A NEW SCREENING TOOL FOR MILD COGNITIVE IMPAIRMENT (MCI)
Prof D.W Molloy
Dr Rónán O’CaoimhCENTRE FOR GERONTOLOGY & REHABILITATION
ST FINBARRS HOSPITAL/UNIVERSITY COLLEGE CORKCORK CITYIRELAND
The Quick Mild Cognitive Impairment Screen:
(Qmci)
Overview
1. Demographics of cognitive impairment.2. Challenges in screening, diagnosing &
assessing cognition…..How to choose appropriate cognitive screening tools.
3. Introduction to the Qmci.4. Development of the Qmci.5. The Quick Memory Check.6. RAPCOG (The Rapid Community COGnitive
screening programme.)
Challenges in Diagnosing Cognitive Impairment
Challenges in screening, diagnosing & assessing cognition.
- How to choose appropriate cognitive & functional screening tools.
- When is it dementia? Normal <> Mild Cognitive Impairment (MCI) <> Mild Dementia
- What separates people who live well with dementia from those that do not?
- What does a typical person with dementia behave like...in their own environment?
Do people want/benefit from this?
Natural History of Cognitive Decline
Normal AAMI Mild Cognitive Impairment (MCI) Dementia
No Age Subjective Memory loss Functionalmemory Associated No/subtle Functional Impairment Impairment loss Memory Some changes on screening tests Impairment Deficiencies on more in-depth Neuro-psych testing
5-10% convert each year In 8 years, 100% conversion
http://www.alz.org/alzheimers_disease_facts_and_figures.asp
http://www.alz.org/alzheimers_disease_facts_and_figures.asp
Cognitive Impairment in Ireland
Four million citizens440,000 aged over 6544,000 with dementiaBy 2050 there will be 90,000-120,000 cases
Approx 10% >65 with MCI
Population is ageing: 2006- 11% over 65
2041-22% (1.4 million)
ref: McGill, P. Illustrating Ageing in Ireland North and South: Key Facts and Figures. Belfast: Centre for Ageing Research and Development in Ireland, 2010
Cognitive Impairment in the Community in
Ireland
• Common in Community.• 16.7% of patients followed by their public health • nurse, living in the community have CI.• Mean age 81.8 years, SD +/-7 (Older, p=0.001).• Predominantly female (70%).• More functionally impaired,(Median Barthel score 14 Vs 18, p<0.001).• Half live alone (p=0.11).
Screening for Cognitive Impairment
Strong evidence that cognitive impairment is under-recognised: 40-75% attending their GP are missed.
ref Valcour et al, Arch Int Med 2000
Detecting & distinguishing mild dementia & MCI is a challenge. ref Löppönen et al Age and Ageing 2003
MCI represents a heterogeneous group of disorders of memory impairment.
How can the “worried well” be differentiated from patients with MCI?
Need sensitive & specific cognitive screening instruments.
Screening for Cognitive Impairment
Sensitivity: If I have the disease, is the test positive?
Specificity: If I do not have the disease, is the test negative?
Too Sensitive: over diagnose, normal's misdiagnosed (including the “worried well”).
Too Specific: under diagnose, people with the disease called normal (“miss highly educated”).
Many screening instruments with different sensitivities & specificities.
Why Screen?
Early Dx allows Prompt initiation of appropriate care.Important to differentiate between MCI, NC &
Dementia – Different Rx & prognosis.Allows identification of reversible causes – 10 -50 %
of referrals to memory clinics ( <1.5% cases of mild to mod dementia).
Screening for vascular risk factors.Initiation of lifestyle advice.Planning for the future (advanced care plans/power of
attorney). Potential for future treatments & reducing stigma.
What we currently use…
1. Abreviated Mental Test Score (MTS)2. Mini-Mental State Examination/Standardised
MMSE widely used.Bias is for more education.
-Score 29 for 70 yr old with 12 yrs education-Score 23 for 70 yr old with 3 yrs education
Low ceiling and high floor. Does not pick up early cognitive losses or MCI.
3. Other newer, cognitive screens specific for MCI include the Montreal Cognitive Assessment.-Narrow scoring range.-Too difficult for measuring progression of dementia.-Too long (10-12 minutes) for busy clinical practice.
Standardised Mini-Mental State ExaminationMaximum Score
Patient`s Score
Questions
5 Orientation“What is the year? Season? Date? Day? Month?”
5 “Where are we now? State? County? Town/city? Hospital? Floor?”
