR "cognitive impairment" OR "hearing impair*"
2233
236
OR "visual impair*" OR "precipitat* factor*" OR "predispos*
factor*") OR AB = (Fever OR "urinary tract infection" OR
dementia OR depression OR dehydration OR infection OR
stroke OR "hip fractures" OR "cognitive impairment" OR
"hearing impair*" OR "visual impair*" OR "precipitat*
factor*" OR "predispos* factor*")
#8 #6 OR #7 2643
#9 #1 AND #5 AND #8 3
#10 #1 AND #5 AND #8 Limiters Published Date from:
19960101 20121231
3
Embase
Search Search Parameters Results
#1 'medical ward':ti OR 'hospital inpatient':ti OR 'medical
inpatient':ti OR 'general medical inpatient':ti OR 'medical
admission':ti OR hospitalisation:ti OR hospitalization:ti OR
medical:ti OR 'medical ward':ab OR 'hospital inpatient':ab
OR 'medical inpatient':ab OR 'general medical inpatient':ab
OR 'medical admission':ab OR hospitalisation:ab OR
hospitalization:ab OR medical:ab AND [embase]/lim
684,877
#2 'delirium'/exp 14,582
#3 'confusion'/de 18,613
#4 'acute confusional state':ab OR confusion:ab OR 'acute
confusion':ab OR 'acute confusional state':ti OR
confusion:ti OR 'acute confusion':ti
26,663
#5 #2 OR #3 OR #4 53,041
#6 'risk factor'/exp OR 'risk factors'/exp 518,021
237
#7 fever:ab OR 'urinary tract infection':ab OR dementia:ab OR
depression:ab OR dehydration:ab OR infection:ab OR
stroke:ab OR 'hip fractures':ab OR 'cognitive
impairment':ab OR 'hearing impairment':ab OR 'visual
impairment':ab OR 'precipitating factor':ab OR
'predisposing factor':ab OR fever:ti OR 'urinary tract
infection':ti OR dementia:ti OR depression:ti OR
dehydration:ti OR infection:ti OR stroke:ti OR 'hip
fractures':ti OR 'cognitive impairment':ti OR 'hearing
impairment':ti OR 'visual impairment':ti OR 'precipitating
factor':ti OR 'predisposing factor':ti
1,423,844
#8 #6 OR #7 1,851,466
#9 #1 AND #5 AND #8 2,001
#10 #9 AND [1996 2012]/py 1,760
Scopus
Search Search Parameters Results
#1 (ABS("medical ward*" OR "hospital inpatient*" OR "medical
inpatient*" OR "general medical inpatient*" OR "medical
admission*" OR hospitali?ation OR medical) OR
TITLE("medical ward*" OR "hospital inpatient*" OR "medical
inpatient*" OR "general medical inpatient*" OR "medical
admission*" OR hospitali?ation OR medical))
1,046,364
#2 (KEY(delirium) OR KEY(confusion)) 41,129
#3 (TITLE("acute confus*" OR confu*) OR ABS("acute confus*"
OR confu*))
77,617
#4 #2 OR #3 111,402
#5 KEY("Risk factor*") 687,933
238
#6 (TITLE(fever OR "urinary tract infection" OR dementia OR
depression OR dehydration OR infection OR stroke OR "hip
fractures" OR "cognitive impairment" OR "hearing impair*"
OR "visual impair*" OR "precipitat* factor*" OR "predispos*
factor*") OR ABS(fever OR "urinary tract infection" OR
dementia OR depression OR dehydration OR infection OR
stroke OR "hip fractures" OR "cognitive impairment" OR
"hearing impair*" OR "visual impair*" OR "precipitat*
factor*" OR "predispos* factor*"))
1,705,643
#7 #5 OR #6 2,264,134
#8 #1 AND #4 AND #7 2,750
#9 #1 AND #4 AND #7 Limiters Published Date from:
19960101 20121231
2,173
Proquest Dissertation and Theses
Search Search Parameters Results
S1 ab("medical ward*" OR "hospital inpatient*" OR "medical
inpatient*" OR "general medical inpatient*" OR "medical
admission*" OR hospitali?ation OR medical ) OR ti("medical
ward*" OR "hospital inpatient*" OR "medical inpatient*" OR
"general medical inpatient*" OR "medical admission*" OR
hospitali?ation OR medical )
43093
S2 su(Delirium) 40
S3 ab("acute confus*" OR confu* ) OR ti("acute confus*" OR
confu* )
12042
S4 S2 OR S3 12069
S5 su("risk factor*") 1753
S6 su(dementia) 804
239
S7 ab(Fever OR "urinary tract infection" OR dementia OR
depression OR dehydration OR infection OR stroke OR "hip
fractures" OR "cognitive impairment" OR "hearing impair*"
OR "visual impair*" OR "precipitat* factor*" OR "predispos*
factor*") OR ti(Fever OR "urinary tract infection" OR
dementia OR depression OR dehydration OR infection OR
stroke OR "hip fractures" OR "cognitive impairment" OR
"hearing impair*" OR "visual impair*" OR "precipitat*
factor*" OR "predispos* factor*")
70942
S8 S5 OR S6 OR S7 72377
S9 S1 AND S4 AND S8 59
S10 S1 AND S4 AND S8 Limiters Published Date from:
19960101 20121231
33
Mednar
Search Search Parameters Results
S1 Delirium 1932
S2 S1 refined by Topic: “risk factors” 40
S1 refined by Topic: “risk factors” and Limiters Published
Date from: 19960101 20121231
34
Cochrane Library
Search Search Parameters Results
S1 Delirium 18
Similar systematic reviews on risk factors for delirium were
not identified
0
240
JBI Library of Systematic Reviews
Search Search Parameters Results
S1 Delirium 2
1 Similar systematic review on risk factors for delirium was
identified, however is relevant to ICU delirium.
1
241
Appendix 3 – Joanna Briggs Institute Critical Appraisal Instrument
242
Appendix 4 – Joanna Briggs Institute Data Extraction Tool
Page 1
243
Page 2
244
Appendix 5 – Case control study research questions, hypothesis and statistical tests
Medical records Audit case control study
Question Variables Hypothesis Statistical test
Is there a relationship between age anddevelopment of delirium?
Age – continuous: age in years (mean ageof groups)Delirium diagnosis – categorical: yes orno
Advanced age is related to the developmentof delirium.
Independent samples t tests
Is there a relationship between gender anddevelopment of delirium?
Gender – categorical: male/femaleDelirium diagnosis – categorical: yes/no
Gender will not influence deliriumdevelopment
Chi square for independence
Is there a relationship between reason foradmission and delirium diagnosis?
Reason for admission categorical: e.g.Fall, respiratory infectionDelirium Diagnosis – categorical: yes/no
There may be a relationship between somereasons for admission but not others.
Chi square for independence
Is there a relationship between past historyand delirium diagnosis?
Past History categorical: e.g. Dementia,Hypertension: yes/noDelirium Diagnosis – categorical: yes/no
There may be a relationship between somepast history and not others.
Chi square for independence
Is there a relationship between recentadmission to hospital and deliriumdiagnosis?
Recent admission to hospital –categorical: yes/noDelirium Diagnosis – categorical: yes/no/no
Patients who developed delirium are morelikely to have had a recent previous admissionto hospital.
Chi square for independence
Were patients with delirium more likely tocome from home alone or with others?
Delirium diagnosisResidence prior to admission
Patients with delirium are more likely tocome from home alone.
Description: Frequencies ofresidence
Were patients with delirium more likely tohave functional problems?
