‘I just take them because I know the people that give them to me’: A theory‐informed interview study of community‐dwelling people with dementia and carers' perspectives of medicines management

Identify facilitators and barriers to successful medicines management for people with dementia (PwD) in primary care from the perspectives of community‐dwelling PwD and carers.


| INTRODUCTION
Medicines are important for people with dementia (PwD), for management of the condition and co-morbidities. 1,2 Medicines management, comprising 'the entire way medicines are selected, procured, delivered, prescribed, administered and reviewed to optimise the contribution they make to producing informed and desired outcomes of patient care', 3 raises unique challenges for PwD. Due to multimorbidity, PwD are likely to be prescribed multiple medications (polypharmacy), 4 resulting in complex medication regimens. Polypharmacy is a risk factor for potentially inappropriate prescribing, 5,6 associated with increased risk of adverse drug events, hospitalisation, mortality, increased healthcare costs, and reduced quality of life. 7,8 PwD's beliefs about medicines and treatment expectations may also influence prescribing. 9 Management of non-cognitive symptoms may challenge prescribers and those assisting with medication administration. [10][11][12] Impairment of cognitive and communication abilities in PwD, together with their beliefs about medicines, may result in nonadherence and medication-related hospital admissions. [13][14][15] Prudent medicines management is therefore vital for improving outcomes for PwD and reducing medicines-related harm.
With cognitive and functional decline, increased assistance with medicines from carers (formal/informal) is needed. [16][17][18] Studies report informal carers may have difficulty managing complicated medication regimens typical in PwD, 17,18 adding to their considerable burden. [17][18][19][20] If carers become unable to cope, alternative options, such as care home placement, may be required. 17 There is limited research on medicines management in community-dwelling PwD; carers' perspectives often dominate, with fewer PwD participating in previously published interview studies. 15,[17][18][19][20][21] Additionally, experiences of PwD living alone have not been established; such individuals are at greater risk of medication self-administration errors and nonadherence than PwD with support. 22,23 Interventions to improve medicines management for PwD are needed. 13,24 This study was part of a larger project, the overall aim of which was to develop an intervention to improve medicines management for PwD in primary care. The focus was those living in the community, who constitute the majority of PwD, 25 many of whom live alone. 26 In line with Medical Research Council guidance, 27 a theory-base was incorporated during intervention development. The Theoretical Domains Framework 28 (TDF) was used to understand behaviours and processes associated with medicines management amongst key stakeholders (i.e., PwD, carers, general practitioners (GPs) and community pharmacists).
The 14 theoretical domains may act as barriers or facilitators to an individual's behaviour (Table S1). 29 Therefore, the aim of this study was to explore perspectives of community-dwelling PwD and their carers about medicines management (prescribing, dispensing, administration, adherence and review) using a theory-informed approach, to identify facilitators and barriers to successful medicines management. Findings from healthcare professional interviews are reported elsewhere. 30

| Memory clinics
PwD/carer dyads were recruited through two memory clinics in Belfast Health and Social Care Trust. Research nurses screened PwD and carers against inclusion criteria and clinicians confirmed potential participants' eligibility. PwD were eligible if they had a dementia diagnosis (any type), living in their own home, prescribed ≥4 regular medications (this numerical cut-off was used in preceding pharmacoepidemiological research 5 ), and capable of undertaking an interview. Carers were eligible if they had contact with the PwD at least three times a week and assisted with medicines. Potential participants were approached by a clinician or research nurse, provided with study information, and followed up by telephone after one week by a researcher.

| Primary care
Ten GP surgeries across NI were recruited as part of the larger project. 30 We anticipated this recruitment method would identify PwD living on their own who may not receive assistance with medicines. Inclusion criteria were the same as those described above, however, PwD had to be living alone in their own home. Research nurses screened practice records against inclusion criteria, and GPs reviewed the resultant list to confirm eligibility. Study information was mailed to potential participants, who were followed up by telephone after one week by a research nurse.

| Join dementia research
Following slow recruitment of PwD through primary care, the Join Dementia Research (JDR; www.joindementiaresearch.nihr.ac.uk) database was used. This United Kingdom-based service allows people to volunteer to participate in dementia research. 35 The online portal matched potentially eligible PwD, using the same inclusion criteria for those recruited through primary care. A researcher telephoned individuals to confirm eligibility. Written information was sent to those who expressed interest in participating, and the researcher followed-up by telephone after one week.