3 RegistrationAsk the subject if you may test his/her memory. Then say "ball", "flag", "tree" clearly and slowly, about 1 second for each. After you have said all 3 words, ask him/her to repeat them - the first repetition determines the score (0-3):
5 Concentration & CalculationSerial 7sAsk the subject to begin with 100 and count backwards by 7. ( 93,86,79,72,65....). Stop after 5 subtractions. Score the correct subtractions.Or Spell WORLD backwardsAsk the subject to spell the word "WORLD" backwards. The score is the number of letters in correct position. For example, "DLROW" is 5, "DLORW" is 3, "LROWD" is 0.
3 RecallAsk the subject to recall the 3 words you previously asked him/her to remember.
2 NamingShow the subject 2 objects, e.g. a wrist watch & pencil and ask him/her what it is.
1 RepetitionAsk the subject to repeat the following :"No ifs, ands, or buts"
3 3 Stage Command followingGive the subject a plain piece of paper and say, "Take the paper in your hand, fold it in half, and put it on the floor."1 point for each TAKES, FOLDS, PUTS
1 ReadingHold up the card reading, "Close your eyes", so the subject can see it clearly. Ask him/her to read it and do what it says. Score correctly only if the subject closes eyes CLOSE YOUR EYES
1 WritingAsk him/her to write a sentence, It must be spontaneous and contain a subject and verb and be sensible.______________________________________________________________
1 CopyingGive subject a piece of paper and ask him/he to copy a design of two intersecting shapes. One point is awarded for correctly copying it. All angles on both figures must be present, and the figures must have one overlapping angle.
30 Total
??
Clock Drawing
RegistrationOrientation Recall
Verbal Fluency Logical Memory
© R O Caoimh, D W Molloy 2012
Quick Mild Cognitive Impairment Screen (Qmci)
Methods
Sampling Recruited from four memory clinics in Ontario Canada, by convenience sampling between 2004 -2010.
Dementia Subjects
Diagnosed clinically by a consult geriatrician after full work-up for alternative causes, based upon DSM-IV criteria.Severity correlated to Reisberg FAST scaleDepression excluded clinically supported by GDS score > 7.
Mild Cognitive Impairment subjects
Diagnosed clinically if1.Subjective & corroborated memory loss2.No obvious loss of function3.No requirement for objective impairment in memory tests
Controls (Normal)
Caregivers / those attending with subjects, with normal cognition.
Excluded Subjects with Lewy body dementia or Parkinsons disease.Depression excluded clinically (supported by GDS score > 7).Aged < 55 years & if English was not their first language.
Analysis Subjects were subdivided into age (>, < 75 years), education (>, < 12 years).Accuracy ( sensitivity & specificity ) determined using ROC curves.
Group Normal MCI DementiaNumber of subjects 630 154 181Age
Mean 67.4 73.6 78.1Median 67 75.5 79range 44 - 92 50 - 88 49 - 93
Education (years in education)mean 13.8 12.2 11.0median 13 12 10range 5 - 29 5 - 26 3 - 20
Qmci(median with IQR)
76 (83-69=14)
62 (68-53=15)
36 (45-23=22)
SMMSE(median with IQR)
29 (30-28=2)
28 (29-27=2)
22 (25-18=7)
Results
Group Normal MCI DementiaNumber of subjects 630 154 181Age
Mean 67.4 73.6 78.1Median 67 75.5 79range 44 - 92 50 - 88 49 - 93
Education (years in education)mean 13.8 12.2 11.0median 13 12 10range 5 - 29 5 - 26 3 - 20
Qmci(median with IQR)
76 (83-69=14)
62 (68-53=15)
36 (45-23=22)
SMMSE(median with IQR)
29 (30-28=2)
28 (29-27=2)
22 (25-18=7)
Results
Group Normal MCI DementiaNumber of subjects 630 154 181Age
Mean 67.4 73.6 78.1Median 67 75.5 79range 44 - 92 50 - 88 49 - 93
Education (years in education)mean 13.8 12.2 11.0median 13 12 10range 5 - 29 5 - 26 3 - 20
Qmci(median with IQR)
76 (83-69=14)
62 (68-53=15)
36 (45-23=22)
SMMSE(median with IQR)
29 (30-28=2)
28 (29-27=2)
22 (25-18=7)
Results
ROC curve demonstrating sensitivities and specificities of the Qmci, ABCS 135 and SMMSE in
differentiating (a). MCI and (b) dementia
O‘ Caoimh R et al. Age Ageing 2012;ageing.afs059
© The Author 2012. Published by Oxford University Press on behalf of the British Geriatrics Society.