Delirium diagnosisLevel of Function prior to admission
Patients with delirium are more likely to haveprevious functional problems beforeadmission.
Description: Frequencies offunctional ability
What words were used most often todescribe delirium
Words used to describe delirium Description: Frequencies ofwords used.
245
Who was most likely to recognise anddocument the symptoms of delirium?
Who first recognised delirium symptoms? Nurses are the most likely to document thefirst symptoms of delirium
Description: Frequencies of whodocumented delirium
What were the first symptoms of delirium Words used to describe delirium Confusion is the most frequently documentedfirst symptom of delirium
Description: Frequencies ofreported symptoms of delirium
What was the patients reported cognitionon admission
Patients reported cognition on admission Patients with delirium are more likely to bedescribed as having a cognitive issue onadmission.
Description: Frequencies ofreported cognition on admission
Is there a relationship between delirium anda cognitive issue?(Anyone not described as no cognitiveissues on admission)
Cognitive Issue categorical: e.g. yes/noDelirium Diagnosis – categorical: yes/no
Delirium will be significantly related to if thepatients had a prior cognitive issue asdescribed.
Chi square for independence
Is there a relationship between delirium anddementia?
Dementia categorical: e.g. yes/noDelirium Diagnosis – categorical: yes/no
There is a strong relationship betweendelirium and dementia
Chi square for independence
Is there a relationship between use of anIDC during admission and development ofdelirium?
Use of IDC categorical: e.g. yes/noDelirium Diagnosis – categorical: yes/no
Use of IDC will be related to the developmentof delirium
Chi square for independence
Is there a relationship between use ofrestraint during admission and developmentof delirium?
Use of restraints categorical: e.g. yes/noDelirium Diagnosis – categorical: yes/no
Use of restraints will be related to thedevelopment of delirium
Chi square for independence
Is there a relationship between adding morethan 3 medications during admission anddevelopment of delirium?
Adding more than 3 medicationscategorical: e.g. yes/noDelirium Diagnosis – categorical: yes/no
Adding more than 3 medications duringadmission will be related to the developmentof delirium.
Chi square for independence
Is there a relationship between Sodiumlevel on admission and development ofdelirium?
Sodium – continuous: Sodium asmeasured on admission (mmol/L)Delirium diagnosis – categorical: yes orno (only 2 groups)
Patients who developed delirium will be morelikely to have an abnormal sodium level.
Independent samples t tests
What predisposing factors predict thelikelihood that a patient will developdelirium during admission?
Dependent Variable: Delirium –categorical Yes/No
Logistic regression
246
Independent variables Categorical:Dementia, cognitive impairment,functional impairment, visionimpairment.
What precipitating factors predict thelikelihood that a patient will developdelirium during admission?
Dependent Variable: Delirium –categorical Yes/NoIndependent variables:Categorical Use of IDC, Use of Restraints,and Given Benzodiazepine duringadmission.Continuous variable: Sodium onadmission and Sodium Day 3 ofadmission
Logistic regression
Were patients with a delirium more likely tohave a code grey called?
Delirium diagnosis Yes/NoCode grey called Yes /No
Patients who developed delirium will be morelikely to have a code grey called.
Description: Frequencies of codegreys called.
Is there a relationship between delirium andnumber of falls prior to admission?
Delirium: Categorical YES/ NoNumber of falls: Categorical Less than 2,more than 2
Chi square for independence
Did the patients with delirium have morefalls during admission?
Delirium: Categorical YES/ NoNumber of falls during admission
Patients with delirium more likely to have afall during admission.
Description: Frequencies of fallsduring admission
Were patients with delirium more likely todevelop a pressure injury during admission?
Delirium: Categorical YES/ NoNumber of pressure injuries duringadmission
Patients with delirium more likely to developa pressure injury during admission.
Description: Frequencies ofpressure injuries duringadmission
Were patients with delirium more likely tohave a change in functional status?
Delirium: Categorical YES/ NoChange in functional status: categoricalYES / NO
Patients with delirium were more likely tohave a change in their functional status
Chi square for independence
Were patients with delirium more likely tohave a change in continence status?
Delirium: Categorical YES/ NoChange in continence status: categoricalYES / NO
Patients with delirium were more likely tohave a change in their continence status
Chi square for independence
247
Were patients that developed deliriummore likely to have a met call duringadmission?
Delirium: Categorical YES/ NoMet call during admission: categoricalYES / NO
Patients with delirium were more likely tohave a met call during admission
Chi square for independence
Was the Length of stay for patients thatdeveloped delirium longer than those thatdid not?
Length of stay (days) ContinuousDelirium diagnosis – categorical: yes orno
Patients with delirium more likely to havelonger length of stay
Independent samples t tests
Were patients with delirium more likely topass away during admission?
Delirium: Categorical YES/ NoPassed away during admission:Categorical Yes/ No
Patients with delirium were more likely topass away during admission
Chi square for independence
Were patients with delirium more likely topass away since admission?
Delirium: Categorical YES/ NPassed away since admission: CategoricalYes/ No
Patients with delirium more likely to havepassed away since admission
Chi square for independence
Were patients with delirium more likely tobe discharged to care facility?
Delirium: categorical YES / NoDischarged to a care facility: CategoricalYes/No
Patients with delirium more likely to bedischarged to a care facility
Chi square for independence
248
Appendix 6 – Case control study audit tool
Clinical Records Audit toolDate of audit _____/_____/_____ Initials __________Date of admission _____/_____/_____ Day__________________Admission unit __________________________________________
A. Demographic detailsA1. Primary diagnosis on admission (describe)
A2. Number of Co morbidities_________ list
A3. Age at admission A4. Gender M FA5. Usual place of residence on admissionHome alone Home with spouse/carerLow level residential care High level residential care
B. Diagnosis and Detection of Delirium
ClinicalGuidelinenumber
B1. Date confusion/increasedconfusion first noted_____/_____/____
B2 Date delirium first noted_____/_____/_____
B3. Was person admitted with a diagnosis of delirium?Documented in initial ED or unit assessment notesYES NO unclearadmitted with ‘confusion’ or ‘increased confusion’B4. Was a diagnosis of ‘delirium’ ever documented?YES NOB5. Was a medical cause for the delirium ever documented?YES NO unclearB8.Was a tool used to assist in the diagnosis of delirium? YES NO If yes,what tool _____________
1.6.1
B9.What words were used to describe delirium? (please list)
C. Cognitive Impairment screeningC1.Was a cognitive assessment test performed on admission?YES NOIf yes, what tool _________
1.5.1
C2.Was the score of the cognitive assessment documented in history?YES NO NA
UR Number
249
If yes, what was the score?
C3. Is there any evidence from the medical record that a family member wasconsulted in regard to cognitive status?YES NOC4. Did the patient experience a sudden change in cognitive status duringadmission?YES NOIf yes, was a cognitive assessment performed?YES NO
1.7
D. Risk factors for delirium: assessment and prediction 2.1
D1. Did the patient have documentation of any of the following risk factors fordelirium? Tick relevant factorsPredisposing factors Precipitating factorsPre existing dementia orcognitive impairment Abnormal serum sodium
Depression Use of an IDC duringadmission
Visual impairment Use of physical restraintsPrevious episode of delirium Severe medical illness
Hearing impairment Adding >3 medications duringhospitalisation
Functional impairment(dependence in 2 ADLs) Other:
D2. Is there evidence of a risk factor assessment of delirium in older persons?YES NO
If Yes, describe_____________
2.4
E. Prevention of Delirium 3.1
E1. Is there any evidence of the following environmental prevention strategiesbeing incorporated into the care plan?
