JAMDA 19 (2018) 948e958

JAMDA

journal homepage: www.jamda.com

Original Study

Exploring Antipsychotic Prescribing Behaviors for Nursing HomeResidents With Dementia: A Qualitative Study

Kieran A. Walsh MPharm a,b,c,*, Carol Sinnott PhD d, Aoife Fleming PhD b,e,Jenny Mc Sharry PhD f, Stephen Byrne PhD b, John Browne PhD c,Suzanne Timmons MD a

aCenter for Gerontology and Rehabilitation, School of Medicine, University College Cork, Cork, Irelandb Pharmaceutical Care Research Group, School of Pharmacy, University College Cork, Cork, Irelandc School of Public Health, University College Cork, Cork, Irelandd The Healthcare Improvement Studies (THIS) Institute, School of Clinical Medicine, University of Cambridge, Cambridge, United Kingdome Pharmacy Department, Mercy University Hospital, Cork, IrelandfHealth Behavior Change Research Group, School of Psychology, National University of Ireland Galway, Galway, Ireland

Keywords:Theoretical Domains Framework (TDF)dementiaantipsychoticsqualitativenursing homebehavioral and psychological symptoms ofdementia (BPSD)

KW is funded by the Health Research Board (Ithropies, a limited life foundation, and this researchSPHeRE Program under Grant No. SPHeRE/2013/1. CSlectureship, by the National Institute for Health ReseCare Research (NIHR SPCR).

The investigators were solely responsible for threcruitment, data collections, analysis and preparatiosources did not participate in this process.

https://doi.org/10.1016/j.jamda.2018.07.0041525-8610/� 2018 AMDA e The Society for Post-Acu

a b s t r a c t

Objectives: Caution is advised when prescribing antipsychotics to people with dementia. This studyexplored the determinants of appropriate, evidence-based antipsychotic prescribing behaviors fornursing home residents with dementia, with a view to informing future quality improvement efforts andbehavior change interventions.Design: Semistructured qualitative interviews based on the Theoretical Domains Framework (TDF).Setting and Participants: A purposive sample of 27 participants from 4 nursing homes, involved in the careof nursing home residents with dementia (8 nurses, 5 general practitioners, 5 healthcare assistants, 3family members, 2 pharmacists, 2 consultant geriatricians, and 2 consultant psychiatrists of old age) in aSouthern region of Ireland.Measures: Using framework analysis, the predominant TDF domains and determinants influencing thesebehaviors were identified, and explanatory themes developed.Results: Nine predominant TDF domains were identified as influencing appropriate antipsychotic pre-scribing behaviors. Participants’ effort to achieve “a fine balance” between the risks and benefits of anti-psychotics was identified as the cross-cutting theme that underpinned many of the behavioraldeterminants. On one hand, neither healthcareworkers nor familymembers wanted to see residents over-sedated andwithout a quality of life. Conversely, the reality of needing toprotect staff, familymembers, andresidents from potentially dangerous behavioral symptoms, in a resource-poor environment, wasemphasized. The implementation of best-practice guidelineswas illustrated through3 explanatory themes(“human suffering”; “the interface between resident and nursing home”; and “power and knowledge:complex stakeholder dynamics”), which conceptualize how different nursing homes strike this “finebalance.”Conclusions: Implementing evidence-based antipsychotic prescribing practices for nursing home resi-dents with dementia remains a significant challenge. Greater policy and institutional support is requiredto help stakeholders strike that “fine balance” and ultimately make better prescribing decisions. Thisstudy has generated a deeper understanding of this complex issue and will inform the development of anevidence-based intervention.

� 2018 AMDA e The Society for Post-Acute and Long-Term Care Medicine.

reland) and Atlantic Philan-was conducted as part of theis funded, through a clinicalarch UK, School for Primary

e design, methods, subjectn of paper and the funding

The authors declare no conflicts of interest.* Address correspondence to Kieran A. Walsh, MPharm, Pharmaceutical Care

Research Group, School of Pharmacy, University College Cork, Cork, Ireland, T12YN60.

E-mail address: kieranwalsh@umail.ucc.ie (K.A. Walsh).

te and Long-Term Care Medicine.

K.A. Walsh et al. / JAMDA 19 (2018) 948e958 949

Guidelines advise against antipsychotics for the first-line man-agement of behavioral and psychological symptoms of dementia(BPSD)1,2 because of the increased risks of stroke and mortality.3e5

However, antipsychotics can be appropriate when behavioral symp-toms are severe, dangerous, or distressing to the person with de-mentia.1,2 Despite the existence of guidelines for over a decade andnational level efforts to improve dementia care, antipsychotic pre-scribing is still common, especially in nursing home (NH) settings.6e8

Global estimates of antipsychotic prescribing prevalence in NH resi-dents vary from 16% in the US,9 19% in England,6 to 27% acrossWesternEurope.7

A systematic review examining the effectiveness of interventionsto reduce inappropriate prescribing of antipsychotics to NH residentswith dementia reported that the majority of interventions wereeffective in the short term.10 However, the long-term effects wereassessed in only 4 studies, with prescribing returning to baselinelevels in 2 studies.11,12

Successful implementation of evidence-based practice requireseffective and sustained behavior change, beginning with a thoroughunderstanding of the problem.13 A body of qualitative research hasexplored problematic clinical decision-making in this area. Werecently published a systematic review of this literature and foundthat the use of antipsychotics in NHs is the culmination of a range ofhealthcare professional behaviors.14 The 2 main behaviors identifiedwere appropriate requesting and prescribing of antipsychotics. How-ever, there has been a lack of exploration of these behaviors asstandalone processes and in terms of how they influence each other.Furthermore, there has been limited exploration of how differentstakeholders perceive these interacting behaviors. Hence, gaps in ourunderstanding remain, which will be best answered by further qual-itative research.

The Theoretical Domains Framework (TDF) is an integrativeframework of influences on behavior, identified by synthesizingmultiple behavior change theories.15 The TDF consists of 14 domains(Table 1) and provides a comprehensive, theory-informed approach toidentifying the determinants (ie, barriers and facilitator), which in-fluence clinical behaviors.15 Utilization of the TDF will help us toidentify the determinants that influence prescribing behaviors andhence, support progression from exploration to intervention.16

The aim of this qualitative study was to explore and interpret thedeterminants of appropriate prescribing behaviors (requesting andprescribing) among a range of individuals involved in the care of NHresidents with dementia, with a view to informing future qualityimprovement efforts and behavior change interventions.

Table 1Theoretical Domains Framework

Domain Definition

Behavioral regulation Anything aimed at managing or cBeliefs about capabilities Acceptance of the truth, reality, orBeliefs about consequences Acceptance of the truth, reality, oEmotion A complex reaction pattern, invol

attempts to deal with a personEnvironmental context and resources Any circumstance of a person’s si

and abilities, independence, socGoals Mental representations of outcomIntentions A conscious decision to performKnowledge An awareness of the existence ofMemory, attention, and decision processes The ability to retain information,

alternativesOptimism The confidence that things will hReinforcement Increasing the probability of a re

response and a given stimulusSkills An ability or proficiency acquiredSocial influences Those interpersonal processes thSocial/professional role and identity A coherent set of behaviors and d

Methods

Study Design

We conducted semistructured interviews, based on the TDF, with arange of healthcare workers and family members involved in the careof NH residents with dementia, in Cork, Ireland. Ethics approval wasgranted by the local ethics committee. The consolidated criteria forreporting qualitative research (COREQ) statement guided studyreporting (Supplementary Table S1).17 Two Patient and PublicInvolvement (PPI) advisory groups composed of 4 people with de-mentia in 1 group, and 2 family members in the other group, providedinput into topic guide development and recruitment. Advisor eligi-bility criteria included being a person with dementia affiliated withthe Alzheimer Society of Ireland or a family member of any NH resi-dent with dementia, and having an interest in research aimed atimproving the quality of medication usage in NHs. Written informedconsent was obtained from all advisors.

Study Setting and Sampling

NHs were chosen as the focus of this study because the prevalenceof antipsychotic use is highest in these settings.18,19 Participants werepurposively sampled to ensure a heterogeneous groupwithmaximumvariation according to 2 main predetermined criteria (professional/social role and NH type) (Supplementary Table S2). We also usedsnowball sampling to fulfil our sampling framework requirements.

Six different NH sites were selected based on our samplingframework, through publicly available directories of registered NHs onthe Health Information and Quality Authority (HIQA)20 and NursingHome Ireland websites.21 The Directors (Nursing or Medical) of eachNH were contacted about the study. Once access was agreed, the Di-rector and other consenting participants connected to that NH wereinterviewed. The Directors approached family members initiallybefore recommending that they were suitable to be contacted.

Eligibility criteria for healthcare workers included being a physi-cian [general practitioner (GP), geriatrician, or psychiatrist of old age],a nurse, a pharmacist or a healthcare assistant (HCA) who wasinvolved in the care of NH residents with dementia. Eligibility criteriafor family members included being a relative of a NH resident withdementia (alive or deceased) who had been prescribed an antipsy-chotic for BPSD.

hanging objectively observed or measured actionsvalidity about an ability, talent, or facility that a person can put to constructive user validity about outcomes of a behavior in a given situationving experiential, behavioral, and physiological elements, by which the individualally significant matter or eventtuation or environment that discourages or encourages the development of skillsial competence, and adaptive behaviores or end states that an individual wants to achieve

a behavior or a resolve to act in a certain waysomethingfocus selectively on aspects of the environment and choose between 2 or more

appen for the best or that desired goals will be attainedsponse by arranging a dependent relationship, or contingency, between the

through practiceat can cause individuals to change their thoughts, feelings, or behaviorsisplayed personal qualities of an individual in a social or work setting

K.A. Walsh et al. / JAMDA 19 (2018) 948e958950

Data Collection

We developed separate topic guides for healthcare professionals,HCAs, and family members. Topic guides were iteratively developedusing findings from our systematic review,14 the TDF, advisor recom-mendations, and 5 pilot interviews. The topic guides underwent re-visions throughout the study (Supplementary Table S3) to ensure thatemerging themes were captured in subsequent interviews. All in-terviews were conducted by the primary author. Written informedconsent was obtained prior to interviews. All interviews were audio-recorded and transcribed verbatim. The author wrote detailed fieldnotes immediately after interviews to refine topic guides and informdata analysis. We sampled until no new ideas emerged and conducted3 more interviews without any new ideas emerging to ensure thatdata saturation had been reached.22 The interviews were conductedbetween July 2016 and April 2017.

Data Analysis

Data analysis followed the principles of framework analysis23 andused NVivo version 11 (QSR International, Melbourne, Australia).24

Data collection and analysis phases occurred concurrently to enablethe exploration of emergent themes in subsequent interviews and toidentify when data saturation occurred.22 We used both inductive anddeductive approaches to analysis. A detailed description of the anal-ysis is available online (Supplementary Material S4). In summary, wefamiliarized ourselves with each transcript and coded emergingconcepts inductively. Simultaneously, we coded data from the tran-scripts into 1 or more TDF domains according to the definitions for

"A FiBalan

Power Knowle

ComplStakeho

Dynam

Human Suffering

Knowledge of the resident

takes precedence

Demen�a causes

suffering

The medica�on

causes suffering

Fig. 1. Conceptual model of explanatory themes: Opposing perspectives and trade-offs (in wprescribe vs not prescribe). The perspective of each NH toward these 3 explanatory themesand benefits of antipsychotics.

each domain (Table 1). We then created distilled summaries of eachinterview to identify the predominant TDF domains and the de-terminants (ie, barriers and facilitators) of the target behaviors(appropriate requesting and prescribing).16 Finally, we developed aconceptual model of explanatory findings, by exploring possible re-lationships between determinants, predominant domains, categories,and theory (Figure 1). In essence, the behavioral determinants werethe “building blocks” for the explanatory themes, and an overarchingtheme was identified, explaining the relationship between de-terminants and explanatory themes. The research group (consisting ofpharmacists, a GP, a health psychologist, a methodologist, and ageriatrician) held regular meetings throughout the study to discussdifferences in interpretation and to identify themes.

Results

We invited 6 NHs to participate and 4 agreed: 2 private NHs,1 withand 1 without a dementia special care unit (SCU); 1 voluntary NH(state-funded but charitable organization governance) without a SCU;and 1 public NH (state-run) without a SCU. Of 38 individuals con-tacted, 27 agreed to participate (8 nurses, 5 GPs, 5 HCAs, 3 familymembers, 2 pharmacists, 2 consultant geriatricians, and 2 consultantpsychiatrists of old age) (Table 2). The median interview length was23 minutes (range 12e56 minutes).