(a). MCI from Normal (b). MCI from Dementia
Box Plots for each subtest of the Qmci showing the median and interquartile range scores for Dementia (D), Mild Cognitive Impairment (MCI) and Normal Cognition (NC)
Comparison of the Qmci Subtests
Orientation Registration Clock drawing
Delayed Recall Verbal Fluency Logical Memory
Qmci SubtestsDiagnosing MCI
QQmcimci Subtests- MCI versus Normal Subtests- MCI versus Normal
AUC
Variable(s) Area
Qmci Orient .57
Qmci Reg .56
Qmci Clock .66
Qmci DR .73
Qmci VF .77
Qmci LM .80
Qmci SubtestsDiagnosing Dementia
QQmcimci Subtests- MCI versus Dementia Subtests- MCI versus Dementia
AUC
Variable(s) Area
Qmci Orient .88
Qmci Reg .64
Qmci Clock .76
Qmci DR .84
Qmci VF .83
Qmci LM .82
Age & Education
Group(MCI & NC)
Test variablesArea Under Curve
(95% CI)Age <= 75 with education < 12 years N = 127
SMMSE 0.65 (0.54 – 0.76)
LM 0.72 (0.62 – 0.82)
Age <= 75 with education > = 12 yearsN = 449
SMMSE 0.66 (0.57 – 0.75)
LM 0.79 (0.73 – 0.86)
Age > 75 with education < 12 years N = 71
SMMSE 0.64 (0.51 – 0.77)
LM 0.74 (0.62 – 0.85)
Age > 75 with education > = 12 years N = 127
SMMSE 0.55 (0.44 – 0.66)
LM 0.79 (0.71 – 0.88)
Overall SMMSE 0.67 (0.62 – 0.72)
LM 0.80 (0.76 – 0.84)
Comparison of the best performing Qmci Subtest, LM, with SMMSEIn differentiating MCI from Normal Cognition.
Age & Education
Group(MCI & NC)
Test variablesArea Under Curve
(95% CI)Age <= 75 with education < 12 years N = 127
SMMSE 0.65 (0.54 – 0.76)
LM 0.72 (0.62 – 0.82)
Age <= 75 with education > = 12 yearsN = 449
SMMSE 0.66 (0.57 – 0.75)
LM 0.79 (0.73 – 0.86)
Age > 75 with education < 12 years N = 71
SMMSE 0.64 (0.51 – 0.77)
LM 0.74 (0.62 – 0.85)
Age > 75 with education > = 12 years N = 127
SMMSE 0.55 (0.44 – 0.66)
LM 0.79 (0.71 – 0.88)
Overall SMMSE 0.67 (0.62 – 0.72)
LM 0.80 (0.76 – 0.84)
Comparison of the best performing Qmci Subtest, LM, with SMMSEIn differentiating MCI from Normal Cognition.
Developing Cut-Off Scores
Developing Cut-Off Scores
Large variety of cognitive scores available for any test based upon sensitivity & specificity.
Represent “transition points” between different cognitive states.
Placing emphasis on cut-offs, rather than profiles of impairment or clinical judgement, can therefore be misleading.
Individual variations in age and educational level (Crum et al, JAMA, 1993).
Developing Cut-Off Scores
Normal Cognition (NC)Mild Cognitive Impairment (MCI)Dementia (D)
Developing Cut-Off Scores
Normal Cognition (NC)Mild Cognitive Impairment (MCI)Dementia (D) CI
Developing Cut-Off Scores
Normal Cognition (NC)Mild Cognitive Impairment (MCI)Dementia (D) D
Developing Cut-Off Scores
Normal Cognition (NC)Mild Cognitive Impairment (MCI)Dementia (D) D
Age(> or ≤75)
Education(≥ or <12)
Developing Cut-Off Scores
Developing Cut-Off Scores
Developing Cut-Off Scores
Developing Cut-Off Scores
Cut-off scores for the Qmci based upon age (> or < 75) and education (< or> 12 years) level.
Cut-off scores
Cut-off scores for the Qmci
Diagnosis Score Sensitivity Specificity
MCI <60 89% 86%
Dementia <50 87% 88%
Cut-off scores
Cut-off scores are important….using cut-off for young well educated adults on older adults with less formal education.