3.1
Provision of a single roomProvision of clock and calendarAvoid room changesQuiet environmentEncourage family involvementE2. Is there any evidence of the following clinical practice prevention strategiesbeing incorporated into care?
3.1
Encouragement of food and fluidintakeRegulation of bowel function
250
Ensuring patients wear hearing aids
Encourage regular mobilisation
Encourage independenceMedication reviewPain managementF. Management of DeliriumF1. Is there evidence in the medical records that the underlying cause ofdelirium was investigated and treated?
4.2
Examination – is there evidence of the following:Obtain vital signs Chest (Auscultation, cough)Mental state examination(decreased arousal orattention)
Abdomen(palpation of bladder and bowel)
Neurological examination Skin (signs of dehydration)Investigations – is there evidence the following was undertaken?Urinalysis or MSU Chest x rayFull Blood Examination Cardiac enzymesUrea and Electrolytes ECGGlucose Liver function testsF2. Is there evidence of the use of non pharmalogical strategies to managedelirium symptoms?
4.3
Use of a support person or one on one nurse(who has been trained in how to care for people with delirium)
Allowing family members to stay with patientProviding relaxation strategies to assist with sleepModification of environment to minimise risk of injuryF.3Were pharmalogical interventions used to treat delirium symptoms?YES NO go to question G.1
4.3.2
F.4What medication was prescribed for the management of severe behaviouraldisturbances?Haloperidoldose____freq_____
Olanzapinedose_______ freq_______
Risperidonedose___ freq____
Quetiapinedose_______ freq_______
Droperidoldose_______ freq__
Other name_________dose_______ freq_______
F.5Was the indication for its use documented? YES NO 4.3.5F.6. Was the commencement dose low?(0.25mg haloperidol; 2.5mg olanzapine; 0.25mg risperidone)YES NO
251
F.8. Number of times a stat dose of antipsychotic medication was given andwhen
4.3.5
Date Time given DateTimegiven
Haloperidol Olanzapine
Risperidone Quetiapine
DroperidolOthername_________
F.9. Was antipsychotic medication reviewed at any stage of admission? YESNO
4.3.5
F.10 If the patient was severely agitated were they reviewed by medical staff4hrly? YES NOIf the patient had less significant agitation were they reviewed by medical staff8hrly? YES NOG. Discharge planning and follow upG.1 Is there evidence that information was provided to the patients and theirfamilies regarding delirium?YES NO
4.4
G2.Was follow up care including professional monitoring of deliriumimplemented on discharge?YES NO
H. Medications on admissionH1. Admitted on antipsychotic medication? YES Document dose, frequency, &how long pt has been taking (if known)NO go to question C4Haloperidoldose___ freq___ duration__________
Olanzapinedose__ freq______ duration________
Risperidonedose_______ freq_______ duration______
Quetiapinedose_______ freq_____duration_____
Droperidoldose_______ freq____duration________H2. Were any of the antipsychotic medications ticked above ceased during the episodeof care?
YES Ceased without adequate documentation of reason (does not includeintermittent administration of medication)
YES Ceased as part of documented management planNO administered as per preadmission
If YES note date this first occurred_____/_____/_____
252
H3. Was dosage/frequency of the antipsychotic medications ticked above altered duringthe episode of care?
YES Modified without adequate documentation of reason (includes intermittentadministration of medication)
YES Modified as part of documented management planNO administered as per preadmission
If YES note date this first occurred_____/_____/_____H4. Admitted on benzodiazepine medication?YES Document dose, frequency, & how long pt has been taking (if known)NO go to Section FDiazepamdose_______ freq_____ duration________
Temazepamdose_____ freq_______duration_____
Nitrazepamdose_______ freq____ duration___
Oxazepamdose_______ freq_____ duration______
Lorazepamdose_______ freq_____ duration_____
Other (please list)__________________________
H5. Was any of the benzodiazepine medications ticked above ceased during the episodeof care?
YES Ceased without adequate documentation of reason (does not includeintermittent administration of medication)
YES Ceased as part of documented management planNO administered as per preadmission
If YES note date this first occurred_____/_____/_____H6. Was dosage/frequency of the benzodiazepine medications ticked above alteredduring the episode of care?
YES Modified without adequate documentation of reason (includes intermittentadministration of medication)
YES Modified as part of documented management planNO administered as per preadmission
If YES note date this first occurred_____/_____/_____
H7. Prescription of benzodiazepine medications during hospitalisationH8Were any benzodiazepine medications newly prescribed during admission? YESNO go to Section G
H9 Date first prescribed _____/_____/_____
H10. Benzodiazepines newly prescribed during admissionDiazepam TemazepamNitrazepam OxazepamLorazepam Other (please list) __________________________
253
Other delirium management issues
1. Use of specialling YES NO Number of days __________2. Use of mechanical restraint YES NOWhich type of restraint?Bed rail/cot side Table attached to chairWrist restraints Ankle restraintsOther specify_____________________
3. Code grey YES NO Number of code grey episodes4. Adverse eventsFall Number _______ Harm from fall? ______ Skin tear Number _______Pressure area Number _______ Other specify
5. Unit from which patient was discharged ________________________________Discharge destination _______________________________________Length of stay ___________(days)
254
Appendix 7 Case control study audit tool on iPad application Tap
Forms
Opening a new record displays screen as displayed below. Arrows indicate a drop down
section that will open when pressed.
255
This screen appears when you click on demographic details.
256
Entering in Age prompts the number pad to appear.
257
The keyboard will appear if typing is required.
258
Predetermined answers that can be chosen from this side menu for some
questions.
259
260
Examples of questions for the section regarding the diagnosis and detection of
delirium.
261
Appendix 8 – Ethical approval letters
Case control study Ethics Approval
263
Deakin University ethics approval
265
Appendix 9 – Delirium management survey
Date: Position of respondent (eg NUM or DON):
Organisation:
Hospital dataHow many patients on average does your hospital treat per year?
What is the estimated number of patients that develop a delirium during admission?(% per year)
Delirium managementCircle
Is there a hospital policy for Delirium Management in your facility? Yes / NoIf your organisation includes more than one health service, pleaseindicate if this policy covers all of the included health services/hospitals.
Yes / NoNA
Are you aware of the guidelines published by the Department of Healthand Ageing in relation to delirium management “Clinical PracticeGuidelines for the Management of delirium in Older people” or the“Delirium Care Pathways”?http://www.health.gov.au/internet/main/publishing.nsf/Content/9E46460CFDAFBA03CA25732B004C4331/$File/Delirium_CPGforMODIOP_web.pdf
Yes / No
Was the Delirium management policy developed using these guidelines? Yes / NoIf yes, please advise how these guidelines were used:
Screening and assessmentIs there a policy recommendation for how to screen and diagnosedelirium?
Yes / No
If Yes, please describe the procedures or tools recommended for delirium screening ordiagnosis
If a tool is recommended, how is this documented in the patient’s medical records? (forexample is this documented on a specific form)
Cognitive assessmentIs there a policy recommendation regarding a formal cognitiveassessment for all adults over age of 65 years admitted to hospital?
Yes / No
If so, what tool is used to assist with this screening?
266
Is training provided for the use of these tools? Yes / NoIf yes, please describe the training:
Risk factorsDoes the delirium management policy recommend a delirium riskassessment in all older people admitted to the health care setting?
Yes / No
What training is provided to staff caring for older patients to increase their knowledge ofthe risk factors of delirium?
Pharmacological managementIs there a hospital policy for the pharmacological management ofdelirium or the management of aggressive behaviour and/or severeagitation?