We identified 9 predominant TDF domains, encompassing 38behavioral determinants that influenced our target behaviors(Table 3). We also developed 3 explanatory themes and 1 over-archingtheme, which are discussed below and illustrated in a conceptualmodel (Figure 1). The 9 predominant TDF domains and the more

ne ce"

and dge: ex lder ics

The Interface Between

Resident and Nursing Home

Impact of the resident on the nursing home

Impact of the nursing home

on the resident

Posi�on of power takes precedence

hite) can tip the “fine balance” in favor of undertaking one behavior over another (eg,(in blue), determines how they strike a “fine balance” (in the center) between the risks

Table 2Characteristics of Interview Participants (n ¼ 27)

Characteristics of Total Participants (n ¼ 27) Participants, n

Professional/social roleNurse 8GP 5HCA 5Family member 3Pharmacist 2Consultant geriatrician 2Consultant psychiatrist of old age 2

GenderFemale 17Male 10Other 0

Category of NH participant worked in*Private only 9Public only 4Voluntary only 3Multiple 8

Years of professional experience (since qualification)*<10 y 310e19 y 10�20 y 10Information not provided 1

Received specialist dementia training*Yes 16No 8

Presence of dementia SCU in any NH participant worked in*Yes 7No 17

Characteristics of Family Member Participants (n ¼ 3) Participants, n

GenderFemale 2Male 1Other 0

Category of NH person with dementia resides/residedPrivate 3

RoleCurrent carer 1Former carer 2

Age of participant40e49 y 150e59 y 160e69 y 1

Relationship to person with dementiaSon/daughter 2Nephew/niece 1

*N/A for n ¼ 3 family members.

K.A. Walsh et al. / JAMDA 19 (2018) 948e958 951

seminal determinants are discussed below; detail on the remainingdeterminants is presented in Table 3.

Predominant TDF Domains

Behavioral regulationParticipants believed that HIQA, the independent NH regulator in

Ireland, has put antipsychotics under scrutiny. Regulation now re-quires NHs to notify HIQA on a quarterly basis of any occasion whenrestraint (chemical or physical) is used.25 Some participants believedthat these regulationsmade them re-evaluate how theymanage BPSD,with positive outcomes for residents.

“I think HIQA is brilliant. because I really think they forcepeople to look at their practice, and to challenge their ownpractice and to change.” [HCA 1]

However, GPs in particular felt that there was over-regulation byHIQA, resulting in increased administrative burden, which did notnecessarily translate into good care.

Furthermore, some participants were confused by the regulatoryrequirements and were concerned about unintended negative con-sequences because of the mistaken belief that only psychotropicmedications used for acute episodes were reportable.

“Now, conversely, what it has made some nursing homes do is, ifsomebody was on a PRN psychotropic, because the residentmight only need it once or twice per month and because it be-comes reportable, they get prescribed regularly.” [Nurse 5]

Healthcare workers reported that interdisciplinary medicationreviews, audits, and internal registries also provided an opportunityfor self-monitoring. When in place, these systems assisted with theidentification of patterns of inappropriate usage. Prescribers foundinternational guidelines helpful in their decision-making.2 However,succinct guidelines specific to the Irish context were sought.

Beliefs about capabilitiesParticipants struggled to find solutions to BPSD other than anti-

psychotics in part because they felt that they lacked necessarytraining. NH staff struggled with the daily management of BPSD andsome admitted that they needed antipsychotics to cope. GPs often feltout of their comfort zone and regularly needed input from specialists.

“In someways I don’t feel I have the sufficient expertise to makethose decisions so I’ll look to specialists at that point if I’mstruggling with something.” [GP 3]

Beliefs about consequencesBoth healthcare workers and family members were worried about

side effects such as sedation and falls. Some viewed these side effectsas undignified and inhumane, and hence were reluctant to request orprescribe antipsychotics.

A fear of negative consequences (ie, adverse behavioral eventsfrom residents) if antipsychotics were not prescribed was expressedby prescribers. They were conscious of the safety of their NH col-leagues who were often at the receiving end of behaviors.

“Because you don’t know what precipitated the [behavior], andthen, whenyou’re trying to pull back and youwalk away, are youleaving your colleagues in the height of it then?” [GP 4]

EmotionParticipants, particularly family and NH staff, spoke emotively

about BPSD, and how these symptoms deeply impacted thempersonally. Sometimes participants believed that antipsychotics werethe only solution to alleviating this distress.

“It was very hard to listen to [the BPSD]. so as far as I’m con-cerned, if there was a medication that would sort this thinganyway, I certainlywas completelyopen to it.” [Familymember2]

NH staff were deeply affected by behaviors leading to burn-out,frustration, and poor morale. Staff sometimes took behaviors person-ally, which could increase the propensity to request prescribing of anti-psychotics. Empathy as opposed to sympathy was viewed as animportant trait when dealing with BPSD. It was seen to be important tobe able to stepback, evaluate thesituation anddetermine thebest courseof action for the resident, without emotions clouding one’s judgement.

“I feel that certain people take huge offense if a person who iscognitively impaired lashes out, punches, screams, whatever,and you have to let it go.” [Nurse 8]

Environmental context and resourcesThe overall picture was one of poor resources in NHs. Although

nonpharmacologic interventions were generally seen as the gold

Table 3Determinants of Appropriate Antipsychotic Prescribing Behaviors (Requesting and Prescribing)

Predominant TDF Domain Determinants (ie, Barriers and/or Facilitators) ofAppropriate Antipsychotic Prescribing Behaviors

Illustrative Quotes

Behavioral regulation� HIQA regulation as a stimulus for change (facilitator) � I think HIQA is brilliant. because I really think they

force people to look at their practice, and to challengetheir own practice and to change.” [HCA 1]

� Perception of HIQA over-regulation by GPs (barrier) � “I think HIQA are a scurge. I wonder what they bringto the table. I think they’re self-fulfilling. Ya I thinkmost GPs would not [be happy with them]. I don’tthink they bring a whole lot to the table unfortu-nately. I think they bully private nursing home andprivate institutions.Ya I think it’s all very, very goodand ivory tower stuff and politically correct. But,could I think [sic] the money spent on HIQA could bespent better on direct services? Probably.” [GP 1]

� Uncertainty regarding HIQA reporting requirements(barrier)

� “Now, conversely, what it has made some nursinghomes do is, if somebody was on a PRN psychotropic,because the resident might only need it once or twiceper month and because it becomes reportable, theyget prescribed regularly.” [Nurse 5]

� Self-monitoring (using local systems) of antipsy-chotic prescribing (facilitator)

� “So, for me it would be to monitor the scripts as theycome in and maybe their charts and we do at therequest of the Director of Care, we do a psychotropicaudit every month. So we see where they’re beingreviewed.” [Pharmacist 2]

� Guidelines for monitoring the appropriateness ofantipsychotic prescribing (facilitator)

� “Guidelines is a good thing, and licensing, becauseyou know there isn’t any license. Grade 1, grade 2evidence, meta-analyses. You can certainly usethem to say why you’re not prescribing an antipsy-chotic. You just say there’s no evidence and it’s notnational policy.” [Consultant Psychiatrist of Old Age2]

Beliefs about capabilities� Poor self-efficacy in the management of BPSD among

nonspecialists (barrier)� “So I suppose in some ways I don’t feel I have the

sufficient expertise to make those kind of decisions soI’ll look to specialists at that point if I’m strugglingwith something.” [GP 3]

� Belief that assessing whether an antipsychotic pre-scription is appropriate or not is challenging (barrier)

� “It’s a difficult one to decipher. When it’s appropriateand when it’s not appropriate.” [Nurse 6]

� Belief that deprescribing antipsychotics is difficult(barrier)

� “And it’s very easy starting these things but thediscontinuation of them is not quite so clear cut.”[Consultant Geriatrician 2]

Beliefs about consequences� Concerns about side effects (facilitator) � “She was just asleep looking, absolutely drugged out

of her tree looking, sitting in a chair.” [Family mem-ber 1]

� Belief that antipsychotics are highly effective(barrier)

� “I know the drugs can fix these things. Now notcompletely right. But I know that drugs can fix thesethings.” [Family member 2]

� Belief that NPIs are not a feasible alternative (barrier) � “But if you have somebody at 2 o clock in themorningthat you’re pacing the floor with until 6 o clock in themorning, where are your therapies then?” [HCA 2]

� Belief that the return of symptoms are caused by thereduction of antipsychotic dosage (barrier)

� “I think people often think, that if something doesn’twork straight way or if there happens to be a coin-cidental problem as soon as you start to reduce it,suddenly there is this complete fear that this hascaused it they expect more immediate, they see theimmediate things as being either absent or present sowhen you start a new drug if it hasn’t worked straightaway there is a bit of ‘oh it’s not working.’” [GP 3]

� Anticipated regret (barrier) � “Because you don’t know what precipitated the[behavior], and then, when you’re trying to pull backand you walk away, are you leaving your colleaguesin the height of it then?” [GP 4]

Emotion� Fear of dementia (barrier) � “It was very hard to listen to [the BPSD]. so as far as

I’m concerned, if there was a medication that wouldsort this thing anyway, I certainly was completelyopen to it.” [Family member 2]

� Taking behaviors personally (barrier) � “I feel that certain people take huge offence if a per-son who is cognitively impaired lashes out, punches,screams, whatever and you know, you have to let itgo.” [Nurse 8]

(continued on next page)

K.A. Walsh et al. / JAMDA 19 (2018) 948e958952

Table 3 (continued )

Predominant TDF Domain Determinants (ie, Barriers and/or Facilitators) ofAppropriate Antipsychotic Prescribing Behaviors

Illustrative Quotes

� Burn-out and frustration (barrier) � “You’ll get staff who are burned out, they just can’tcope. They’re sick of saying X, Y, and Z and they’re notbeing listened to, and they just don’t care anymore.”[Nurse 3]

� Empathy toward people with dementia (facilitator) � “I think people with a very empathetic view of de-mentia would be less likely to encourage, prescrip-tion of antipsychotics, because there is that, ‘oh it’s,you know, you don’t have to give them drugs for it,it’s just their dementia, we can get around it,’ andthen, some people. will see the more negative sideof the dementia, and be like, ‘isn’t it awful for them,God wouldn’t you just give them something to relaxthem.’ [Nurse 6]

� Emotions of healthcare professionals tend to reflectthose of family members (barrier)

� “I’ll get [a phone call], ‘The family were in todaythey’re very worried about mammy. She’s very upsetand agitated’. I never get those phone calls to say thatthey’re worried that’s she’s just sitting there staringinto space.” [GP 1]

� Personal experience of dementia (barrier/facilitator)* � “We’re all human, we all bring our own stuff.” [HCA 3]Environmental context and resources

� Lack of adequate resources (barrier) � “You need to have the time to be with somebody,staffing levels don’t really give you the opportunity tosit with somebody all day long or all afternoon. youcan come and go but you can’t stay with the person.”[Nurse 4]

� Perception that it’s cheaper to give antipsychoticsthan deliver NPIs (barrier)

� “They haven’t enough staff and they seem to thinkthat the cheapest way is to dose them, and keep themquiet” [Family member 1].

� Impact of the built environment on the person withdementia (facilitator/barrier)*

� “I think if we had properly designed purpose builtmodern dementia units that allowed us to offer adifferent environment than the standard ward envi-ronment. I do think that would be far more humaneand you’ll probably get better overall results thanresorting to the old fashioned chemical restraints.”[Consultant geriatrician 2]

� Each NH is different (facilitator/barrier)* � “You go to different nursing homes and attitudes arevery different.” [Nurse 3]

� Impact of treatment culture on residents (facilitator/barrier*)

� “Sometimes it can feel like the person is there as. Idon’t know how to say this politely, but they’re in thebed and they have to acquiesce or be compliant withthe system around them, be good children or goodgrown-ups and play the game. And if you don’t dothat, then you get labeled and your behavior getslabeled.” [Consultant Psychiatrist of Old Age 1]

Knowledge� Knowledge of antipsychotics (facilitator) � “If you can tell someone what the potential compli-

cations [of antipsychotics] are, they may be a little bitless likely to ask for them.” [GP 1]

� Knowledge on the cause and nature of BPSD(facilitator)

� “I think if people understood. why [residents] havebehaviors that challenge I think that would go a longway for a lot more understanding and people notwanting just to sedate somebody.” [Nurse 3]

� Knowledge of the resident (facilitator) � “I think just knowing the person. Knowing that theyhave been on them [antipsychotics] for years. Look-ing at them now, their state of deterioration and youknow in your heart and soul they don’t need them.”[Nurse 5]

Memory, attention and decision-processes� Decision-making based on a thorough assessment

(facilitator)� “Then with the physical as well, we do the PINCH ME

acronym so we. pain, infection, constipation, hy-dration, nutrition, medications, environment, welook at real holistic view of the person and try andrule out any triggers there [sic].” [Nurse 6]

� Paying attention to where the challenge lies withregards to the behavioral symptoms (facilitator)

� “Sometimes it just ultimately again it takes me back,you need to take a step back, who are you treating?Are you treating the carer who wants a certainamount given so somebody is peaceful or a certainamount of investigation is done, or are we treatingthe staff members who are treating the patientbecause they want a peaceful night or a peaceful dayon the ward, or are wemaking a decision to make ourown lives easier. And we just have to take a step backsometimes.” [GP 5]

(continued on next page)

K.A. Walsh et al. / JAMDA 19 (2018) 948e958 953

Table 3 (continued )

Predominant TDF Domain Determinants (ie, Barriers and/or Facilitators) ofAppropriate Antipsychotic Prescribing Behaviors