Diagnosis>75 age <12 edu
Correct(N)
Incorrect(N)
Misclassified(%)
CI 148 49 25%
Dementia 80 89 9%
Cut-off scores
Qmci Vs MoCA
Qmci Vs MoCA
Sampling Recruited from a university hospital memory clinics in Cork, by convenience sampling between 2012 -2013.
Dementia Subjects
Diagnosed clinically by a consult geriatrician after full work-up for alternative causes, based upon DSM-IV criteria.Severity correlated to Reisberg FAST scaleDepression excluded clinically supported by GDS score ≥ 7.
Mild Cognitive Impairment subjects
Diagnosed clinically if1.Subjective & corroborated memory loss2.No obvious loss of function3.No requirement for objective impairment in memory tests
Controls (Normal)
Subjects referred with memory loss but with normal cognition.
Excluded Subjects with Lewy body dementia or Parkinsons disease.Depression excluded clinically (supported by GDS score > 7).Aged < 45 years & if English was not their first language.
Analysis Subjects were subdivided into age (>, < 75 years), education (>, < 12 years).Accuracy ( sensitivity & specificity ) determined using ROC curves.
Group Normal MCI DementiaNumber of assessments 78 93 250Age
Median 72 76 77Interquartile range 79-65=14 80-70=10 82-72=10range 46-89 58-91 48-97
Qmci(median with IQR)
67.5(73-63=10)
58(64-52=12)
36(45-22=23)
MoCA(median with IQR)
25(27-23=4)
22(24-20=4)
13(17-8=9)
SMMSE(median with IQR)
29(30-28=2)
28(29-26=3)
22(25-16=9)
Results
Group Normal MCI DementiaNumber of assessments 78 93 250Age
Median 72 76 77Interquartile range 79-65=14 80-70=10 82-72=10range 46-89 58-91 48-97
Qmci(median with IQR)
67.5(73-63=10)
58(64-52=12)
36(45-22=23)
MoCA(median with IQR)
25(27-23=4)
22(24-20=4)
13(17-8=9)
SMMSE(median with IQR)
29(30-28=2)
28(29-26=3)
22(25-16=9)
Results
Group Normal MCI DementiaNumber of assessments 78 93 250Age
Median 72 76 77Interquartile range 79-65=14 80-70=10 82-72=10range 46-89 58-91 48-97
Qmci(median with IQR)
67.5(73-63=10)
58(64-52=12)
36(45-22=23)
MoCA(median with IQR)
25(27-23=4)
22(24-20=4)
13(17-8=9)
SMMSE(median with IQR)
29(30-28=2)
28(29-26=3)
22(25-16=9)
Results
Qmci V MoCA - Preliminary data
a. MCI versus NC b. MCI versus Dementia
Variable
AUC
Qmci 0.82
MoCA 0.74
SMMSE 0.71
Variable
AUC
Qmci 0.96
MoCA 0.91
SMMSE 0.90
Qmci Vs SADAS-cog
Qmci Vs sADAS-cog
Alzheimer`s Disease Assessment Scale-cognitive section and its standardised version (SADAS-cog) are the current standard for assessing cognitive outcomes in clinical trials of dementia.
11 domains, including: word recall, object naming, command following, construction, orientation, word recognition, language, speech comprehension, word finding and recall: score from 0-70.
ADAS-cog…45 minutes to administer, questions over IRR, need for training.
Retrospective comparison of Qmci Vs SADAS-cog in a clinical trial – DARAD trial.
Also compared to clinical dementia rating (CDR) scale and the Quick activities of daily living screen (Qadl).
Qmci Vs sADAS-cog
Multicentre randomised clinical trial of 406 patients with mild to moderate Alzheimer’s dementia.
Correlations were estimated using regression:
- at each time point- all time points- mean values across time
Qmci Vs sADAS-cog
Qmci Qadl CDR
Scatter plot depicting correlations between a) SADAS-cog and Qmci, b) SADAS-cog and Qadl and c) SADAS-cog and CDR ,at each time point, (1,3,6,9,12 months).
Qmci Vs sADAS-cog
Strong and significant correlation between the SADAS-cog and Qmci, r=-0.75, p<0.001.
SADAS-cog and Qmci also correlated (moderate) with CDR and Qadl scores .
Qmci correlated strongly with the SADAS-cog and both were equally responsive to deterioration.
Qmci could be substituted or used in conjunction with longer tests to allow for more frequent assessment of cognition.
Qmci Summary
Qmci more accurately differentiated MCI from NC than the original ABCS 135,the SMMSE & MoCA.