Yes / No
What medications are recommended in this policy and what is the recommended dose?(tick)
Haloperidol Dose: Risperidone Dose:
Olanzipine Dose: Other (name) Dose:
Is there a recommended guideline for the frequency of medical review fora severely agitated patient?
Yes / No
If so, what is the recommended time frequency?
What training is provided to medical staff regarding medication management forseverely agitated or aggressive patients with delirium?
Have you experienced any barriers in relation to the implementation of protocolsrelating to delirium management? If so, please describe.
Any further comments:
Thank you for your time.
267
Appendix 10 – Email to Directors of Nursing and participants
Transcript of email to Director of Nursing
Dear Executive/Director of Nursing,
My name is Emily Cull and I am a PhD student in the School of Nursing andMidwifery at Deakin University. The PhD investigates the incidence of delirium inacute care settings. Part of my PhD project involves conducting surveys of hospitalsin Melbourne regarding the management of patients with delirium in acutehospitals. More specifically, the survey will be used to explore policies that havebeen developed by health care organisations to manage delirium and therecommendations that the policies provide.
I am writing to seek your approval for organisational consent to participate in thesurvey and to nominate a potential respondent; someone who you think would besuitable to participate in the survey. I am seeking a participant who has a goodunderstanding of policies used in the organisation for delirium management andthe content of these policies. The survey should take around 10 minutes tocomplete. Participant’s information will be included with the survey but identifiableinformation with respect to the respondent and the organisation will not be madeavailable in the reporting of the results. Attached to this email is a copy of the PlainLanguage Statement and the survey instrument that you can forward to thepotential participant. Please inform the participant that they are invited to emailthe researcher to register interest in completing the survey and to also arrange asuitable time for the survey to be completed over the phone. Alternatively pleasesend me the email address of the potential participant so that I may contact themto invite them to participate.
The ethical aspects of this research project have been approved by a HumanResearch Ethics Committee at Deakin University.
If you require further information please send me an email.
Thank you for your time
Emily CullPhD CandidateDeakin [email protected] 9244 6958
268
Transcript of email to participant
Dear Participant
My name is Emily and I am a PhD student at Deakin University. For my project, I amexamining delirium management in acute hospitals, more specifically examining atpolicies that have been developed by health care organisations, to manage deliriumand the recommendations that the policies provide. Because you have beensuggested as a representative of the organisation I am inviting you to participate inthis research study by completing the attached survey.
The following questionnaire will require approximately 10 minutes to complete.There is no compensation for responding nor is there any known risk. In order toensure that all information will remain confidential, please do not include yourname. Copies of the project will be provided to my Deakin University supervisor. Ifyou choose to participate in this project, please answer all questions as honestly aspossible and return the completed questionnaires promptly by email.
Participation is strictly voluntary and you may refuse to participate at any time.Thank you for taking the time to assist me in my educational endeavours. The datacollected will provide useful information regarding the management of delirium inacute hospitals. If you would like a summary copy of this study please email mewith a Request for Information. Completion and return of the questionnaire willindicate your willingness to participate in this study. If you require additionalinformation or have questions, please contact me at the number listed below.
If you are not satisfied with the manner in which this study is being conducted, youmay report any complaints to the Deakin University Human Research Ethics office(research [email protected])
Kind Regards
Emily CullPhD CandidateSchool of Nursing and MidwiferyDeakin [email protected] 342 043
Professor Alison HutchinsonSchool of Nursing and MidwiferyDeakin University
269
Appendix 11 Plain language statements
Director of Nursing
DEAKIN UNIVERSITY PLAIN LANGUAGE STATEMENT
Plain Language Statement
Date:Full Project Title: Incident delirium in acute medical patients
Principal Researcher: Alison Hutchinson
Student Researcher: Emily Cull
Associate Researcher(s): Nikki Philips
Dear Director of Nursing,
This letter is to invite you to participate in a research project, which examines the policies and procedures used in health-care organisations for patients with delirium. This will help to provide information regarding how delirium is managed in hospitals and the process of care that is recommended for these patients. The project is being undertaken as part of a PhD degree.
You have been selected as a potential respondent at this organisation. We are seeking a respondent who has a good understanding of policies used in the organisation and the content of these policies. Whether you agree to take part in the project is completely up to you. If you are not able to participate in the study we ask you to forward the attached plain language statement to a person who think would be suitable.
This study will involve completing a questionnaire regarding policies and procedures for delirium management. The questionnaire is anonymous and will only require stating the respondent’s position in the organisation. It should take around 5 – 10 minutes of your time to complete. Surveys are being sent to all public and private hospitals in Melbourne for potential participation. If you agree to participate, please return your questionnaire by [date]. Implied consent is obtained via the completion of the survey. Surveys can be returned by email (from a generic and non-identifiable email address), by post, or respondents can arrange to have the survey completed over the phone.
Participants are invited to contact the researchers should they wish to obtain a summary of the results.
270
The ethical aspects of this research project have been approved by a human ethics panel at Deakin University. If you have any complaints about any aspect of the project, the way it is being conducted or any questions about your rights as a research participant, then you may contact: The Manager, Office of Research Integrity, Level 1, Building EA, Deakin University, Elgar Road, Burwood Victoria 3125, Telephone: 9251 7129, Facsimile: 9244 6581; [email protected]. Please quote project number
If you require further information, wish to withdraw your participation or if you have any problems concerning this project, you can contact the research supervisor.
Thank you for your time.
271
Participant/potential respondent
DEAKIN UNIVERSITY PLAIN LANGUAGE STATEMENT
Plain Language Statement
Date:Full Project Title: Incident delirium in acute medical patients
Principal Researcher: Alison Hutchinson
Student Researcher: Emily Cull
Associate Researcher(s): Nikki Philips
Dear potential respondent,
This letter is to invite you to participate in a research project which examines the policies and procedures used in health-care organisations for patients with delirium. This will help to provide information regarding how delirium is managed in hospitals and the process of care that is recommended for these patients. The project is being undertaken as part of a PhD degree.
You have been identified as a potential respondent at this organisation by the director of nursing and is someone who has a good understanding of delirium management policies used by the organisation and the content of these policies. Whether you agree to take part in the project is completely up to you.
This study will involve completing a questionnaire regarding policies and procedures for delirium management. The questionnaire is anonymous and will only require stating your position in the organisation (e.g NUM). It should take around 5 – 10 minutes of your time to complete. Surveys are being sent to all public and private hospitals in Melbourne for potential participation. If you agree to participate, please return your questionnaire by [date]. Implied consent is obtained via the completion of the survey. You may return the survey by email (from a generic and non-identifiable email address), by post, or you can arrange to have the survey completed over the phone.
Participants are invited to contact the researchers should they wish to obtain a summary of the results.
The ethical aspects of this research project have been approved by a human ethics panel at Deakin University. If you have any complaints about any aspect of the project, the way it is being conducted or any questions about your rights as a research participant, then you may contact:
272
The Manager, Office of Research Integrity, Level 1, Building EA, Deakin University, Elgar Road, Burwood Victoria 3125, Telephone: 9251 7129, Facsimile: 9244 6581; [email protected]. Please quote project number
If you require further information, wish to withdraw your participation or if you have any problems concerning this project, you can contact the research supervisor.
Thank you for your time.