Illustrative Quotes

Social influences� Social pressure to prescribe (barrier) � “So I feel under pressure to knock this person out,

anesthetize this patient, who they see as, shouldn’t bechallenging. And they’re already completely over-sedated and the staff want them to be even moresedated.” [Consultant psychiatrist of old Age 2]

� Reliance on accurate information from NH staff(facilitator/barrier)*

� “I think people can be a little bit biased in how theycan present a case to you at times to get to the endsthat they want. I know there has been one incidentwhere. a staff member [was overheard] saying ‘surejust tell her she’s had hallucinations.’” [GP 3]

� Modelling of prescribing behavior(facilitator/barrier)*

� “A lot of our learning seems to come from the con-sultations and referrals that we actually see what thepsychiatry of the elderly prescribe in these situations,and we have been led by that, so quetiapine justseems to be one they seem to use.” [GP 5]

� Prevailing culture of care (facilitator/barrier)* � “Medication comes first in Ireland. ‘Give it to them asmuch as possible.’” [Family member 1]

Social/professional role and identity� Advocacy role of NH staff and family members

(facilitator)� “See mom didn’t have a voice, nobody would listen to

her even when she was speaking, she wasn’t listenedto and I was her voice.” [Family member 1]

� Professional identity (facilitator/barrier)* � “It depends on what background you are comingfrom and when you trained, how you view themedications and the use of medications. I think thereis a difference, between the younger generation ofnurses and the older generation of nurses. There ap-pears to be more of a reluctance, I think, in theyounger generation of nurses with giving out, I sup-pose the high risk medications like [antipsychotics].And I think there is a difference there then becauseyou’re not seeing your nursing profession as a med-ical profession, you’re almost a facilitator. andwhenyou see it from that perspective then medication isn’talways the first kind of thing that pops into yourhead.” [Nurse 6]

� Variable sense of responsibility for prescribing de-cisions (facilitator/barrier)*

� “But I suppose it’s up to the prescriber to be able tosort the wheat from the chaff and see what’s a goodgrounded opinion and what’s maybe not as reliableyou know.” [HCA 3]

� Leadership role of NH manager (facilitator) � “You need a manager who is supporting staff and isknowledgeable and roles out good training to thestaff. And has good experience, and ideally goodmental health experience because not all of themhave good mental health experience but it is impor-tant for the manager, if you meet the manager, youcan usually see the tone of the home.” [ConsultantPsychiatrist of Old Age 2]

� Traditional hierarchy (barrier) � “As it stands and we’re talking about the real world,it’s really the nursing staff and the GP. I don’t have aninfluence there. If I get the script, we just have tohand it over.” [Pharmacist 2]

NPIs, nonpharmacologic interventions.*This determinant could be a barrier or a facilitator depending on the individual circumstance.

K.A. Walsh et al. / JAMDA 19 (2018) 948e958954

standard, there was consensus that these interventions were staff-intensive and not always feasible.

“You need to have the time to be with somebody, staffing levelsdon’t really give you the opportunity to sit with somebody allday long or all afternoon. you can come and go but you can’tstay with the person.” [Nurse 4]

The physical environment was believed to have a profound impacton residents. Some participants believed that if the environment wasbetter suited to meet the needs of the resident, then there would beless of a need to prescribe.

“I think if we had properly designed purpose-built modern de-mentia units that allowed us to offer a different environmentthan the standard ward environment. I do think that would be

far more humane and you’ll probably get better overall resultsthan resorting to the old fashioned chemical restraints.”[Consultant geriatrician 2]

Participants described how treatment culture impacted theresident in terms of prescribing, both positively (eg, beingresident-centered) and negatively (eg, being task-orientated).There was a general agreement that every NH was completelydifferent, and what may be acceptable in 1 NH may not beacceptable in another.

KnowledgeBoth healthcare workers and family members were aware that

antipsychotics cause side effects. However, nonconsultants in

K.A. Walsh et al. / JAMDA 19 (2018) 948e958 955

particular acknowledged their own limited knowledge on this topic,and welcomed further education. Furthermore, GPs believed that abetter understanding of the risk/benefit profile among NH staff wouldreduce requests for antipsychotics.

“If you can tell someone what the potential complications [ofantipsychotics] are, they may be a little bit less likely to ask forthem.” [GP 1]

In-depth knowledge of the resident was believed to be paramount.Knowing the resident and understanding their life story helped NHstaff to adapt the environment to meet the needs of the resident andoften prevented unnecessary prescribing.

“I think just knowing the person. Knowing that they have beenon them [antipsychotics] for years. Looking at them now, theirstate of deterioration and you know in your heart and soul theydon’t need them.” [Nurse 5]

Memory, attention, and decision processesThe importance of conducting a holistic assessment of the resident

was emphasized by participants. There was agreement that antipsy-chotics were only appropriate after all potential reversible causes ofBPSD were ruled out. In one NH, where a comprehensive assessmentprotocol was recently introduced, nurses explained how this protocolassisted them with their decision-making.

Social influencesPrescribers were based off-site so they relied on accurate and

objective information about residents from nurses. Prescribers largelyvalued and trusted the nurses’ judgements and tended to make pre-scribing decisions based on the information provided. However, thiscould lead to a perception that behavioral symptoms were beingexaggerated to increase the likelihood of prescription.

“I think people can be a little bit biased in how they can presenta case to you at times to get to the ends that they want. I knowthere has been one incident where. a staff member [wasoverheard] saying ‘sure just tell her she’s had hallucinations.”[GP 3]

Prescribers reported that pressure to prescribe antipsychoticsarose from many sources including individual staff members, familymembers, the NH organization, and from society itself.

“So I feel under pressure to knock this person out, anesthetizethis patient, who they see as, shouldn’t be challenging. Andthey’re already completely over-sedated and the staff wantthem to be even more sedated.” [Consultant psychiatrist of oldAge 2]

There was a perception by some of a prevailing culture where allbehaviors may be attributed to the disease rather than an unmet need.However, other participants felt that because of the influence of HIQA,NHs were moving toward a more social model of care. This shift inculture was broadly welcomed. However, some physicians feared thatthe pendulum had “swung too far” [Consultant psychiatrist of old age1], and that GPs, in particular, may be fearful of using antipsychoticsbecause of the perceived antimedication climate.

Social/professional role and identityNH staff and family members viewed themselves as the resident’s

advocate. This role empowered them to speak up on behalf of theresident.

“See mom didn’t have a voice, nobody would listen to her evenwhen she was speaking, she wasn’t listened to and I was hervoice.” [Family member 1]

There was a hierarchy described by participants in the NH envi-ronment. HCAs were often not involved in any degree of decision-making despite their in-depth knowledge of residents. Furthermore,one pharmacist felt disregarded in this area, despite her pharmaco-logic expertise. Decisions were perceived as being made between GPsand nurses, with input from consultants when needed.

“As it stands and we’re talking about the real world, it’s reallythe nursing staff and the GP. I don’t have an influence there. If Iget the script, we just have to hand it over.” [Pharmacist 2]

The importance of leadership from the NH manager was empha-sized. Good leaders were perceived as those with experience whoprovided adequate training and support to staff.

Explanatory Themes

We identified “a fine balance” [HCA 1] as the over-arching theme.On one hand, neither healthcare workers nor family members wantedto see residents oversedated and without a quality of life. Conversely,the reality of needing to protect staff, family members, and residentsfrom potentially dangerous behavioral symptoms, in a resource-poorenvironment, was emphasized. We found that NH staff and pre-scribers struggled with this constant tension throughout their dailypractice.

Beneath the over-arching theme of “a fine balance,” we developed3 explanatory themes as a means of illustrating why this imple-mentation issue, nonadherence to best-practice guidelines, persists.Within these themes, opposing perspectives and trade-offs wereevident which can tip the “fine balance” in favor of undertaking onebehavior over another (eg, prescribe versus not prescribe). We arguethat the perspective of each NH toward these 3 explanatory themes,determines how they strike this “fine balance” (Figure 1).

Human Suffering

Participants described suffering related to both the disease andantipsychotic medications. Some viewed dementia as a terribleaffliction: “I think it’s the hardest disease out there, tomanage. It’s oneI would NOT like to get myself” [HCA 2]. Not only was dementiaperceived to cause suffering to the resident, but often participantsreported being physically and emotionally affected themselves. Anti-psychotics were viewed through this perspective as a way of allevi-ating suffering for everyone. Conversely, others acknowledged thatantipsychotics can cause severe side effects for the resident, and wereused primarily for “staff-focused” [Consultant psychiatrist of old age 2]as opposed to resident-focused purposes. From this perspective, theuse of antipsychotics was frowned upon.

The Interface Between Resident and NH

The perceived effect that the resident has on the NH, and viceversa, was the second explanatory theme. A resident exhibiting BPSDwas perceived by some to have a negative impact on the NH envi-ronment, ultimately requiring additional staff and money: “Theyhaven’t enough staff and they seem to think that the cheapest way isto dose them, and keep them quiet” [Family member 1]. From thisperspective, antipsychotics were perceived as necessary to enable staffto care for all residents in an efficient manner. Conversely, the NHenvironment was perceived by others to have an important impact onthe resident. From this perspective, placing the resident in “the rightplace” [Nurse 3] (ie, a more dementia-friendly environment) wasperceived to be more beneficial to the resident than any medication.

K.A. Walsh et al. / JAMDA 19 (2018) 948e958956

Power and Knowledge: Complex Stakeholder Dynamics

The final theme refers to the complex interplay between the manydifferent stakeholders involved in the care of residents. The symbioticconcepts of power and knowledge can help us to understand thesecomplex stakeholder dynamics. There were different types ofknowledge valued by participants: knowledge of the disease, the drug,and the resident. Often primacy was given to the latter. Hence, fromthis perspective, nurses’ in-depth knowledge of residents legitimizedtheir power to request that an antipsychotic be started or stopped:“The GPs will do it [deprescribe], no problem, we need to instigate it,and it’s just the experience of knowing the person” [Nurse 5].Conversely, others argued that those in higher positions of power hadknowledge that was more important (ie, knowledge of drug and dis-ease) in determining the best outcomes for residents: “Old age psychusually make a recommendation and then the GP will sign the pre-scription” [Nurse 8]. From this perspective, those in positions of powerwere perceived to have themost important knowledge in determiningthe appropriateness of antipsychotic prescribing.

Discussion

Using a novel multiperspective approach, we have generated adeeper understanding of the behavioral components of antipsychoticuse in NH residents with dementia, the professional interactions thatoccur between different stakeholders, and the determinants ofimplementation of best-practice guidelines. Our findings highlighthow implementing evidence-based practice in this area remains asignificant challenge, despite advances in knowledge and stricterregulations. We identified that stakeholders strive to strike “a finebalance” but ultimately, as humans, are influenced by interactingemotional, environmental, organizational, and societal issues.

Comparison with Previous Research

This study builds on the findings of our systematic review14 wherewe identified 5 key concepts influencing decision-making: organiza-tional capacity; individual professional capacity; communication andcollaboration; attitudes; and regulations and guidelines. In this cur-rent study, we found all of these concepts also play a role in imple-menting evidence-based practice. With regards to organizationalcapacity, the fundamental issue of inadequate resources was discussedin almost all of our interviews. This current study also extends ourunderstanding of the influence of regulations on practice. Our studyconfirms the important role of regulations, but also highlights unin-tended negative consequences that may occur as NHs undertakevarious workarounds. Similar workarounds have been reported in theUS, where increasing diagnoses of schizophrenia in NH residents havebeen observed, in a suspected attempt to exempt antipsychotics fromregulatory reporting requirements.26

We identified 9 TDF domains that influenced our target behaviors,which are similar to those found in previous TDF studies exploringprescribing behaviors for various conditions.27e31 The key difference isour identification of “emotion” as a predominant domain which isabsent in themajority of other prescribing studies.27e30 The emotionalimpact of BPSD on family members32 and NH staff33 is established inthe literature. The concept that people with dementia inevitably losetheir identity to dementia and, thus, become “dehumanized” has beenhypothesized as a rationale for why family members often strugglewith BPSD.32 In our study, this fear of dementia emerged as animportant issue. It is evident that this impacts not only family mem-bers, but also NH staff. Prescribers believe that sometimes it is chal-lenging to decipher who precisely is distressed by the BPSD.

Foucault wrote that power and knowledge are not independent en-tities but are inextricably linkedd “knowledge is always an exercise of

power and power always a function of knowledge.”34 This theory mayhelp us to understand the complex dynamics between hierarchicalstakeholders and how different types of knowledge are valued bydifferent stakeholders. Knowledgeof the resident tends tobeprioritized,and sometimes this can contradict treatment goals set by those inhigherpositions of power (with different types of knowledge). Hence, advo-cating on behalf of the resident, particularly by nurses, is central todecision-making and a key target for potential intervention.35,36

Previous studieshaveexplored the challengesGPsexperiencewhenmanaging BPSD.37e39 Jennings et al identified 3 main challenges: lackof clinical guidance; stretched resources; and difficulties managingexpectations.37 Our study corroborates these findings by highlightingthe multitude of difficulties GPs face when deciding whether to pre-scribe antipsychotics or not. However, our study goes further byexploring theperspectives of awider rangeof stakeholders, allowingusto gain a more holistic insight into this implementation problem.