| Procedure
Interviews were conducted by two researchers (Heather E. Barry,

Máiréad McGrattan; both pharmacists) in patients' homes between
October 2015 and November 2016. All participants provided informed consent and were offered £50 for participation. Interviews with PwD/carer dyads were designed to be conducted separately.
However, if PwD preferred to be interviewed with their carer present, this was respected.
Interview topic guides, based on the TDF, 28

| Analysis
PwD and carer interviews were analysed separately using NVivo 11. 37 Two researchers (Heather E. Barry, Máiréad McGrattan) independently analysed each transcript, based upon the approach used in the larger study. 30 The framework method 38 was used initially to deductively analyse data using TDF domains as coding categories. A framework matrix was generated for each participant group (PwD, carers) to summarise data. Content analysis 39 of each framework matrix was performed inductively to identify barriers and facilitators within each TDF domain. The researchers met frequently BARRY ET AL.
-885 to discuss coding; any disagreements were resolved through discussion with another researcher (Carmel M. Hughes).
Quantitative data were coded, entered into SPSS Statistics v22, 40 and descriptive analyses (frequencies, proportions, means and standard deviations [SD]) conducted. A mean and median score was calculated for each BMQ subscale (necessity, concerns, overuse and harm). Missing data were excluded from analyses.

| Participant characteristics
Eighteen PwD and 15 carers participated in the study (Table 1).

| Summary of findings from TDF analysis
PwD felt their responsibility was to take their medicines as prescribed and adhere to their medication regimen, whilst carers emphasised their supportive role to ensure that the PwD had taken their medications ('social/professional role and identity'). Carers identified a number of medicines management activities they undertook, including prescription ordering and acquisition, collection of medicines from the pharmacy, dispensing medicines into an adherence aid, administration of medicines and monitoring patient adherence.

| Beliefs about medicines
Completed BMQs were received from 15 PwD. Table 2  -887 overuse and harms, relative to the mid-points of these three scales.
Standard deviations calculated for each of these four sub-scales for the BMQ also provide evidence for variability, and hence diversity, in beliefs about medicines across the PwD participants.

| DISCUSSION
This is one of only two studies, to our knowledge, to use a theorybased approach to explore experiences of community-dwelling PwD and their carers about medicines management, and determine beliefs of PwD about medicines. 41 It extends a limited evidence base to provide valuable findings for future intervention development to improve medicines management for PwD. This study emphasises that carers provide critical assistance with medicines for many PwD and their input will be integral to future interventions. 27,41 In addition, intervention components may need to be tailored according to disease trajectory, cognitive function, and personal circumstances of PwD (if they have carer support or are living alone). Having to take medicines worries me (C) 2 (11. This study included a greater number of PwD (n = 18) than reported previously; other studies were dominated by carer participation. 5,[17][18][19][20]21 This is a key strength; the 'patient voice' is missing from the literature, 42 and it is important to portray and consider views of PwD to inform intervention development. Participants revealed they did not experience any major issues with medicines management at the time of interview and were happy with the status quo, contrasting with other studies, 16,17,18 and concerns expressed by healthcare professionals, particularly about adherence. 16,17,30 Whilst cognitive status was not assessed, PwD appeared to have mild-moderate cognitive impairment, received appropriate support from their GP surgery and community pharmacy, and all had a formal/informal carer to assist with medicines. In addition, PwD displayed positive beliefs about medicines, particularly medication necessity, which likely explains the observed intrinsic motivation towards medicine-taking and adherence. 9,43 Beliefs about medication are shown to be potentially important determinants of adherence in cognitively intact older adults. [44][45][46][47] Our findings suggest that such beliefs may also exert influence on PwD, although this would need to be confirmed with larger sample sizes. Conversely, carers anticipated future difficulties with medicine administration and adherence, when refusal of medicines and difficulties with swallowing may be apparent. 16,43 Disparity between PwD and carer views may also be due to the nature of the condition and impact that dementia may have on PwD's problem-solving, judgement and decision-making abilities. However, it has been reported that self-ratings by those with early-stage dementia about their medication management abilities were more accurate than those provided by carers. 48  While carer support is important throughout the disease course, this will become more critical with disease progression to avoid additional burden and stress placed upon them by medicines management.

| Strengths and limitations
A major strength of this study was that most accounts from PwD were triangulated with those of their carer, which allowed needs and experiences of both key stakeholders to be considered. It should be noted, however, that findings represent the views and perspectives of PwD and carers who were interviewed, and are therefore subject to any reporting biases pertinent to PwD and carers in this context and at that time. Interviewer bias was minimised by adopting a reflexive interviewing style, conducting regular debriefing sessions during data collection, and having regular meetings with the wider research team to discuss and agree upon analysis. We acknowledge the self-reported dementia diagnosis provided by the participant recruited through JDR is a limitation. In addition, participants' level of cognitive impairment was not measured in this study; further work is needed to fully ascertain medicines management issues experienced by community-dwelling PwD and their carers at moderate to advanced stages of the condition, enabling any future intervention to be adapted and refined accordingly. One limitation was the difficulty in trying to recruit PwD living alone, despite taking a number of approaches to sampling. 31 Such patients are at risk of social isolation, receiving inadequate medical supervision and having unmet medical needs. 26,49 It is still unknown as to how they cope with medicines management; future work must identify their needs so that appropriate support can be provided.

| CONCLUSION
This study provides a rich insight into the experiences and views -889