Emily Cull PhD Candidate Deakin University [email protected] 9244 6958
273
Plain language statement
DEAKIN UNIVERSITY, THE ALFRED & EASTERN HEALTHPLAIN LANGUAGE STATEMENT
Date: 19th February 2014
Full Project Title: Incident delirium in the acute general medical setting
Principal Researcher: Prof Alison Hutchinson
Student Researcher:Ms Emily Cull
Associate Researcher(s): Dr Nikki Philips
Dear potential participant,
This letter invites you to participate in a research project, which examines the policies andprocedures used in health care organisations for patients with delirium. You have beenspecifically nominated as a potential participant by the Executive/Director of Nursing inyour hospital. This research will help to provide information regarding how delirium ismanaged in hospitals and the process of care that is recommended for these patients. Theproject is being undertaken as part of a PhD degree. Whether you agree to take part in theproject is completely up to you; the Executive/Director of Nursing will not be informed ofyour involvement. If you decide to take part and later change your mind, you are free towithdraw from the project at any stage. The DON will not be made aware of your decisionto participate.
This study involves participating in a survey regarding policies and procedures for deliriummanagement. It will be conducted via telephone with the researcher, taking approximately10 minutes. A copy of the survey is attached for your information. There are no costsassociated with participating in this research project, nor will you be paid. There are nodirect benefits for you in participating in the research. However, if you choose toparticipate, you will provide important information about how patients with delirium aremanaged in health care organisations.
If you are willing to participate in the study please send an email to the researcher toarrange a suitable time for you to undertake the survey via telephone. Your consent will beobtained verbally at the beginning of the telephone interview and will be recorded using anaudio recording device. By verbally consenting to participate in the research, you areconsenting to the research team collecting information from you in regards to deliriumpolicies and procedures. No personal information will be collected from you and your name
Health care organization
logo
274
will not be kept with the survey response. Specific organisational information will not beidentifiable in the reporting of the results.
You may inform the researchers if wish to obtain a summary of the results at the conclusionof the research. It is anticipated that the results of this research project will be publishedand/or presented in a variety of forums. In any publication and/or presentation,information will be provided in such a way that you cannot be identified. The study findingswill also be presented in the form of a thesis.
The ethical aspects of this research project have been approved by a Human ResearchEthics Committee at Eastern Health, Alfred Health and Deakin University. This project willbe carried out according to the National Statement on Ethical Conduct in Human Research(2007). This statement has been developed to protect the interests of people who agree toparticipate in human research studies. If you have any complaints about any aspect of theproject, the way it is being conducted or any questions about your rights as a researchparticipant, then you may contact:
Eastern HealthChairpersonEastern Health Human Research and Ethics CommitteeEmail: [email protected] quote project number LR05 1314
Deakin UniversityThe Manager, Office of Research Integrity, Level 1, Building EA, Deakin University, ElgarRoad, Burwood Victoria 3125research [email protected] quote project number 2013 201
Alfred HealthMs Emily Bingle,Ethics & Research Governance Office
If you require further information, wish to withdraw your participation or if you have any problemsconcerning this project, you can contact the research supervisor.
Thank you for your time.
Emily CullPhD CandidateDeakin [email protected]
275
Appendix 12 – Reasons for study exclusion from systematic review
Excluded Studies Citation Reason for Exclusion
1 Adamis, D., et al., Cytokines and IGF I in deliriousand non delirious acutely ill older medical inpatients.Age And Ageing, 2009. 38(3): p. 326 32; discussion251
Presence of delirium not assessed onadmission.No differentiation of incident andprevalent delirium in reporting ofresults.
2 Brauer, C., et al., The cause of delirium in patientswith hip fracture. Archives Of Internal Medicine,2000. 160(12): p. 1856 1860.
Patients had a surgical intervention(possible confounding risk factor formedical induced incident delirium)
3 Caeiro, L., et al., Delirium in the first days of acutestroke. Journal Of Neurology, 2004. 251(2): p. 171178.
Presence of delirium not assessed onadmission.No differentiation of incident andprevalent delirium in reporting ofresults.
4 Caeiro, L., et al., Delirium in acute stroke: apreliminary study of the role of anticholinergicmedications. European Journal Of Neurology: TheOfficial Journal Of The European Federation OfNeurological Societies, 2004. 11(10): p. 699 704.
Presence of delirium not assessed onadmission.No differentiation of incident andprevalent delirium in reporting ofresults.
5 Crawley, E.J. and J. Miller, Acute confusion amonghospitalized elders in a rural hospital.MedsurgNursing: Official Journal Of The Academy OfMedical Surgical Nurses, 1998. 7(4): p. 199 206.
No differentiation of incident andprevalent delirium in reporting ofresults.Poor reporting of delirium patientcharacteristics.
6 Dahl, M.H., O.M. Rønning, and B. Thommessen,Delirium in acute stroke—Prevalence and riskfactors. Acta Neurologica Scandinavica, 2010.122(Suppl 190): p. 39 43.
Presence of delirium not assessed onadmission.No differentiation of incident andprevalent delirium in reporting ofresults.
7 de Rooij, S.E., et al., Cytokines and acute phaseresponse in delirium. Journal Of PsychosomaticResearch, 2007. 62(5): p. 521 525.
Presence of delirium not assessed onadmission.No differentiation of incident andprevalent delirium in reporting ofresults.
8 Edlund, A., et al., Delirium in older patients admittedto general internal medicine. Journal Of GeriatricPsychiatry And Neurology, 2006. 19(2): p. 83 90.
Assessing prevalent delirium(delirium present on admission tohospital)
9 Elie, M., et al., Delirium risk factors in elderlyhospitalized patients. Journal Of General InternalMedicine, 1998. 13(3): p. 204 212.
Systematic reviewIncluded surgical, medical andpsychiatric patients
10 Feldman, J., et al., Delirium in an acute geriatric unit:clinical aspects. Archives of Gerontology andGeriatrics, 1999. 28(1): p. 37 44.
Presence of delirium not assessed onadmission.No differentiation of incident andprevalent delirium in reporting ofresults.
276
Excluded Studies Citation Reason for Exclusion
11 Ferreyra, A., G. Belletti, and M. Yorio, [Acuteconfusional state in hospitalized patients].Medicina,2004. 64(5): p. 385 389.
Included surgical and ICU patients
12 Flowers, S.R., Predisposing Factors of Delirium inPatients on a General Medical Nephrology Unit,2012, University of Nevada, Reno: United States –Nevada
Assessing prevalent delirium.
13 Formiga, F., et al., Predisposing factors of delirium inhip fractured patients older than 84 years. Factoresfavorecedores de la aparición de cuadro confusionalagudo en pacientes mayores de 84 años confractura de fémur, 2005. 124(14): p. 535 537.
Patients admitted to surgical unitpost surgery
14 Givens, J.L., R.N. Jones, and S.K. Inouye, The overlapsyndrome of depression and delirium in olderhospitalized patients. Journal Of The AmericanGeriatrics Society, 2009. 57(8): p. 1347 1353.
Secondary analysis of prior study.Outcome measure not delirium riskfactors but measuring outcomes ofoverlap between delirium anddepression
15 Grover, S., et al., Prevalence and clinical profile ofdelirium: a study from a tertiary care hospital innorth India. General Hospital Psychiatry, 2009.31(1): p. 25 29.
Patients admitted to psychiatry units.Included medical and surgicalpatients, no differentiation betweenpatients in reporting of results.
16 Han, L., et al., Use of medications withanticholinergic effect predicts clinical severity ofdelirium symptoms in older medical inpatients.Archives Of Internal Medicine, 2001. 161(8): p.1099 1105.
Assessing prevalent delirium.Poor differentiation of incident andprevalent delirium in reporting ofresults.