Implications

It is evident that greater policy and institutional support is requiredto help stakeholders strike that “fine balance” and ultimately makebetterprescribingdecisions.Developmentofnational clinical guidelinesmay be one appropriate policy intervention. Such guidelines arecurrently being developed in Ireland as a priority action point of thenational dementia strategy.40 An important implication of our study isthe need to clarify existing regulations for stakeholders, as it is evidentthat theyareunsureas towhichprescribingscenariosare reportable andwhicharenot, and residentsmaybeadverselyaffectedby this confusion.

Further consideration should also be given to the design of futureNHs. Our findings highlight the importance stakeholders attribute todementia SCUs in terms of meeting the needs of residents with de-mentia. However, resident outcomes from SCUs have been mixed,along with concern over higher levels of antipsychotic usage.41,42

Therefore, although SCUs may be desired by stakeholders, more evi-dence of the quality and safety of this approach is required beforewidespread adoption.

The perceived impact of treatment culture on antipsychotic usagefeatured heavily throughout this study. In line with previous sys-tematic review findings,14,43 the NH managers were seen as a keydeterminant of NH treatment culture, as they possessed both a posi-tion of power and knowledge of the resident. We recommend that NHmanagers take advantage of their influential role by providing/orga-nizing ongoing training to staff as well as encouraging the involve-ment of peripheral stakeholders (ie, HCAs, pharmacists, familymembers) in decision-making.

Despite guidance on avoiding antipsychotics in dementia, they canplay an essential role in certain situations.1,2 Our study shows thatbecause of the stigma attached to antipsychotics, some prescribers arefearful of prescribing them at all, risking unnecessary distress for aresident for whom the medications are indicated. A recent studydemonstrated that discontinuation of antipsychotics, without non-pharmacologic substitution, can have a detrimental impact on resi-dents’ health-related quality of life.44 Our findings suggest that anevidence-based, standardized approach involving interdisciplinarycollaboration, careful documentation, and regular review is needed toensure the most appropriate use of both pharmacologic and non-pharmacologic interventions.45 One such model program is the DICE(describe, investigate, create, and evaluate) approach, which promotesa holistic, person-centered approach to managing BPSD.45,46

Educational programs are themost common intervention typeusedto tackle inappropriate antipsychotic prescribing10 (eg, the OASISprogram,47 the Halting Antipsychotic use in Long-Term care [HALT]study,48 and the Reducing Use of Sedatives [RedUSe] project).11

Ongoing education and training to both NH staff and prescribers is animportant aspect of ensuring appropriate antipsychotic prescribing,

K.A. Walsh et al. / JAMDA 19 (2018) 948e958 957

but is not sufficient on its own. Drawing from existing pro-grams11,45,47,48 as well as our own findings, we recommend that futureprograms should include training on the assessment andmanagementof BPSD, dealing with emotions, and managing expectations. It isimportant for prescribers to be empathetic and acknowledge theemotional and physical impact of BPSD, while assertively conveying,the limited benefit and serious risks associated with antipsychotics.Likewise nurses, as the key influencer on prescribing, should be awareof and communicate these issues to otherswithin the NH and to familymembers. In particular, the OASIS communication training programenforces these key messages.47 Future research should focus ondetermining how best to deliver educational interventions, andalongside what, to achieve sustainable results.

Strengths and Limitations

The trustworthiness of our findings are underpinned by theinvolvement of different disciplines within our research team and theparticipation of multiple stakeholders from different organizationsduring the interviews. Triangulation of analysts and participants alsocontributed toward the credibility of the results. Interviews took placein one region in Ireland, but transferability is supported by the provi-sionof sufficient contextual information toenable readers todeterminehow applicable our findings are to their own situation. Detailedreporting of well-established methods with diagrammatical audit-trails contributed toward the dependability of our findings. Finally, interms of confirmability, detailed reporting of participants’ quotationshelped ensure that our findings were primarily borne from the data.49

Although 66% (4/6) of NHs and 71% (27/38) of individuals agreed toparticipate in our study, it is possible that only thosewith strong viewson this topic took part. Furthermore, although we employed a pur-posive sampling approach, Directors may have recommended in-dividuals for participation who were more likely to provide favorableresponses about practices in their NH. Hence, the possibility of se-lection bias cannot be excluded. Random sampling of participantsalong with a larger sample may have reduced this problem and mayhave allowed us to explore differences in perceptions betweenrespondent groups and settings in greater detail.49

Another limitation was the small number of family membersrecruited. The challenges of recruiting family members of residentswith dementia to research studies have been previously reported.50

Despite engaging with our advisors on this issue and reminding Di-rectors to identify potential participants, we only managed to recruit 3family members. It is possible that family members were apprehen-sive about taking part because of the emotive nature of this topic.Furthermore, it is possible that the Directors may have been over-protective of family members.

Conclusions

Implementing evidence-based antipsychotic prescribing practicesfor NH residents with dementia remains a significant challenge, despiteadvances in knowledge and stricter regulations. In striving to strike “afine balance”, stakeholders are influenced by interacting emotional,environmental, organizational, and societal issues. Greater policy andinstitutional support is required to help stakeholders strike that “finebalance” and ultimately make better prescribing decisions. This studyprovides uswith a deeper understanding of this complex issue andwillinform the development of a theory and evidence-based intervention.

Acknowledgments

The authors wish to thank all participants who kindly participatedin this qualitative study. In additionwewish to extend our gratitude toCarmel Geogheghan, Dr Emer Begley, Dr Bernadette Rock, the Irish

DementiaWorking Group, the Alzheimer Society of Ireland, and to ourPPI advisory groupmembers for their helpful contributions.Wewouldalso like to thank Dr Justin Presseau and Dr Andrea Patey, OttawaHospital Research Institute for their advice on the analysis.

The investigators were solely responsible for the design, methods,subject recruitment, data collections, analysis and preparation of pa-per and the funding sources did not participate in this process.

References

1. Reus VI, Fochtmann LJ, Eyler AE, et al. The American Psychiatric Associationpractice guideline on the use of antipsychotics to treat agitation or psychosis inpatients with dementia. Am J Psychiatry 2016;173:543e546.

2. National Institute for Health and Clinical Excellence (NICE). Dementia: Assess-ment, management and support for people living with dementia and theircarers. 2018. Available at: https://www.nice.org.uk/guidance/ng97. AccessedJune 26, 2018.

3. Maust DT, Kim HM, Seyfried LS, et al. Antipsychotics, other psychotropics, andthe risk of death in patients with dementia: Number needed to harm. JAMAPsychiatry 2015;72:438e445.

4. Schneider LS, Dagerman KS, Insel P. Risk of death with atypical antipsychoticdrug treatment for dementia: Meta-analysis of randomized placebo-controlledtrials. JAMA 2005;294:1934e1943.

5. Hsu W-T, Esmaily-Fard A, Lai C-C, et al. Antipsychotics and the risk of cerebro-vascular accident: A systematic review and meta-analysis of observationalstudies. J Am Med Dir Assoc 2017;18:692e699.

6. Szczepura A, Wild D, Khan AJ, et al. Antipsychotic prescribing in carehomes before and after launch of a national dementia strategy: An obser-vational study in English institutions over a 4-year period. BMJ Open 2016;6:e009882.

7. Janus SI, van Manen JG, IJzerman MJ, et al. Psychotropic drug prescriptions inWestern European nursing homes. Int Psychogeriatr 2016;28:1775e1790.

8. Westbury J, Gee P, Ling T, et al. More action needed: Psychotropic prescribing inAustralian residential aged care. Aust N Z J Psychiatry; 2018. 0004867418758919.

9. Gurwitz JH, Bonner A, Berwick DM. Reducing excessive use of antipsychoticagents in nursing homes. JAMA 2017;318:118e119.

10. Thompson-Coon J, Abbott R, Rogers M, et al. Interventions to reduce inappro-priate prescribingof antipsychoticmedications inpeoplewith dementia residentin care homes: A systematic review. J Am Med Dir Assoc 2014;15:706e718.

11. Westbury J, Tichelaar L, Peterson G, et al. A 12-month follow-up study of“RedUSe”: A trial aimed at reducing antipsychotic and benzodiazepine use innursing homes. Int Psychogeriatr 2011;23:1260e1269.

12. Monette J, Monette M, Sourial N, et al. Effect of an interdisciplinary educationalprogram on antipsychotic prescribing among residents with dementia in twolong-term care centers. J Appl Gerontol 2013;32:833e854.

13. Michie S, Atkins L, West R. The Behaviour Change Wheel: A Guide to DesigningInterventions. London, United Kingdom: Silverback Publishing; 2014.

14. Walsh KA, Dennehy R, Sinnott C, et al. Influences on decision-making regardingantipsychotic prescribing in nursing home residents with dementia: A sys-tematic review and synthesis of qualitative evidence. J Am Med Dir Assoc 2017;18:897.e.

15. Cane J, O’Connor D, Michie S. Validation of the theoretical domains frameworkfor use in behaviour change and implementation research. Implement Sci 2012;7:37.

16. Atkins L, Francis J, Islam R, et al. A guide to using the Theoretical DomainsFramework of behaviour change to investigate implementation problems.Implement Sci 2017;12:77.

17. Tong A, Sainsbury P, Craig J. Consolidated criteria for reporting qualitativeresearch (COREQ): A 32-item checklist for interviews and focus groups. Int JQual Health care 2007;19:349e357.

18. Walsh KA, O’Regan NA, Byrne S, et al. Patterns of psychotropic prescribing andpolypharmacy in older hospitalized patients in Ireland: The influence of de-mentia on prescribing. Int Psychogeriatr 2016;28:1807e1820.

19. Zhang Y, Letuchy EM, Carnahan RM. Where are antipsychotics prescribed innursing homes initiated? J Am Geriatr Soc 2018;66:1082e1088.

20. Health Information and Quality Authority (HIQA). Health Information andQuality Authority (HIQA). 2016. Available at: https://www.hiqa.ie/find-a-centre. Accessed June 1, 2016.

21. Nursing Home Ireland. Nursing Home Ireland. 2016. Available at: http://www.nhi.ie/. Accessed June 1, 2016.

22. Francis JJ, Johnston M, Robertson C, et al. What is an adequate sample size?Operationalising data saturation for theory-based interview studies. PsycholHealth 2010;25:1229e1245.

23. Ritchie J, Spencer L, O’Connor W. Carrying out qualitative analysis. In: Ritchie J,Lewis J, editors. Qualitative research practice: A guide for social science stu-dents and researchers. London, United Kingdom: Sage; 2003:219e262.

24. QSR International Pty Ltd. NVivo Qualitative Data Analysis Software version11. 2017.

25. Health Act 2007 (Care and Welfare of Residents in Designated Centres for OlderPeople) Regulations 2013, Irish Statute Book, 2013.

26. Westbury J. Antipsychotic drug prescribing in nursing homes. JAMA 2017;318:1829.

K.A. Walsh et al. / JAMDA 19 (2018) 948e958958

27. Fleming A, Bradley C, Cullinan S, et al. Antibiotic prescribing in long-term carefacilities: A qualitative, multidisciplinary investigation. BMJ Open 2014;4:e006442.

28. CadoganCA,RyanC, Francis JJ, et al. Improvingappropriatepolypharmacy forolderpeople in primary care: Selecting components of an evidence-based interventionto target prescribing and dispensing. Implementation Science 2015;10:161.

29. Cullinan S, Fleming A, O’mahony D, et al. Doctors’ perspectives on the barriersto appropriate prescribing in older hospitalized patients: A qualitative study. BrJ Clin Pharmacol 2015;79:860e869.

30. O’Riordan D, Byrne S, Fleming A, et al. GPs’ perspectives on prescribing forolder people in primary care: A qualitative study. Br J Clini Pharmacol 2017;83:1521e1531.

31. Sargent L, McCullough A, Del Mar C, et al. Using theory to explore facilitatorsand barriers to delayed prescribing in Australia: A qualitative study using thetheoretical domains framework and the behaviour change wheel. BMC FamPract 2017;18:20.

32. Feast A, Orrell M, Charlesworth G, et al. Behavioural and psychological symp-toms in dementia and the challenges for family carers: Systematic review. Br JPsychiatry 2016;208:429e439.

33. Brodaty H, Draper B, Low LF. Nursing home staff attitudes towards residentswith dementia: Strain and satisfaction with work. J Adv Nurs 2003;44:583e590.

34. Foucault M. Power/Knowledge: Selected Interviews and Other Writings, 1972-1977. New York, NY: Pantheon; 1980.

35. Mc Gillicuddy A, Crean AM, Kelly M, et al. Oral medicine modification for olderadults: A qualitative study of nurses. BMJ Open 2017;7:e018151.

36. Walent RJ, Kayser-Jones J. Having a voice and being heard: Nursing homeresidents and in-house advocacy. J Gerontol Nurs 2008;34:34e42.

37. Jennings AA, Foley T, McHugh S, et al. ‘Working away in that Grey Area.’Aqualitative exploration of the challenges general practitioners experience whenmanaging behavioural and psychological symptoms of dementia. Age Ageing2018;47:295e303.

38. Foley T, Boyle S, Jennings A, et al. “We’re certainly not in our comfort zone”: Aqualitative study of GPs’ dementia-care educational needs. BMC Fam Pract2017;18:66.