Ability of the Qmci to identify MCI was better for those over 75 years.
It is as accurate as the SMMSE in distinguishing MCI from Dementia (monitoring progression).
LM is the best performing subtest for differentiating MCI from NC, irrespective of age or educational status.
Orientation was the best performing subtest for identifying/monitoring progression to dementia.
QMT (based upon Qmci) is accurate for detecting both MCI and dementia.
Qmci Conclusion
The Qmci is….SHORT (Median time 4.24 mins,range 3-5 mins).Measures the whole range from normal through
ageing changes, MCI and through dementia.Reliable & Valid.Cut-offs easy to remember:60 for CI50 for DementiaEASY to Learn and USE!QMT has the potential to increase the up take.…
move more towards a consumer driven healthcare (apps etc).
Caregiver TestingFew informant questionnaires available.The Informant Questionnaire on Cognitive Decline in
the Elderly (IQCODE) is the most widely used.Recently, the AD8 has been advocated as a short
screen. One “patient” administered test developed to date…the
TYM- Test Your Memory screenReliability affected by informant characteristics such
as depression, anxiety and the quality of their relationship.
Evidence that informant testing complements direct screening/testing. Value of direct screening by informants, prior to clinic is unknown.
QMT
“Quick Memory Check” or QMC.New Caregiver administer cognitive screen.Preliminary data (ongoing data collection).Currently being validated in our memory clinic. Two
components: A. Caregiver report ( validated against IQcode/AD8).B. Caregiver assessment (against Qmci/MoCA).Scored out of 100
Orientation – 15Verbal Fluency – 40Logical Memory – 45
QMT
Preliminary data (ongoing data collection).
Group Normal MCI DementiaNumber of subjects 46 50 142Age
Median 73 76 77Interquartile range (81-65=16) (80-70=10) (81-71=10)range 46-89 60-91 48-92
QMC(median with IQR)
62(73-53=20)
50(58-39=19)
31(39-16=23)
Qmci(median with IQR)
67(72-62=10)
56(62-52=10)
33(43-20=23)
SMMSE(median with IQR)
29(30-28=2)
28(29-26=3)
21(25-16=9)
MoCA(median with IQR)
25(26-23=3)
22(24-20=4)
12(17-7=10)
QMC - Preliminary dataNormal Cognition versus Cognitive Impairment
Variable
AUC
Qmci 0.95
MoCA 0.89
SMMSE 0.90
QMC 0.87
QMC
The QMC, administered by untrained raters in an informal setting, compares favourably to established cognitive screening tests administered by trained raters.
Potential use of caregiver administered screening tests in identifying cognitive impairment.
1. In clinics, reducing testing time & patient stress (white coat effect).
2. Use as a screen for MCI & early dementia in the community.
Positive feedback from carers and some patients who screened themselves.
Community Screening
Screening programmes for detecting CI are advocated (Boustani 2003).
Limited evidence supporting routine screening in clinical practice (Boustani 2005).
Insufficient evidence for sensitive and specific cognitive tests to differentiate between normal cognition (NC), MCI and dementia (Winblad 2004, Boustani 2005).
RAPCOG Pathway
RAPCOG Pathway
RAPCOG
Pilot study, screening 700 older adults, will being in July 2013.
RAPCOG has the potential to improve integration between the general public, primary & secondary care.
Qmci is now translated and being validated in several different languages
Qmci-P (Portuguese)Qmci-D (Dutch)Qmci-T (Turkish)Qmci-G (German)Qmci-PL (Polish)
RAPCOG
Need to develop a stronger evidence base for early diagnosis AND treatment.
Screening people in the community for decline in cognition and function is a delicate matter and not everyone will be amenable to it.
Improving knowledge & reducing stigma will help.See: www.ucc.ie/alzcafe
Conclusion
There is a need to diagnose MCI early.We need to study the natural history of this disease
more to understand it better.We need better, shorter instruments to detect it and to
measure change over time in the clinic and in General practice.
Qmci may be that instrument.Ongoing research:
-examining its performance in GP, (IRR & Validity).-use in the community to rapidly screen and triage those with CI.-IRR and Validity of Qmci App.-Comparison of QMT (caregiver report to Iqcode & AD8).
RAPCOG: The Rapid Community Cognitive Screening program.
Questions?
http://www.hseland.ie/coursefiles/packages/
OpenBrookes_FS12/player.html
Thank you
RAPCOG Pathway