17 Henon, H., et al., Confusional state in stroke:Relation to preexisting dementia, patientcharacteristics, and outcome. Stroke, 1999. 30(4): p.773 779.
Unable to determinepresence/absence of delirium onadmission.No differentiation of incident andprevalent delirium in reporting ofresults.
18 Holden, J., S. Jayathissa, and G. Young, Deliriumamong elderly general medical patients in a NewZealand hospital. Internal Medicine Journal, 2008.38(8): p. 629 634.
Included patients admitted torehabilitation wardNo differentiation of incident andprevalent delirium in reporting ofresults.
19 Inouye, S.K., Predisposing and precipitating factorsfor delirium in hospitalized older patients. DementiaAnd Geriatric Cognitive Disorders, 1999. 10(5): p.393 400.
Summary article. Results to be usedfrom original publication
20 Iseli, R.K., et al., Delirium in elderly general medicalinpatients: a prospective study. Internal MedicineJournal, 2007. 37(12): p. 806 811.
Assessing only prevalent delirium. Noanalysis carried out for incident riskfactors, only those for prevalentdelirium.
21 Khurana, P.S., P.S.V.N. Sharma, and A. Avasthi, Riskfactors in delirious geriatric general medical
Unable to determinepresence/absence of delirium onadmission.
277
Excluded Studies Citation Reason for Exclusion
inpatients. Indian Journal Of Psychiatry, 2002. 44(3):p. 266 272.
No differentiation of incident andprevalent delirium in reporting ofresults.
22 Korevaar, J.C., B.C. van Munster, and S.E. de Rooij,Risk factors for delirium in acutely admitted elderlypatients: a prospective cohort study. BMC Geriatrics,2005. 5: p. 6 6.
Unable to determinepresence/absence of delirium onadmission. Assessed at 48 hours afteradmission.
23 Lin, R.Y., L.C. Heacock, and J.F. Fogel, Drug Induced,Dementia Associated and Non Dementia, Non DrugDelirium Hospitalizations in the United States, 19982005. Drugs & Aging, 2010. 27(1): p. 51 61.
No differentiation of incident andprevalent delirium in reporting ofresults.Poor specification of hospital settingpatients admitted to.
24 Lindsay, C.A., et al.,Medications associated withdelirium in hospitalized subjects. Pharmacotherapy,2011. 31(10): p. 432e.
Conference abstract. No full textarticle available.
25 Ljubisavljevic, V. and B. Kelly, Risk factors fordevelopment of delirium among oncology patients.General Hospital Psychiatry, 2003. 25(5): p. 345 352.
Unable to determinepresence/absence of delirium onadmission.
26 Margiotta, A., et al., Clinical characteristics and riskfactors of delirium in demented and not dementedelderly medical inpatients. The Journal Of Nutrition,Health & Aging, 2006. 10(6): p. 535 539.
No differentiation of incident andprevalent delirium in reporting ofresults.
27 Martin, N.J., The impact of environmental factors onthe development of delirium, 1997, University ofWaterloo (Canada): Canada. p. 236 p.
Included surgical patientsResults not differentiated intomedical or surgical risk factor analysis
28 McCusker, J., et al., Environmental risk factors fordelirium in hospitalized older people. Journal Of TheAmerican Geriatrics Society, 2001. 49(10): p. 13271334.
Unable to determinepresence/absence of delirium onadmission.Examining environmental risks inrelation to delirium severity notpresence of delirium
29 McManus, J., et al., The course of delirium in acutestroke. Age and Ageing, 2009. 38(4): p. 385 389.
No differentiation of incident andprevalent delirium in reporting ofresults.Unable to determinepresence/absence of delirium onadmission.
30 Mussi, C., et al., Importance of SerumAnticholinergic Activity in the Assessment of ElderlyPatients with Delirium. Journal of GeriatricPsychiatry and Neurology, 1999. 12(2): p. 82 86.
No measure of incident delirium.Only measured delirium prevalence
31 Nastri, L., et al., Delirium: Incidence and risk factorsin a group of hospitalized old inpatients. Rivista diPsichiatria, 2007. 42(4): p. 255 262.
Included surgical patients. Nodifferentiation betweenmedical/surgical patients in reportingof results. Unable to determinemedical only risk factors.
278
Excluded Studies Citation Reason for Exclusion
32 Oldenbeuving, A.W., et al., Delirium in the acutephase after stroke: incidence, risk factors, andoutcome. Neurology, 2011. 76(11): p. 993 999.
Unable to determinepresence/absence of delirium onadmission.
33 Rigney, T., Allostatic load and delirium in thehospitalized older adult. Nursing Research, 2010.59(5): p. 322 330.
Included patients admitted to ICU
34 Robinson, S. and C. Vollmer, Undermedication forpain and precipitation of delirium.Medsurg nursing :official journal of the Academy of Medical SurgicalNurses, 2010. 19(2): p. 79 83; quiz 84.
Included medical and surgicalpatients. No differentiation betweenmedical/surgical patients in reportingof results. Unable to determinemedical only risk factors.
35 Srinonprasert, V., et al., Risk factors for developingdelirium in older patients admitted to generalmedical wards. Journal Of The Medical AssociationOf Thailand = Chotmaihet Thangphaet, 2011. 94Suppl 1: p. S99 S104.
No differentiation of incident andprevalent delirium in reporting ofresults.
36 Takeuchi, T., et al., Delirium in inpatients withrespiratory diseases. Psychiatry And ClinicalNeurosciences, 2005. 59(3): p. 253 258.
Included patients admitted to ICU
37 van Minister, B.C., et al., Serum S100B in elderlypatients with and without delirium. InternationalJournal Of Geriatric Psychiatry, 2010. 25(3): p. 234239.
Unable to determinepresence/absence of delirium onadmission.No differentiation of incident andprevalent delirium in reporting ofresults.
38 van Munster, B.C., et al., Polymorphisms in thecatechol o methyltransferase gene and delirium inthe elderly. Dementia And Geriatric CognitiveDisorders, 2011. 31(5): p. 358 362.
Included medical and surgicalpatients. No differentiation betweenmedical/surgical patients in reportingof results. Unable to determinemedical only risk factors.
39 van Munster, B.C., et al., Genetic polymorphisms inthe DRD2, DRD3, and SLC6A3 gene in elderlypatients with delirium. American Journal Of MedicalGenetics. Part B, Neuropsychiatric Genetics: TheOfficial Publication Of The International Society OfPsychiatric Genetics, 2010. 153B(1): p. 38 45
Included medical and surgicalpatients. No differentiation betweenmedical/surgical patients in reportingof results. Unable to determinemedical only risk factors.
40 van Munster, B.C., A.H. Zwinderman, and S.E. deRooij, Genetic variations in the interleukin 6 andinterleukin 8 genes and the interleukin 6 receptorgene in delirium. Rejuvenation Research, 2011.14(4): p. 425 428.
Included medical and surgicalpatients. No differentiation betweenmedical/surgical patients in reportingof results. Unable to determinemedical only risk factors.
41 Villalpando Berumen, J.M., et al., Incidence ofDelirium, Risk Factors, and Long Term Survival ofElderly Patients Hospitalized in a Medical SpecialtyTeaching Hospital in Mexico City. InternationalPsychogeriatrics, 2003. 15(4): p. 325 336.
Unable to determinepresence/absence of delirium onadmission.Included medical and surgicalpatients. No differentiation betweenmedical/surgical patients in reportingof results. Unable to determinemedical only risk factors.
279
Excluded Studies Citation Reason for Exclusion
42 Vollmer, C.M., et al., Incidence, prevalence, andunder recognition of delirium in urology patients.Urologic Nursing, 2010. 30(4): p. 235.