39. Jennings AA, Foley T, Walsh KA, et al. General practitioners’ knowledge, atti-tudes and experiences of managing behavioural and psychological symptomsof dementia: A mixed methods systematic review. Int J Geriatr Psychiatry 2018;33:1163e1176.

40. Department of Health. The Irish national dementia strategy 2014. Available at:http://health.gov.ie/blog/publications/the-irish-national-dementia-strategy/.Accessed June 18, 2018.

41. Cioltan H, Alshehri S, Howe C, et al. Variation in use of antipsychotic medica-tions in nursing homes in the United States: A systematic review. BMC Geriatr2017;17:32.

42. Gruneir A, Lapane KL, Miller SC, et al. Is dementia special care really special? Anew look at an old question. J Am Geriatr Soc 2008;56:199e205.

43. Sawan M, Jeon Y-H, Chen TF. Relationship between organizational culture andthe use of psychotropic medicines in nursing homes: A systematic integrativereview. Drugs Aging 2018;35:189e211.

44. Ballard C, Orrell M, Sun Y, et al. Impact of antipsychotic review and non-pharmacological intervention on health-related quality of life in people withdementia living in care homes: WHELDda factorial cluster randomisedcontrolled trial. Int J Geriatr Psychiatry 2017;32:1094e1103.

45. Kales HC, Gitlin LN, Lyketsos CG, et al. Management of neuropsychiatricsymptoms of dementia in clinical settings: Recommendations from a multi-disciplinary expert panel. J Am Geriatr Soc 2014;62:762e769.

46. Kales HC, Gitlin LN, Lyketsos CG. Assessment and management of behavioraland psychological symptoms of dementia. BMJ 2015;350:h369.

47. Tjia J, Hunnicutt JN, Herndon L, et al. Association of a communication trainingprogram with use of antipsychotics in nursing homes. JAMA Intern Med 2017;177:846e853.

48. Brodaty H, Aerts L, Harrison F, et al. Antipsychotic deprescription for olderadults in long-term care: The HALT study. J Am Med Dir Assoc 2018;19:592e600.e7.

49. Shenton AK. Strategies for ensuring trustworthiness in qualitative researchprojects. Educ Iinform 2004;22:63e75.

50. Barry HE, Parsons C, Passmore AP, et al. Pain in care home residents with de-mentia: An exploration of frequency, prescribing and relatives’ perspectives. IntJ Geriatr Psychiatry 2015;30:55e63.

Supplementary Table S1Consolidated criteria for reporting qualitative research (COREQ) Checklist

Domain 1: Research Team and ReflexivityPersonal characteristics1. Interviewer/facilitator Which author/s conducted the interview or focus group? KW conducted the interviews2. Credentials What were the researcher’s credentials (eg, PhD, MD)? At the time of undertaking the interviews KW’s tials were BPharm, MPharm, MPSI3. Occupation What was their occupation at the time of the study? KW is an Irish registered pharmacist, who was un

Research, when this study was conducted.4. Sex Was the researcher male or female? Male5. Experience and training What experience or training did the researcher have? KW completed training in utilization of NVivo so

at Oxford University, UK. KW has also conductqualitative evidence previously.

Relationship with participants6. Relationship established Was a relationship established prior to study commencement? There were no established relationships between

team prior to study commencement.7. Participant knowledge of

the interviewerWhat did the participants know about the researcher (eg,personal goals, reasons for doing the research)?

KW had disclosed to all participants that he was ato conducting the interviews.

8. Interviewer characteristics What characteristics were reported about theinterviewer/facilitator? (eg, bias, assumptions, reasonsand interests in the research topic)

KW is a registered pharmacist with communityconducting this study as part of his PhD explordementia. This information was disclosed to p

In order to minimize the potential for KW’s info(clinical and nonclinical) were involved in thetopic.

Our research team consisted of a broad range of d(CS); a health psychologist (JMcS); a methodolcontributed toward the credibility of the result

Domain 2: Study DesignTheoretical framework9. Methodological orientation

and TheoryWhat methodological orientation was stated to underpinthe study? (eg, grounded theory, discourse analysis,ethnography, phenomenology, content analysis)?

Framework Analysis as described by Ritchie and

Participant selection10. Sampling How were participants selected (eg, purposive,

convenience, consecutive, snowball)?Participants were purposively sampled to ensureto 2 main predetermined criteria (Professional/to fulfil our sampling framework requirements

Six different NH sites were selected based on our sof registered nursing homes on the Health InfoIreland websites.

11. Method of approach How were participants approached (eg, face-to-face,telephone, mail, email)?

The Directors of each NH (Directors of Nursing orand informed about the study, with a follow uaccess, they were interviewed themselves by Kparticipants connected to their nursing home,telephone with information about the study. Aold age, consultant geriatricians and pharmacisthe study. The Directors approached family meKW that they were suitable to be contacted.

12. Sample size How many participants were in the study? 2713. Nonparticipation How many people refused to participate or dropped out? Reasons? Of 6 nursing homes contacted by KW via their re

Of the 4 pharmacists serving the 4 different NH srespond.

Of the 9 GPs serving the 4 different NH sites, 5 pdefinite date for interview and 2 did not respo

Of 10 nurses across the 4 different sites whowereand 1 did not respond.

Of 5 family members who were contacted by KWwith a definite date for interview; 1 initially agrwant to be involved. An unknown number of f

Appendix

creden

dertaking a PhD in Population Health and Health Services

ftware and received qualitative research methods traininged and published a systematic review and synthesis of

any of the 27 participants and the researcher or research

pharmacist undertaking this study as part of his PhD, prior

pharmacy and qualitative research experience, and wasing antipsychotic prescribing in NH residents witharticipants ahead of the interview.rmation bias, entering the analysis, a mix of professionalsanalysis, with varying levels of knowledge on this specific

isciplines: pharmacists (KW, AF, SB); a general practitionerogist (JB); and a geriatrician (ST). Triangulation of analystss, and minimized bias from any one particular researcher.

Lewis, using the TDF as the a priori defined framework.

a heterogeneous groupwithmaximum variation accordingsocial role and NH type). We also used snowball sampling.ampling framework, through publicly available directoriesrmation and Quality Authority (HIQA) and Nursing Home

Medical Directors) were contacted by KW by email initiallyp phone-call if no response. Once the Directors agreedW and they then recommended other potentialwhom KW would approach face-to-face or via email/ll relevant visiting staff (ie, GPs, consultant psychiatrists ofts) serving each of the sites were invited to participate inmbers initially about the study before recommending to

spective Director, 4 participated and 2 did not respond.ites, 2 participated. 1 said they was too busy and 1 did not

articipated. 2 initially agreed but never followed up with and.contacted by KW, 8 participated; 1 said they were too busy

, 3 participated; 1 initially agreed but never followed upeed but then canceled because the rest of the family didn’tamily members were informally approached about the

(continued on next page)

K.A.W

alshet

al./JAMDA19

(2018)948

e958

958.e1

Supplementary Table S1 (continued )

study by the Directors of each NH site, but did not agree to participate.Of 3 consultant geriatricians contacted by KW, 2 participated. 1 initially agreed but never followed upwith adefinite date for interview.

Of 2 consultant psychiatrists of old age contacted by KW, both participated.Of 5 HCAs contacted by KW, all 5 participated.Total nonparticipators: n ¼ 2 nursing homes, n ¼ 11 individuals directly contacted by KW

Setting14. Setting of data collection Where was the data collected (eg, home, clinic, workplace)? All interviews took place either in the participant’s place of work, home or an office in the researcher’s

university, depending on participant’s preference.15. Presence of nonparticipants Was anyone else present besides the participants and researchers? No16. Description of sample What are the important characteristics of the sample

(eg, demographic data, date)?Refer to Table 2

Data collection17. Interview guide Were questions, prompts, guides provided by the authors?

Was it pilot tested?Three types of topic guides were in circulation at any one time. They were broadly similar for content, butdiffered primarily for language:

1 for healthcare professionals (physicians, nurses, and pharmacists),1 for healthcare assistants1 for family members.The topic guides were pilot tested by 5 participants (1 nurse, 1 HCA, 1 pharmacist, 1 GP, and 1 familymember) to ensure appropriate content and language for the different groups. All topic guides wererevised slightly after every pilot interview. Only the latter interview conducted with a family member wassubsequently included in the analysis, as this topic guide was agreed to be close enough to the finalversion.

Throughout the remainder of the study, the topic guides underwent iterative revision to ensure thatemerging themes were captured in subsequent interviews.

18. Repeat interviews Were repeat interviews carried out? If yes, how many? No19. Audio/visual recording Did the research use audio or visual recording to collect the data? All interviews were audio recorded.20. Field notes Were field notes made during and/or after the interview

or focus group?Field notes were written immediately after the interviews, and were referred to during analysis, andrefinement of topic guides.

21. Duration What was the duration of the interviews or focus group? The median interview length was 23 minutes and the range was 12e56 minutes.22. Data saturation Was data saturation discussed? The Francis et al method was used to determine when data saturation had been reached. We sampled until

no new ideas emerged from the interviews and then conducted a further 3 interviews without any newideas emerging to ensure that data saturation had been reached.

23. Transcripts returned Were transcripts returned to participants for commentand/or correction?

No.

Domain 3: Analysis and FindingsData analysis24. Number of data coders How many data coders coded the data? Four (KW, CS, AF, JMcS)25. Description of the

coding treeDid authors provide a description of the coding tree? The TDF was used as a basis for the coding tree

26. Derivation of themes Were themes identified in advance or derived from the data? We utilized both deductive and inductive approaches to analysis throughout the 5 framework stages(familiarization, identifying a thematic framework, indexing, charting, and mapping and interpretation).First KW became familiar with the data by reading and re-reading transcripts and field notes and opencoded across the entire dataset. The 14 TDF domains were then deductively applied systematically to thedata during indexing while emerging concepts were coded and categorized inductively. These indexingsteps were conducted independently by at least 2 authors for 7 transcripts (KW and AF/JMcS), who met todiscuss differences in application of the TDF or interpretation of emerging concepts, and came toconsensus. The indexing of the remaining transcripts was conducted by KW using agreed understandingsof the TDF domains.

Charting of the data, with distilled summaries in matrix format was used to identify the predominantdomains influencing the target behaviors (appropriate requesting and prescribing). This activity wasperformed independently by 2 authors (KW and CS), who then discussed any disagreement untilconsensus was reached. From these predominant domains, the determinants (ie, barriers and facilitators)of the target behaviors were identified by KW, with input from the whole team.

For the final mapping and interpretation step, we iteratively developed links between barriers and

K.A.W

alshet

al./JAMDA19

(2018)948

e958

958.e2

facilitators, predominant domains, initial categories and theory to provide overall explanations for thefindings. This was conducted by KW, with input from the whole research team.

27. Software What software, if applicable, was used to manage the data? NVivo 1128. Participant checking Did participants provide feedback on the findings? No

Reporting29. Quotations presented Were participant quotations presented to illustrate the

themes/findings? Was each quotation identified?eg, participant number

Yes

30. Data and findingsconsistent

Was there consistency between the data presentedand the findings?

Quotes are presented in a manner consistent with findings

31. Clarity of major themes Were major themes clearly presented in the findings? Major (explanatory) themes are presented in the results section.32. Clarity of minor themes Is there a description of diverse cases or discussion

of minor themes?The predominant TDF domains that feed into the major (explanatory) themes are explored in detail in theresults section

K.A.W

alshet

al./JAMDA19

(2018)948

e958

958.e3

Supplementary Table S2Sampling Framework

NH Type Total

Professional/Social Role Private NH Voluntary NH Public NH

GP 2 1 2 5Nurse 5 2 1 8Pharmacist 1 1 0 2HCA 2 2 1 5Family member 3 0 0 3Consultant psychiatry of old age 2* 2Consultant geriatrician 2* 2Total 27

The number in each box refers to the number of participants recruited, according to the 2 main predetermined criteria (professional/social role and NH type).*These participants worked across all NH types (private, voluntary and public)

K.A. Walsh et al. / JAMDA 19 (2018) 948e958958.e4

Supplementary Table S3Evolving Topic Guides

1. Pilot Topic Guides (Version 1):1A. Healthcare professional (HCP)

In your own words, tell me what your views are regarding the prescribing of antipsychotics and other psychotropics to Nursing Home (NH) residents with dementia.(Prompts if necessary: Is it appropriately prescribed in all cases? Is it necessary? What are the benefits and harms?)In the context of NH residents with dementia, how you would you define an “appropriate prescription” of an Antipsychotic or Psychotropic (AP/P)?(Prompts if necessary: What would an “appropriate prescription” entail in terms of indication, drug, dosage, frequency of review, duration, who needs to beconsulted?)

What do you believe to be the main barriers to an “appropriate prescription”?(Rephrase if necessary: What prevents the use of alternative nonpharmacologic approaches in residents who may not necessarily need AP/P medications?)Conversely, what do you believe enables “appropriate prescribing”?(Rephrase if necessary: What facilitates the use of alternative nonpharmacologic approaches in residents who may not necessarily need AP/P medications?)What are your views on nonpharmacologic approaches in residents who exhibit behaviors that challenge, such as massage therapy, reminiscence therapy, and musictherapy?