No differentiation of incident andprevalent delirium in reporting ofresults.
43 Weinrebe, W., et al., Low muscle mass of the thigh issignificantly correlated with delirium and worsefunctional outcome in older medical patients [2].Journal Of The American Geriatrics Society, 2002.50(7): p. 1310 1311.
Unable to determinepresence/absence of delirium onadmission. Unable to determine ifdelirium developed duringhospitalisation
44 Yang, F.M., et al., Participation in activity and risk forincident delirium. Journal Of The AmericanGeriatrics Society, 2008. 56(8): p. 1479 1484.
Unable to determinepresence/absence of delirium onadmission. Unable to determine ifdelirium developed duringhospitalisation
45 Formiga, F., et al., Acute confusional syndrome inelderly patients hospitalized due to medicalcondition. Revista Clínica Española, 2005. 205(10): p.484 488.
Article not available in English
46 Formiga, F., et al., Prevalence of delirium in patientsadmitted because of medical conditions.MedicinaClínica, 2007. 129(15): p. 571 573.
Article not available in English
47 Gotor, P., J.I. González Montalvo, and T. Alarcón,Factors associated to the appearance of delirium ingeriatric patients with hip fracture Factoresasociados a la aparición de delirium en pacientesgeriátricos con fractura de cadera. RevistaMultidisciplinar de Gerontologia, 2004. 14(3): p.138 148.
Article not available in English
48 Regazzoni, C.J., M. Aduriz, and M. Recondo, Deliriumdevelopment during hospitalization of elderlypatients. Sindrome confusional agudo en el ancianointernado, 2000. 60(3): p. 335 338.
Article not available in English
280
Appendix 13 – Joanna Briggs Institute individual study critical
appraisal results for included studies
Alagiakrishnan et al. 2009 Gaps in patient care practices to prevent hospital acquired deliriumCriteria Judgement Comments/Description1 Is sample representative of patients in the
population as a whole?Yes Investigated all patients >65
years who were admitted tomedical unit.
2 Are the patients at a similar point in thecourse of their condition/illness?
Yes All patients were investigatedon admission to hospital.Although different illnesses,were all acutely ill at the time.
3 Has bias been minimised in relation toselection of cases and of controls?
Yes Cases selected if they scoredpositive on CAM, all patientstested using CAM
4 Are confounding factors identified andstrategies to deal with them stated?
No All possible risk factors wereidentified and measured forassociation and also monitoredpractices that may increasedelirium risk. However, nodocumentation of possibleconfounding factors.
5 Are outcomes assessed using objectivecriteria?
Yes All possible risk factors andoutcomes were assessed usingvalid tools. E.g Katz et al ADLscale, CAM, Mini Cog
6 Was follow up carried out over a sufficienttime period?
NA Study period was conducted foronly time spent in hospital. Nofollow up after discharge wasconducted
7 Were the outcomes of people who withdrewdescribed and included in the analysis?
Unclear 5 patients died. Unclear if theseresults were included inanalysis.
8 Were outcomes measured in a reliable way? Yes All possible risk factors andoutcomes were assessed usingvalid tools. E.g Katz et al ADLscale, CAM, Mini Cog
9 Was appropriate statistical analysis used? Yes Mann Whitney U testcontinuous variables and Fisherexact test for categoricalvariables
Include? Yes
281
Campbell et al 2011 Association between prescribing of anticholinergic medications andincident delirium: a cohort studyCriteria Judgement Comments/Description1 Is sample representative of patients in the
population as a whole?Yes Investigated all patients >65
years who were admitted togeneral medical ward.
2 Are the patients at a similar point in thecourse of their condition/illness?
Yes All patients were investigated onadmission to hospital. Althoughdifferent illnesses, were allacutely ill at the time.
3 Has bias been minimised in relation toselection of cases and of controls?
Yes Cases selected if they scoredpositive on CAM, all patientstested using CAM and SPMSQ
4 Are confounding factors identified andstrategies to deal with them stated?
No Authors state that confoundersmay have been introduced andaffected the findings but do notidentify any particular factors asconfounders
5 Are outcomes assessed using objectivecriteria?
Yes Outcomes were assessed usingvalid tools
6 Was follow up carried out over a sufficienttime period?
NA Study period was conducted foronly time spent in hospital. Nofollow up after discharge wasconducted
7 Were the outcomes of people whowithdrew described and included in theanalysis?
No No reported documentation ofpatients who withdrew.
8 Were outcomes measured in a reliableway?
Yes Delirium assessed using CAM.Cognition using the SPMSQ.Illness assessed using CCI
9 Was appropriate statistical analysis used? Yes Fisher exact test and analysis ofvariance used to compare data.
Include? Yes
282
Franco et al. 2010 Relationship between cognitive status at admission and incident delirium inolder medical patientsCriteria Judgement Comments/Description1 Is sample representative of patients in the
population as a whole?Yes All patients >60 years admitted
to hospital in medical ward wereable to be evaluated
2 Are the patients at a similar point in thecourse of their condition/illness?
Yes All patients were investigated onadmission to hospital. Althoughdifferent illnesses, were allacutely ill at the time.
3 Has bias been minimised in relation toselection of cases and of controls?
Yes Standardised tools were used toselect cases. CAM and DRS R98were used to test for deliriumcases.
4 Are confounding factors identified andstrategies to deal with them stated?
No All possible risk factors wereidentified and measured forassociation. However, nodocumentation of possibleconfounding factors.
5 Are outcomes assessed using objectivecriteria?
Yes Outcomes were assessed usingvalid tools
6 Was follow up carried out over a sufficienttime period?
NA Study period was conducted foronly time spent in hospital. Nofollow up after discharge wasconducted
7 Were the outcomes of people who withdrewdescribed and included in the analysis?
No No reported documentation ofpatients who withdrew.
8 Were outcomes measured in a reliable way? Yes All outcomes were measuredusing valid tools such as MMSE ,CAM and DRS R98
9 Was appropriate statistical analysis used? Yes Mann Whitney test. T tests andchi squares used to evaluatedifferences.
Include? Yes
283
Inouye & Charpentier 1996 Precipitating factors for delirium in hospitalized elderly persons:Predictive model and interrelationship with baseline vulnerabilityCriteria Judgement Comments/Description1 Is sample representative of patients in the
population as a whole?Yes All patients >70 years admitted
to hospital in general medicalward were able to be evaluated
2 Are the patients at a similar point in thecourse of their condition/illness?
Yes All patients were investigated onadmission to hospital. Althoughdifferent illnesses, were allacutely ill at the time.
3 Has bias been minimised in relation toselection of cases and of controls?
Yes Standardised tools were used toselect cases that were positivefor delirium (CAM). Researcherswere blinded to the researchquestion.
4 Are confounding factors identified andstrategies to deal with them stated?
Yes All possible risk factors andoutcomes were assessed usingvalid tools. Delirium onlyassessed every second day butwas supplemented withinterviews with nursing staff.However, no documentation ofpossible confounding factors.
5 Are outcomes assessed using objectivecriteria?
Yes All outcomes were measuredusing valid tools
6 Was follow up carried out over a sufficienttime period?
NA Study period was conducted foronly time spent in hospital. Nofollow up after discharge wasconducted
7 Were the outcomes of people who withdrewdescribed and included in the analysis?
No No reported documentation ofpatients who withdrew.
8 Were outcomes measured in a reliable way? Yes All outcomes were measuredusing valid tools includingAPACHE, MMSE, CAM.
9 Was appropriate statistical analysis used? Yes T test for continuous variablesor x² statistics for categoricalvariables
Include? Yes
284
Jones et al. 2006 Does Educational Attainment Contribute to Risk for Delirium? A PotentialRole for Cognitive ReserveCriteria Judgement Comments/Description1 Is sample representative of patients in the
population as a whole?Yes Evaluated results of 2 studies
that investigated all patients >70years admitted to hospital ingeneral medical ward.
2 Are the patients at a similar point in thecourse of their condition/illness?
Yes All patients were investigated onadmission to hospital. Althoughdifferent illnesses, were allacutely ill at the time.
3 Has bias been minimised in relation toselection of cases and of controls?
Yes Standardised tools were used toselect cases that were positivefor delirium (CAM).
4 Are confounding factors identified andstrategies to deal with them stated?
No Attempted to identify allpossible risk factors andoutcomes were assessed usingpatient data and valid tools formeasurement. Controlled forvariables known to be riskfactors, but did not state how.
5 Are outcomes assessed using objectivecriteria?
Yes All outcomes were measuredusing valid tools
6 Was follow up carried out over a sufficienttime period?
NA Study period was conducted foronly time spent in hospital. Nofollow up after discharge wasconducted
7 Were the outcomes of people who withdrewdescribed and included in the analysis?
No No reported documentation ofpatients who withdrew.
8 Were outcomes measured in a reliable way? Yes All outcomes were measuredusing valid tools including CAM,APACHE, MMSE and BlessedDementia rating scale (BDRS).
9 Was appropriate statistical analysis used? Yes Logistic regression controllingfor known risk factors.
Include? Yes
285
McAvay et al. 2007 Depressive symptoms and the risk of incident delirium in older hospitalizedadultsCriteria Judgement Comments/Description1 Is sample representative of patients in the
population as a whole?Yes Secondary analysis of results of
1 study that investigated allpatients >70 years admitted tohospital in general medicalward.
2 Are the patients at a similar point in thecourse of their condition/illness?
Yes All patients were investigated onadmission to hospital. Althoughdifferent illnesses, were allacutely ill at the time.
3 Has bias been minimised in relation toselection of cases and of controls?
Yes Standardised tools were used toselect cases that were positivefor delirium (CAM).
4 Are confounding factors identified andstrategies to deal with them stated?
Yes To avoid confounding factor ofdepression, patients taking antidepressants on admission wereexcluded. No other confoundingfactors identified.
5 Are outcomes assessed using objectivecriteria?
Yes All outcomes were measuredusing valid tools
6 Was follow up carried out over a sufficienttime period?
NA Study period was conducted foronly time spent in hospital. Nofollow up after discharge wasconducted
7 Were the outcomes of people who withdrewdescribed and included in the analysis?
No No reported documentation ofpatients who withdrew.
8 Were outcomes measured in a reliable way? Yes All outcomes were measuredusing valid tools including CAM,MMSE, Geriatric DepressionScale (GDS)
9 Was appropriate statistical analysis used? Yes Cox proportional hazardsregression model.
Include? Yes
286
O'Keeffe & Lavan 1996 Predicting delirium in elderly patients: development and validation of arisk stratification modelCriteria Judgement Comments/Description1 Is sample representative of patients in the
population as a whole?Yes Investigated all patients
admitted to hospital in acutecare geriatric ward.
2 Are the patients at a similar point in thecourse of their condition/illness?
Yes All patients were investigated onadmission to hospital. Althoughdifferent illnesses, were allacutely ill at the time.
3 Has bias been minimised in relation toselection of cases and of controls?
Yes Standardised tool DeliriumAssessment Scale (DAS) wasused to select cases that werepositive for delirium.
4 Are confounding factors identified andstrategies to deal with them stated?
No Attempted to identify allpossible risk factors andoutcomes were assessed usingpatient data and valid tools formeasurement. However, nodocumentation of possibleconfounding factors.
5 Are outcomes assessed using objectivecriteria?
Yes All outcomes were measuredusing valid tools
6 Was follow up carried out over a sufficienttime period?
NA Study period was conducted foronly time spent in hospital. Nofollow up after discharge wasconducted
7 Were the outcomes of people who withdrewdescribed and included in the analysis?
No No reported documentation ofpatients who withdrew.
8 Were outcomes measured in a reliable way? Yes All outcomes were measuredusing valid tools including Katzet al ADL scale, GDS, BDRS andMMSE
9 Was appropriate statistical analysis used? Yes Unadjusted odds ratios (OR) andconfidence intervals (CI) werecalculated for each variable.
Include? Yes
287
Wakefield 2002 Risk for acute confusion on hospital admissionCriteria Judgement Comments/Description1 Is sample representative of patients in the
population as a whole?No Only investigated male patients
>65 years admitted to VAhospital in general medical units
2 Are the patients at a similar point in thecourse of their condition/illness?
Yes All patients were investigated onadmission to hospital. Althoughdifferent illnesses, were allacutely ill at the time.
3 Has bias been minimised in relation toselection of cases and of controls?
Yes Standardised tool – NEECHAMconfusion scale was used toselect cases that were positivefor delirium.
4 Are confounding factors identified andstrategies to deal with them stated?
No Attempted to identify allpossible risk factors andoutcomes were assessed usingpatient data and valid tools formeasurement. However, nodocumentation of possibleconfounding factors.
5 Are outcomes assessed using objectivecriteria?
Yes All outcomes were measuredusing valid tools
6 Was follow up carried out over a sufficienttime period?
NA Study period was conducted foronly time spent in hospital. Nofollow up after discharge wasconducted
7 Were the outcomes of people who withdrewdescribed and included in the analysis?
No No reported documentation ofpatients who withdrew.
8 Were outcomes measured in a reliable way? Yes All outcomes were measuredusing valid tools including Katzet al ADL scale, GDS and MMSE
9 Was appropriate statistical analysis used? Yes T tests for 2 groups withunequal variances. OR werecalculated and significance wasdetermined by Mantel Haenszelchi square or Fisher’s exact test.
Include? Yes
288
Wilson et al. 2005 Plasma insulin growth factor 1 and incident delirium in older peopleCriteria Judgement Comments/Description1 Is sample representative of patients in the
population as a whole?Yes Investigated all patients >75
years admitted to hospital ingeneral medical units
2 Are the patients at a similar point in thecourse of their condition/illness?
Yes All patients were investigated onadmission to hospital. Althoughdifferent illnesses, were allacutely ill at the time.
3 Has bias been minimised in relation toselection of cases and of controls?
Yes Standardised tool – CAM wasused to select cases that werepositive for delirium.
4 Are confounding factors identified andstrategies to deal with them stated?
No Attempted to identify allpossible risk factors andoutcomes were assessed usingpatient data and valid tools formeasurement. However, nodocumentation of possibleconfounding factors.
5 Are outcomes assessed using objectivecriteria?
Yes All outcomes were measuredusing valid tools
6 Was follow up carried out over a sufficienttime period?
NA Study period was conducted foronly time spent in hospital. Nofollow up after discharge wasconducted
7 Were the outcomes of people who withdrewdescribed and included in the analysis?
No No reported documentation ofpatients who withdrew.
8 Were outcomes measured in a reliable way? Yes All outcomes were measuredusing valid tools including Katzet al ADL scale, GDS, MMSE andInformant questionnaire forcognitive decline in the elderly(IQCODE)
9 Was appropriate statistical analysis used? Yes Univariate analysis conducted toexamine relationship betweenvariable data and delirium
Include? Yes
289