(Prompts if required: Are they effective? Whose role is it? Are they being used first-line for challenging behavior?)Can you describe your general approach to:prescribing (physician)requesting or administering (nurses)dispensing (pharmacist)

a prescription for AP/P medications to a typical resident with dementia, who may be exhibiting behaviors that challenge?(Prompts if necessary: How would you start this process or journey for a NH resident with dementia? What is the first thing you would always do? Use of NPI? Whatwould you do next? Would you always do this? Anything else?)

Do you believe that everyone involved in the care of residents with dementia (including consultants, General Practitioner (GP), nurses, pharmacists, Healthcare as-sistant (HCA), and family members) are sufficiently aware of the risk/benefits of AP/P prescribing?(Prompts if necessary:Why do you think this? Is there any group of people in particular that you feel could benefit frommore training and education?What specificallydo you think they need to know more about?)[If not mentioned from above] Furthermore do you believe that everyone involved in the care of residents with dementia, knows how to effectively manage someonewho is exhibiting behaviors that challenge?

(Prompts if necessary: Why do you think this? Is there any group of people in particular that you feel could benefit from more training and education? What do youthink they need to know more about?)

In the literature, the importance of “knowing the resident” is emphasized in relation to AP/P prescribing. Do you agree with this statement?(Prompts if necessary: Why/why not? Does it enable a person-centered care approach? What is the best way to “get to know” the resident?)Thinking about AP/P prescribing in NH residents with dementia, what would you consider your responsibilities to be as ensuring that the residents receive anappropriate prescription?

(Prompts if required: Is there anything that you would consider to be beyond your responsibility as ensuring an appropriate AP/P prescription?) [if necessary: Youmentioned X as being a barrier to appropriate prescribing earlier]

What (other) aspects of the environment influence whether or not these agents are appropriately prescribed and reviewed?(Prompts if required: Is there any resource issues that could lead to inappropriate prescribing?)Who or what would influence your decision about whether or not to prescribe an AP/P to a resident with dementia? (physician only)ORHow do you think that your views and opinions, and that of others, influence the prescriber, in relation to AP/P prescribing? (nurses and pharmacists)(Prompts if required: Individuals/groups of HCPs/public opinion/guidelines/dementia strategy. Anyone else?)[If not mentioned already] Do you believe that Health Information and Quality Authority (HIQA) have an important role in influencing AP/P prescribing in the NHsetting? What about the Irish Dementia Strategy?(Prompts if required: HIQA have released new updated Standards with an increased emphasis on restraint use in NH residents with dementia, are you familiar withthem? Any thoughts?)

In the literature, “nursing home culture” comes up a lot and is seen to be important in influencing the AP/P prescribing process. From your own experience, would youcare to comment on this finding?

(Rephrase if necessary: A lot of studies found that “the way things are understood, judged and valued”, within the institution may influence the prescribing process.Please comment)

(Prompts if required: Is there conformity within the NH? Is a person-centered care approach adopted? Does it contribute to increased or decreased usage in youropinion?)

Our research group is interested in developing interventions or strategies to support appropriate and rational use of AP/P in NH residents with dementia. Fromreviewing the literature we have found that interventions are quite complex and often involve a number of different components. This makes it difficult to identifywhich components are the most important in terms of improving patient outcomes and achieving appropriate and rational prescribing.

In an ideal world, what would you consider to be crucial components of an intervention to enable the appropriate and rational use of AP/P in NH residents withdementia?(Rephrase if necessary: If resources were not an issue, what strategies would ensure that residents would not be exposed to AP/P unnecessarily?)(Prompts: Who should be involved in delivering these types of interventions in practice? pharmacists, carers, HCAs, GP, consultants, nurse?)(Prompts: What would they have to do?)(Prompts: In reality, what would be the main barrier to implementing this type of intervention?)(Prompts: What do you think would help with the implementation of this intervention?)What do you think are important outcome measures in such an intervention study?(Rephrase if necessary: What do you think is important to measure, so that people can determine for themselves whether they felt the intervention was a success ornot?)

(Prompts if required: How would you be personally persuaded that such an intervention was successful? Number of AP/P prescribed? QoL? Satisfaction level? No. ofchallenging behaviour?)

In recent times, there has been a push to get members of the public (ie, service users, carers, advocacy groups and patients themselves) actively involved in researchrather than being simply research participants. For example, members of the public, who may have dementia, may join an advisory group for a project.

What are your thoughts on Public Involvement in the designing of an intervention, such as one you have just described?(Prompts if required: Should members of the public be actively involved in the decisions about medicines they or their loved ones may be prescribed? Is it feasible?)That brings us to the end of the interview.Is there anything else I haven’t asked you today that you would like to mention?

1B Healthcare assistant1. In your own words, tell me what your views are regarding the prescribing of antipsychotics (such as Zyprexa, Seroquel) and other psychotropics (such as Xanax,

Lustral, and Stilnoct) to NH residents with dementia.

(continued on next page)

K.A. Walsh et al. / JAMDA 19 (2018) 948e958 958.e5

Supplementary Table S3 (continued )

(Prompts if necessary: Is it appropriately prescribed in all cases? Is it necessary? What are the benefits and harms?)2. In the context of NH residents with dementia, how you would you define an “appropriate prescription” of an AP/P?(Prompts if necessary: What would an “appropriate prescription” entail in terms of indication, drug, dosage, frequency of review, duration, who should be consulted?)3. What do you believe to be the main barriers to an “appropriate prescription”?(Rephrase if necessary: What prevents the use of alternative nonpharmacologic approaches in residents who may not necessarily need AP/P medications?)4. Conversely, what do you believe enables “appropriate prescribing”?(Rephrase if necessary: What facilitates the use of alternative nonpharmacologic approaches in residents who may not necessarily need AP/P medications?)5. What are your views on nonpharmacologic approaches in residents, such as massage therapy, reminiscence therapy and music therapy?(Prompts if required: Are they effective? Whose role is it? Are they being used first-line for challenging behaviour?6. Can you describe your general approach to requesting a prescription for AP/P medications to a typical resident with dementia, who may be exhibiting behaviors

that challenge?(Prompts if necessary: Is that something you would normally do as a HCA? How would you start this process or journey for a NH resident with dementia? What is thefirst thing you would always do? Use of NPI? What would you do next? Would you always do this? Anything else?)7. Do you believe that everyone involved in the care of residents with dementia (including consultants, GPs, nurses, pharmacists, family members, and HCAs) are

sufficiently aware of the risk/benefits of AP/P prescribing?(Prompts if necessary:Why do you think this? Is there any group of people in particular that you feel could benefit frommore training and education?What specificallydo you think they need to know more about?)8. [If not mentioned from above] Furthermore do you believe that everyone involved in the care of residents with dementia, knows how to effectively manage

someone who is exhibiting behaviors that challenge?(Prompts if necessary: Why do you think this? Is there any group of people in particular that you feel could benefit from more training and education? What do youthink they need to know more about?)9. In the literature, the importance of “knowing the resident” is emphasized in relation to AP/P prescribing. Do you agree with this statement?(Prompts if necessary: Why/why not? Does it enable a person-centered care approach? What is the best way to “get to know” the resident?)10. Thinking about AP/P prescribing in NH residents with dementia, what would you consider your responsibilities to be as a HCA in ensuring that the residents

receives an appropriate prescription?(Prompts if required: Is there anything that you would consider to be beyond your responsibility as a HCA in ensuring anappropriate AP/P prescription?)

11. [if necessary: You mentioned X as being a barrier to appropriate prescribing earlier]What (other) aspects of the environment influence whether or not these agents are appropriately prescribed and reviewed?(Prompts if required: Is there any resource issues that could lead to inappropriate prescribing?)12. How do you think that your views and opinions, and that of others, influence the prescriber, in relation to AP/P prescribing? (Prompts if required: Individuals/

groups of HCPs/public opinion/guidelines/dementia strategy. Anyone else?)13. [If not mentioned already] Do you believe that HIQA have an important role in influencing AP/P prescribing in the NH setting? What about the Irish Dementia

Strategy? In what way?(Prompts if required: HIQA have released new updated Standards with an increased emphasis on restraint use in NH residents with dementia, are you familiar withthem? Any thoughts?)14. In the literature, “nursing home culture” comes up a lot and is seen to be important in influencing the AP/P prescribing process. From your own experience, would

you care to comment on this finding?(Rephrase if necessary: A lot of studies found that “the way things are understood, judged and valued,” within the institution may influence the prescribing process.Please comment)(Prompts if required: Is there conformity within the NH? Is a person-centered care approach adopted? Does it contribute to increased or decreased usage in youropinion?)Our research group is interested in developing interventions or strategies to support appropriate and rational use of AP/P in NH residents with dementia. Fromreviewing the literature we have found that interventions are quite complex and often involve a number of different components. This makes it difficult to identifywhich components are the most important in terms of improving patient outcomes and achieving appropriate and rational prescribing.15. In an ideal world, what would you consider to be crucial features of an intervention to enable the appropriate and rational use of AP/P in NH residents with

dementia?(Rephrase if necessary: If resources were not an issue, what strategies would ensure that residents would not be exposed to AP/P unnecessarily?)(Prompts: Who should be involved in delivering these types of interventions in practice? pharmacists, carers, HCAs, GP, consultants, nurse?)(Prompts: What would they have to do?)(Prompts: In reality, what would be the main barrier to implementing this type of intervention?)(Prompts: What do you think would help with the implementation of this intervention?)16. What do you think are important outcome measures in such an intervention study?(Rephrase if necessary: What do you think is important to measure, so that people can determine for themselves whether they felt the intervention was a success ornot?)(Prompts if required: Howwould you be personally persuaded that such an intervention was successful? Number of AP/P prescribed? Quality of Life (QoL)? Satisfactionlevel? No. of challenging behaviour?)In recent times, there has been a push to get members of the public (ie, service users, carers, advocacy groups and patients themselves) actively involved in researchrather than being simply research participants. For example, members of the public, who may have dementia, may join an advisory group for a project.17. What are your thoughts on Public Involvement in the designing of an intervention, such as one you have just described?(Prompts if required: Should members of the public be actively involved in the decisions about medicines they or their loved ones may be prescribed? Is it feasible?)That brings us to the end of the interview.18. Is there anything else I haven’t asked you today that you would like to mention?

1C. Family memberIn your own words, can you describe what your views are towards the use of medications in your loved one with dementia? (Prompts if necessary: have they beenbeneficial? Have you noticed any improvements? Have they caused any side effects?)The focus of my PhD research is on the usage of a group of medications called antipsychotics and another broader group of medications called psychotropics in NHresidents with dementia. Common examples of Antipsychotics include Zyprexa, Seroquel, and Serenace�. Common examples of other psychotropics include Xanax,Halcion, and Lustral. These drugs are sometimes prescribed to people with dementia if they are severely distressed or displaying some behaviors that others may findchallenging such as aggressive or agitated behavior.If you have any experience in the use of these medications in your loved one, whether in the NH setting, in the hospital or at home, I’d be very interested to hear yourstory. (If not, then this is absolutely fine we can still talk about medication use in general) (Prompts if necessary: Why was he/she prescribed these drugs? Can youremember what it was he/she was prescribed? Did it help the situation? Were there any side effects? Is he/she still on it? Who stopped it and why?)From your perspective, what would constitute an “appropriate prescription” of an AP/P?(Prompts if necessary: Who needs to be consulted in the process? What would an “appropriate prescription” entail in terms of indication, drug, dosage, frequency ofreview, duration?)What do you believe to be the main barriers to an “appropriate prescription”?

(continued on next page)

K.A. Walsh et al. / JAMDA 19 (2018) 948e958958.e6

Supplementary Table S3 (continued )

(Rephrase if necessary: What prevents the use of alternative nonpharmacologic approaches in residents who may not necessarily need AP/P medications?)Conversely, what do you believe helps “appropriate prescribing” to occur?(Rephrase if necessary: What facilitates the use of alternative nonpharmacologic approaches in residents who may not necessarily need AP/P medications?)What are your views on alternative approaches in residents who may be displaying distressing or challenging behaviors, such as massage therapy, reminiscencetherapy and music therapy?(Prompts if required: Are they effective? Whose role is it? Are they being used first-line for challenging behaviors?)Have you ever requested a prescription for an AP/P or have you ever requested it to be stopped? If yes, could you describe for me in general what happened?(Prompts if necessary: Why did you do this? Is that something you would normally do as a family member? How would you start this process or journey for a NHresident with dementia? What is the first thing you would always do? Use of Non-pharmacologic intervention (NPI)? What would you do next? Would you always dothis? Anything else?)Do you believe that everyone involved in the care of residents with dementia (including consultants, GPs, nurses, pharmacists, HCAs, and family members) knowenough about these drugs?(Prompts if necessary:Why do you think this? Is there any group of people in particular that you feel could benefit frommore training and education?What specificallydo you think they need to know more about?)[If not mentioned from above] Furthermore do you believe that everyone involved in the care of residents with dementia, knows how to effectively manage someonewho is exhibiting behaviors that challenge?(Prompts if necessary: Why do you think this? Is there any group of people in particular that you feel could benefit from more training and education? What do youthink they need to know more about?)“Knowing the resident” is important in relation to AP/P prescribing. What is the best way for those involved in the care of residents with dementia to “get to know” theresident?(Prompts if necessary: How would this help? Would it enable a person-centered care approach?)Thinking about AP/P prescribing in your loved one with dementia, what would you consider your responsibilities to be as a family member in ensuring that he/shereceives an appropriate prescription?(Prompts if required: Is there anything that you would consider to be beyond your responsibility as a family member in ensuring an appropriate AP/P prescription?)[if necessary: Youmentioned X as being a barrier to appropriate prescribing earlier] What (other) aspects of the environment influence whether or not these agents areappropriately prescribed and reviewed?(Prompts if required: Is there any resource issues that could lead to inappropriate prescribing?)How do you think that your views and opinions, and that of others, influence the prescriber, in relation to AP/P prescribing? (Prompts if required: Individuals/groups ofHCPs/public opinion/guidelines/dementia strategy. Anyone else?)Our research group is interested in developing interventions or strategies to support appropriate and rational use of AP/P in NH residents with dementia. Fromreviewing the literature we have found that interventions are quite complex and often involve a number of different components. This makes it difficult to identifywhich components are the most important in terms of improving patient outcomes and achieving appropriate and rational prescribing.In an ideal world, what would you consider important in an intervention to enable the appropriate and rational use of AP/P in NH residents with dementia?(Rephrase if necessary: If resources were not an issue, what strategies would ensure that residents would not be exposed to AP/P unnecessarily?)(Prompts: Who should be involved in delivering these types of interventions in practice? What would they have to do?)That brings us to the end of the interview.Is there anything else I haven’t asked you today that you would like to mention?

2. Topic Guides (Version 5):2A. Healthcare professionals

In your own words, tell me what your views are regarding the use of antipsychotics and other psychotropics in NH residents with dementia. (Prompts: Is it appro-priately prescribed in all cases? Is it necessary?)In the context of NH residents with dementia, what you would you define as an “appropriate” usage of these agents?(Prompts: indication, frequency of review, duration, who needs to be consulted?) [If not mentioned] Can you talk me through your general approach to: prescribing(physician)/requesting (nurses)/dispensing (pharmacist) a prescription for, AP/P medications to a typical resident with dementia, whomay be exhibiting behavior thatchallenge? (Rephrase: Talkme through1situation where this occurred. Prompts: Howwould you start this process or journey for a NH resident with dementia?What isthe first thing you would always do? Use of NPI?What would you do next?Would you always do this? Anything else?What about reviewing?What about Pro Re Nata(PRN) usage)Can you tell me about a case where you were able to successfully reduce someone’s dosage of these agents or manage someone without medications. What do youbelieve were the main facilitators? (What do you believe enables “appropriate” usage?) (Rephrase: What facilitates the use of alternative nonpharmacologic ap-proaches in residents who may not necessarily need AP/P medications?)Now can you tell me about a case where you were perhaps less successful. What do you believe were the main barriers in this case? How is it different? (What do youbelieve to be the main barriers to “appropriate” usage?) (Rephrase: What prevents the use of alternative nonpharmacologic approaches in residents who may notnecessarily need meds?)What are your views on nonpharmacologic approaches? (Prompts: Are they effective? Whose role is it? Are they being used first-line?)Do you believe that everyone involved in the care of residents with dementia knows enough about these medications? (Prompts: Why do you think this? Is there anygroup of people in particular that you feel could benefit from more training and education? What specifically do you think they need to know more about?)What about having the skills to effectively manage someone who is exhibiting behavior that challenge? (Prompts: Why do you think this? Is there any group of peoplein particular that you feel could benefit from more training and education? What specifically do you think they need to know more about?)What would you consider your responsibilities to be as ensuring that the residents receive these medicines appropriately? [If not answered] What strategies arecurrently available to support you in ensuring their usage is appropriate?(Rephrase: What resources would you use/consult with first to ensure appropriateness eg, guidelines, pharmacists, GP. Is there anything else that could be done tosupport you?)Who or what would influence your decision about whether or not to prescribe an AP/P to a resident with dementia? (Physician only) (Prompts: Why/Why not?Individuals/groups of HCPs/finance/NH itself/public opinion/guidelines. Anyone else) (Rephrase: How, if at all, does the team communicate about AntipsychoticMedication (APM) usage?)OR 11. How do you think that your views and opinions, and that of others, influence the prescriber, in relation to AP/P prescribing? (nurses and pharmacists) (Prompts:Individuals/groups of HCPs/finance/NH itself/public opinion/guidelines. Anyone else?) [If not mentioned already] (You may or may not be aware but HIQA haverecently started conducting dementia-themed inspections of NH, and have released updated standards with an increased emphasis on chemical restraints.) What isyour opinion on the influence of HIQA on AP/P prescribing in the NH setting? Prompts: HIQA have released new updated Standards with an increased emphasis onrestraint use in NH residents with dementia, are you familiar with them? Any thoughts? Negative or Positive Light?)That brings us to the end of the interview.Is there anything else I haven’t asked you today that you would like to mention?

(continued on next page)

K.A. Walsh et al. / JAMDA 19 (2018) 948e958 958.e7

Supplementary Table S3 (continued )

2B. HCAsIn your own words, tell me what your views are regarding the use of medications in dementia residents who are exhibiting challenging behaviors. (Prompts: Anti-psychotics such as Zyprexa and Seroquel. Other relaxers such as Xanax or Ativan. Is it appropriately prescribed in all cases? Is it necessary? What are the benefits andharms?)In these residents, what would you define as an “appropriate” use of these kind of medications?(Prompts: indication, frequency of review, duration, who needs to be consulted?)[If not mentioned] Can you talk me through your general approach to requesting a prescription for these agents to a dementia resident, who may be exhibitingchallenging behaviour? (Prompts if necessary: Is that something you would normally do as a HCA? Howwould you start this process or journey for a NH resident withdementia? What is the first thing you would always do? Use of NPI? What would you do next? Would you always do this? Anything else? What about reviewing?)Can you tell me about a case where you were able to manage someone without medications. What do you believe were the main facilitators? (What do you believeenables or facilitates “appropriate” use?) (Rephrase: What facilitates the use of alternative nonpharmacologic approaches in residents who may not necessarily needAP/P medications?)Now can you tell me about a case where you were perhaps less successful. What do you believe were the main barriers in this case? How is it different? (What do youbelieve to be the main barriers to “appropriate” use?) (Rephrase: What prevents the use of alternative nonpharmacologic approaches in residents who may notnecessarily need meds?)What are your views on nonpharmacologic approaches in residents? (Prompts: massage therapy, reminiscence therapy and music therapy. Are they effective? Whoserole is it? Are they being used first-line?)Do you believe that everyone involved in the care of residents with dementia know enough about these meds? (Prompts: Why do you think this is? Is there any groupof people in particular that you feel could benefit from more training and education? Consultants, GPs, nurses, pharmacists, HCAs, and family members. What spe-cifically do you think they need to know more about?)What about having the skills to effectively manage these challenging behaviour? (Prompts:Why do you think this is? Is there any group of people in particular that youfeel could benefit from more training and education? What do you think they need to know more about?)What would you consider your responsibilities to be as a HCA in ensuring that all residents receive these meds appropriately?[If not answered] What strategies are currently available to support this nursing home in ensuring the usage of these meds are appropriate?(Rephrase: What resources would they use/consult with first to ensure appropriateness eg, guidelines, pharmacists, GP. Is there anything else that could be done?)How do you think that your views and opinions, and that of others, influence the prescriber, in relation to AP/P prescribing? (Prompts: Individuals/groups of HCPs/finance/public opinion/guidelines. Anyone else?)In general, do you think that a healthcare professional’s attitude toward dementia can influence their usage? Why/why not?(Rephrase: Some studies in the literature found that HCPs with a positive attitude toward People with Dementia (PwD) were less likely to use APM. Would you agreewith this statement?) [If not mentioned already] (You may or may not be aware but HIQA have recently started conducting dementia-themed inspections of NH, andhave released updated standards with an increased emphasis on chemical restraints.) What is your opinion on the influence of HIQA on AP/P prescribing in the NHsetting? What about the Irish Dementia Strategy? (Prompts: HIQA have released new updated Standards with an increased emphasis on restraint use in NH residentswith dementia, are you familiar with them? Any thoughts? Positive or negative light?)That brings us to the end of the interview.Is there anything else I haven’t asked you today that you would like to mention?

2C. Family membersIn your own words, can you describe what your views are towards the use of medications in the care of your loved one? (Prompts if necessary: have they beenbeneficial? Have you noticed any improvements? Have they caused any side effects?)The focus of my PhD research is on the usage of a group of medications called Antipsychotics and another broader group of medications called Psychotropics in NHresidents with dementia. Common examples of Antipsychotics include Zyprexa, Seroquel and Serenace. Common examples of other Psychotropics include Xanax,Halcion, and Lustral. These drugs are sometimes prescribed to people with dementia if they are severely distressed or displaying some behavior that others may findchallenging such as aggressive or agitated behaviour.If you have any experience in the use of these medications in your loved one, I’d be very interested to hear your story. (If not, then this is absolutely fine we can still talkaboutmedication use in general) (Prompts if necessary:Whywas he/she prescribed these drugs? Can you rememberwhat it was he/shewas prescribed? Did it help thesituation? Were there any side effects? Is he/she still on it? Who stopped it and why?)Have you ever requested a prescription for such amedication or have you ever requested it to be stopped or reviewed? If yes, could you describe for me in general whathappened? (Prompts if necessary: Why did you do this? Is that something you would normally do as a family member? Would you always do this? Anything else?)From your perspective, what would constitute an “appropriate” use of such a medication? (Prompts if necessary: Who needs to be consulted in the process? How longshould they be on it, in general?)What do you believe facilitates the “appropriate” use of these medications?(Rephrase if necessary: Can you tell me about a time where you were able to manage a challenging behavior without medications). What facilitates the use ofalternative non medicine approaches in residents who may not necessarily need AP/P medicationsWhat do you believe to be the main barriers to the “appropriate” use?(Rephrase if necessary: Now can you tell me about a time where you were perhaps less successful and had to resort to medications. What prevents the use ofalternative nonpharmacologic approaches in residents who may not necessarily need AP/P medications?) What strategies would have helped in this situation?What are your views on alternative approaches, such as massage therapy, reminiscence therapy andmusic therapy? (Prompts if required: Do they work?Whose role isit? Are they being used before medications?)Do you believe that everyone involved in the care of residents with dementia know enough about these drugs? (Prompts if necessary: Why do you think this? Is thereany group of people in particular that you feel could benefit from more training and education? What specifically do you think they need to know more about?)What about having the skills to effectively manage someone who is exhibiting behavior that challenge? (Prompts if necessary: Without using medicines. Why do youthink this? Is there any group of people in particular that you feel could benefit frommore training and education?What do you think they need to knowmore about?)What would you consider your responsibilities to be as -family member in ensuring that he/she receives an appropriate prescription of these medications?How do you think that your views and opinions, influence the prescriber, in relation to prescribing of these agents? What about the views of others? (Prompts ifrequired: Individuals/groups of HCPs/financial/public opinion/guidelines/dementia strategy. Anyone else?)In general, do you think that a healthcare professional’s attitude toward dementia can influence their usage of these drugs? Why/why not? (Rephrase: Some studieshave found that HCPs with a positive attitude toward PwD were less likely to use APM. Would you agree with this statement?)That brings us to the end of the interview.Is there anything else I haven’t asked you today that you would like to mention?

3. Topic Guides (Version 10):3A. Healthcare professionals

In your own words, tell me what your views are regarding the use of antipsychotics in nursing home residents with dementia. (Prompts: Is it appropriately prescribedin all cases? Is it necessary?) (What impact, if any, do resources and financial issues have an AP prescribing, in your experience?)In the context of NH residents with dementia, what you would you define as an “appropriate” usage of these agents? (Prompts: indication, frequency of review,duration, who needs to be consulted?)

(continued on next page)

K.A. Walsh et al. / JAMDA 19 (2018) 948e958958.e8

Supplementary Table S3 (continued )

[If not mentioned] Can you talk me through your general approach to: prescribing (physician)/requesting (nurses)/dispensing (pharmacist) a prescription for, APmedications to a typical resident with dementia, who may be exhibiting behavior that challenge?(Rephrase: Talk me through1situation where this occurred. Prompts: How would you start this process or journey for a NH resident with dementia? What is the firstthing you would always do? Use of NPI? What would you do next? Would you always do this? Anything else? What about reviewing? What about PRN usage)Can you tell me about a case where you were able to successfully reduce someone’s dosage of these agents or manage someone without medications. What do youbelieve were the main facilitators? (What do you believe enables “appropriate” usage?) (Rephrase: What facilitates the use of alternative nonpharmacologic ap-proaches in residents who may not necessarily need AP/P medications?)Now can you tell me about a case where you were perhaps less successful. What do you believe were the main barriers in this case? How is it different? (What do youbelieve to be the main barriers to “appropriate” usage?) (Rephrase: What prevents the use of alternative nonpharmacologic approaches in residents who may notnecessarily need meds?)What are your views on nonpharmacologic approaches? (Prompts: Are they effective? Whose role is it? Are they being used first-line?)Do you believe that everyone involved in the care of residents with dementia knows enough about these medications? (Prompts: Why do you think this? Is there anygroup of people in particular that you feel could benefit from more training and education? What specifically do you think they need to know more about?)What about having the skills to effectively manage someone who is exhibiting behaviors that challenge? (Prompts:Why do you think this? Is there any group of peoplein particular that you feel could benefit from more training and education? What specifically do you think they need to know more about?)What would you consider your responsibilities to be as ensuring that the residents receive these medicines appropriately? [If not answered] What strategies or re-sources are currently available to support you in ensuring their usage is appropriate?(Rephrase: What resources would you use/consult with first to ensure appropriateness eg, guidelines, pharmacists, GP.As you may be aware, we are planning to undertake an intervention study in your NH to help support nurses, HCAs and doctors in ensuring prescribing of anti-psychotics is to a high quality. What would you like to see in this intervention program? (Prompts: What would be helpful to you as a X?What would not be helpful toyou?)Who would influence your decision about whether or not to prescribe an AP to a resident with dementia? What about guidelines? (Physician only) (Prompts: Why/Why not? Individuals/groups of HCPs/finance/NH itself/public opinion/guidelines. Anyone else) (Rephrase: How, if at all, does the team communicate about APMusage?)ORHow do you think that your views and opinions, and that of others, influence the prescriber, in relation to AP prescribing? What about guidelines? (nurses andpharmacists)(Prompts: Individuals/groups of HCPs/finance/NH itself/public opinion/guidelines. Anyone else?)Some people say that if a healthcare professional has a greater understanding of dementia then they might be less inclined to use antipsychotics. What do you thinkabout that?(Rephrase: Some studies in the literature found that HCPs with a positive attitude toward PwD were less likely to use APM. Would you agree with this statement?)Do different nursing homes have different cultures? If so, what impact does this have on AP prescribing? [If working in multiple sites][If not mentioned already] (You may or may not be aware but HIQA have recently started conducting dementia-themed inspections of NH, and have released updatedstandards with an increased emphasis on chemical restraints.) What is your opinion on the influence of HIQA on AP prescribing in the NH setting? Prompts: HIQA havereleased new updated Standards with an increased emphasis on restraint use in NH residents with dementia, are you familiar with them? Any thoughts? Negative orPositive Light?)That brings us to the end of the interview.Is there anything else I haven’t asked you today that you would like to mention?Thank you very much for giving up your time to talk to me today. I’ll now stop recording.

3B HCAsIn your own words, tell me what your views are regarding the use of antipsychotic medications in dementia residents who are exhibiting challenging behaviors.(Prompts: Antipsychotics such as Zyprexa� and Seroquel�. Other relaxers such as Xanax� or Ativan�. Is it appropriately prescribed in all cases? Is it necessary? Whatare the benefits and harms?)In these residents, what would you define as an “appropriate” use of these kind of medications?(Prompts: indication, frequency of review, duration, who needs to be consulted?)[If not mentioned] Can you talk me through your general approach to requesting a prescription for these agents to a dementia resident, who may be exhibitingchallenging behaviors? (Prompts if necessary: Is that something you would normally do as a HCA? How would you start this process or journey for a NH resident withdementia? What is the first thing you would always do? Use of NPI? What would you do next? Would you always do this? Anything else? What about reviewing?)Can you tell me about a case where the team were able to successfully reduce someone’s dosage of these agents and you were able to manage them without med-ications? What do you believe were the main facilitators? (What do you believe enables “appropriate” usage?) (Rephrase: What facilitates the use of alternativenonpharmacologic approaches in residents who may not necessarily need AP/P medications?)Now can you tell me about a case where you were perhaps unable to manage the patient without some form of medication? What do you believe were the mainbarriers in this case? How is it different? (What do you believe to be the main barriers to “appropriate” use?) (Rephrase: What prevents the use of alternativenonpharmacologic approaches in residents who may not necessarily need meds?)What are your views on alternative approaches to managing behaviour, such as distraction, massage therapy, reminiscence therapy and music therapy? (Prompts ifrequired: Do they work? Whose role is it? Are they being used before medications?)Do you believe that everyone involved in the care of residents with dementia know enough about these meds? (Prompts: Why do you think this is? Is there any groupof people in particular that you feel could benefit from more training and education? Consultants, GPs, nurses, pharmacists, HCAs, and family members. What spe-cifically do you think they need to know more about?)What about having the skills to effectively manage these challenging behaviors? (Prompts: Why do you think this is? Is there any group of people in particular that youfeel could benefit from more training and education? What do you think they need to know more about?)What would you consider your responsibilities to be as a HCA in ensuring that all residents receive these meds appropriately? [If not answered] What strategies orresources are currently available to support this nursing home in ensuring the usage of these meds are appropriate?(Rephrase: What resources would they use/consult with first to ensure appropriateness eg, guidelines, pharmacists, GP.As you may be aware, we are planning to undertake an intervention study in your NH to help support nurses, HCAs and doctors in ensuring prescribing of anti-psychotics is to a high quality. What would you like to see in this intervention program? (Prompts: What would be helpful to you as a X?What would not be helpful toyou?)How do you think that your views and opinions, and that of others, influence the prescriber, in relation to AP prescribing? (Prompts: Individuals/groups of HCPs/finance/public opinion/guidelines. Anyone else?)Some people say that if a healthcare professional has a greater understanding of dementia then they might be less inclined to use antipsychotics. What do you thinkabout that?(Rephrase: Some studies in the literature found that HCPs with a positive attitude toward PwD were less likely to use APM. Would you agree with this statement?)[If not mentioned already] (You may or may not be aware but HIQA have recently started conducting dementia-themed inspections of NHs, and have released updatedstandards with an increased emphasis on chemical restraints.) What is your opinion on the influence of HIQA on AP prescribing in the NH setting Strategy? (Prompts:HIQA have released new updated Standards with an increased emphasis on restraint use in NH residents with dementia, are you familiar with them? Any thoughts?Positive or negative light?)

(continued on next page)

K.A. Walsh et al. / JAMDA 19 (2018) 948e958 958.e9

Supplementary Table S3 (continued )

That brings us to the end of the interview.Is there anything else I haven’t asked you today that you would like to mention?

3C. Family membersIn your own words, can you describe what your views are towards the use of medications in the care of your loved one? (Prompts if necessary: have they beenbeneficial? Have you noticed any improvements? Have they caused any side effects?)The focus of my PhD research is on the usage of a group of medications called Antipsychotics in NH residents with dementia. Common examples of Antipsychoticsinclude Zyprexa, Seroquel and Serenace. These drugs are sometimes prescribed to people with dementia if they are severely distressed or displaying some behaviorthat others may find challenging such as aggressive or agitated behaviour.If you have any experience in the use of these medications in your loved one, I’d be very interested to hear your story. (If not, then this is absolutely fine we can still talkaboutmedication use in general) (Prompts if necessary:Whywas he/she prescribed these drugs? Can you rememberwhat it was he/shewas prescribed? Did it help thesituation? Were there any side effects? Is he/she still on it? Who stopped it and why?)Whenever your loved one is a bit agitated or distressed, is there anything that helps to put them at ease? (Prompts if necessary: Reminiscing about the past? Activities?What about Medications?)Have you ever requested a prescription for such amedication or have you ever requested it to be stopped or reviewed? If yes, could you describe for me in general whathappened? (Prompts if necessary: Why did you do this? Is that something you would normally do as a family member? Would you always do this? Anything else?)From your perspective, what would constitute an “appropriate” use of such a medication? (Prompts if necessary: Who needs to be consulted in the process? How longshould they be on it, in general?)What are your views on alternative approaches to managing behaviour, such as massage therapy, reminiscence therapy and music therapy? (Prompts if required: Dothey work? Whose role is it? Are they being used before medications?)Do you believe that everyone involved in the care of residents with dementia know enough about these drugs? (Prompts if necessary: Do family members knowenough? Should they know more? Is there any group of people in particular that you feel could benefit from more training and education? What specifically do youthink they need to know more about?)What about having the skills to effectively manage someone who is exhibiting behavior that challenge? (Prompts if necessary: Without using medicines. Why do youthink this? Is there any group of people in particular that you feel could benefit frommore training and education?What do you think they need to knowmore about?)What would you consider your responsibilities to be as -family member in ensuring that he/she receives an appropriate prescription of these medications?How do you think that your views and opinions, influence the GP, in relation to prescribing of these agents? What about the views of others?(Prompts if required: Individuals/groups of HCPs/financial/public opinion/guidelines/dementia strategy. Anyone else?, How are your views and opinions communi-cated to the GP?)Some people say that if a healthcare professional has a greater understanding of dementia then they might be less inclined to use antipsychotics. What do you thinkabout that?(Rephrase: Some studies have found that HCPs with a positive attitude toward PwD were less likely to use APM. Would you agree with this statement?)That brings us to the end of the interview.Is there anything else I haven’t asked you today that you would like to mention?

HCP, Healthcare Professional; NH, Nursing Home; AP/P, Antipsychotic or Psychotropic; APM, Antipsychotic Medication; QoL, Quality of Life; GP, General Practitioner; HCA,Healthcare Assistant; NPI, Non-pharmacologic Intervention.

K.A. Walsh et al. / JAMDA 19 (2018) 948e958958.e10

References

E1. Ritchie J, Spencer L, O’Connor W. Carrying out qualitative analysis. In: Ritchie J,Lewis J, editors. Qualitative Research Practice: A Guide for Social Science Stu-dents and Researchers. Los Angeles, CA: Sage; 2003. p. 219e262.

E2. QSR International Pty Ltd. NVivo Qualitative Data Analysis Software version11, 2017.

E3. Atkins L, Francis J, Islam R, et al. A guide to using the Theoretical DomainsFramework of behaviour change to investigate implementation problems.Implementation Sci 2017;12:77.

E4. Francis JJ, Johnston M, Robertson C, et al. What is an adequate sample size?Operationalising data saturation for theory-based interview studies. PsycholHealth 2010;25:1229e1245.

K.A. Walsh et al. / JAMDA 19 (2018) 948e958 958.e11

Supplementary Material S4. Detailed Description of DataAnalysis

Data Analysis

Data analysis followed the principles of framework analysis,E1

and used NVivo v 11.E2 We used deductive and inductive ap-proaches to analysis throughout the 5 stages (familiarization,identifying a thematic framework, indexing, charting, and mappingand interpretation). First, the author became familiar with the databy reading transcripts and field notes and open-coded the dataacross the entire data set. During indexing, data from the tran-scripts were coded into 1 or more TDF domains according to thedefinitions for each domain (Table 1). Simultaneously, conceptsemerging from the open coding were categorized inductively. Thesesimultaneous indexing steps were conducted independently by 3authors for 7 transcripts, who met to discuss differences in TDFapplication or interpretation of emerging concepts, and came toconsensus. The indexing of the remaining transcripts was con-ducted by the primary author.

Charting of the data, with distilled summaries in matrix formatwas used to identify the predominant TDF domains influencing thetarget behaviors (appropriate requesting and prescribing).E3 This wasperformed independently by 2 authors, who then discussed anydisagreement until consensus was reached. From these predominantdomains, the determinants (ie, barriers and facilitators) of the targetbehaviors were identified.

For the mapping and interpretation step, we iteratively developedlinks between determinants, predominant domains, categories andtheory to provide overall explanations for the findings. This wasachieved by constructing conceptual mind maps exploring possiblerelationships between all these different factors. By iterativelyexamining these evolving conceptual mind maps as an interdisci-plinary research group, we were able to condense our findings into 3explanatory themes and 1 overarching theme (Figure 1). Therefore,the behavioral determinants were the building blocks for the themes.These stages were not linear (Supplementary Figure S1), and the datacollection and analysis phases occurred concurrently, to enable theexploration of emergent themes in subsequent interviews and toidentify when data saturation occurred.E4

Stage 1: Familiariza�on

Read and re-read all the data and associated field notes. Open coded the complete dataset.

Stage 2: Iden�fying a thema�c framework

Selected the 14-domain Theore�cal Domains

Framework (TDF).

Stage 3: Indexing

Applied the 14 domains of the TDF systema�cally to the en�re dataset.

Generated ini�al categories of themes based on the open coding.

Stage 4: Char�ng

Created a matrix to summarize the data from each par�cipant against the 14 TDF domains. Iden�fied the 9 predominant domains influencing target behaviors

using a consensus approach between authors. Re-analysed data to determine the barriers and

facilitators to these target behaviors.

Stage 5: Mapping and Interpreta�on

Itera�vely developed links between barriers and facilitators, domains, ini�al categories and theory to provide overall explana�ons for the findings.

Generated explanatory themes with one overarching theme.

Expl

anat

ory

inte

rpre

ta�o

nDe

scrip

�ve

codi

ng

Supplementary Figure S1. The 5 iterative stages of Framework Analysis, moving from descriptive coding to explanatory interpretation.

K.A. Walsh et al. / JAMDA 19 (2018) 948e958958.e12