Medication management in older people: the MEMORABLE realist synthesis

Co-applicant Group

This was an informal group of five individuals who were invited by the chief investigator to collaborate on the development of the initial research proposal. Support and advice about the need for the research also came from a patient and public involvement (PPI) event in November 2015 run by Dr Maidment and the research’s PPI lead, Mrs S Bailey. The event involved eight older people and their informal carers.

The Co-applicant Group brought together topic, practice, research, information and public participation interests in medication management. Members collaborated on the development and submission of the MEMORABLE proposal to the National Institute for Health Research (NIHR). They generated the initial Protocol31 and MEMORABLE’s first published paper: Developing a framework for a novel multi-disciplinary, multi-agency intervention(s), to improve medication management in community-dwelling older people on complex medication regimens (MEMORABLE) – a realist synthesis. 30 The Co-applicant Group evolved into the Project Group.

Project Group

With an established membership of nine, this group was the monthly reference point for work on MEMORABLE, responsible for overseeing the project and disseminating the research. Members provided expertise in medication management, medicine, nursing, pharmacy, research and information methodology, and PPI. This group was key to maintaining momentum in the research and assuring delivery of the protocol.

Stakeholder Group

This group had a more fluid membership than the Project Group, including experts by experience, practitioners, managers and researchers. It had an independent lead from within the Project Group, who co-chaired it with the Chief Investigator. It met on five occasions to review key outputs from the research. Each meeting lasted for about 1.5 hours. Meetings were scheduled to fit with the evolution of the research and the need to gain diverse input from multiple perspectives. The group provided advice and feedback to the Project Group on the veracity of the emerging evidence, programme theories and proposed interventions, as well as the dissemination strategy. It had an important role in making sure that recommendations from the research were appropriate, practicable and likely to make a difference for older people, informal carers and practitioners.

Research Team

This small team of five included the chief investigator and the research associate. It was one of the key sources of topic, practice, research and information expertise for MEMORABLE. It met five times to consider discussion papers prepared by the research associate. Meetings were usually face to face and lasted for half a day, with follow-up contact by telephone and e-mail, if needed. It was the forum to engage in extended discussion and debate, and to resolve topic and research issues. The team significantly augmented the expertise available to the chief investigator and research associate, and provided detailed support to and scrutiny of the research, its content and outputs.

Patient and public involvement

While this study was being carried out, the Project Group regularly checked what was being done. The group had a PPI lead who was a co-applicant for the study and who contributed to all meetings, and is a co-author of this report. The PPI lead also provided advice on the way MEMORABLE was engaging with older people and informal carers, including advice on the types and the content of information, as well as facilitating links with potential participants in the research. In addition, the Stakeholder Group provided advice on findings and recommendations as these emerged from the study. This group had several older people and informal carers on it, along with people working in health, care and research. Both groups contributed to the dissemination strategy for MEMORABLE. The researchers valued and benefited from people’s expert, experiential input to the many discussions and debates. However, they also acknowledge the difficulties of ensuring that there is PPI ‘representation’ and the need to consider the workload (volume and time scale) that such a complex study can create for a single PPI lead.

Realist approach

This theory-driven approach helps explain the way complex interventions work in real-world situations. It supports the identification of what works, for whom, why and in which circumstances: the methodology’s four ‘W’s. 49–53,55,56 Thus, it enables researchers to address complexity53 by unpacking the context-sensitive mechanisms that are the generative causal processes found within an intervention. 49,51,57 Realist approaches are interested in the behaviours of those involved, from individuals to groups and organisations; the influential factors that bring about, prevent or modify those behaviours; and the outcomes that result, intended and unintended. One of the purposes of this approach is ‘to improve the thinking that goes into service building’. (Reproduced from Pawson et al. 50 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 3.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/3.0/.)

As medication management can be understood as a complex intervention,58,59 the characteristics and purpose of realist methodology are particularly well aligned to meeting MEMORABLE’s aims and objectives.

Realist review and realist evaluation

Realist methodology has two arms. The first, realist review, explores data from secondary sources such as published studies and reports60 and the second, realist evaluation, engages with primary data. 61 MEMORABLE was set up as a realist synthesis, combining realist review with realist evaluation. In this approach, knowledge from the literature was analysed with the first-hand experiences of those engaged in medication management, augmenting the potential contribution of both data sets to theorising about how this complex intervention works.

Theory in realist methodology

Two levels of theory are associated with the realist approach: programme theory and substantive theory.

Programme theory is generated within a study to explain what the intervention has been set up to do and how that intervention might work. 51,60 Initial or candidate programme theories developed at the start of a realist research project are then confirmed, refined or refuted through iterations of data synthesis. 50,51,53 Finalised, evidence-based programme theories can inform the development of recommendations or specific proposals. These might involve modifying an existing intervention, or perhaps suggesting a new intervention to try to achieve the outcomes that matter, typically by addressing contextual influences.

In addition to the iterative way programme theories evolve, they have another characteristic which is that they are finally formulated at a ‘middle-range’ level of abstraction. 51 This means that the programme theory is specified in a way that permits empirical testing. Both realist reviews and realist evaluations aim to produce middle-range realist programme theories. Being realist in nature contributes to the transferability of theories to other settings in which that same intervention may be delivered, or even to other interventions. This transferability is based on the assumption that in other settings or interventions the same mechanisms may be in operation. This is the way that realist methodology contributes to the accumulation of knowledge. 55,62 Candidate or initial programme theories for MEMORABLE were drafted at the beginning of the research: see Step 4: Candidate programme theories and substantive theory. Final programme theory for a key stage in medication management is contained in Figure 5.

Underpinning the analytical work on programme theories are the context–mechanism–outcome (CMO) configurations. 51,53,56 Generated from the data, they have explanatory potential in their configured format. It is, therefore, important that they are set out and reported in their entirety rather than disaggregated into lists of single components. 63 CMO configurations contain:

• Contexts – influential factors that activate people’s responses, enabling or inhibiting their decisions and actions. In MEMORABLE, contexts can include older people’s diagnoses or medication, the interpersonal relationships they have with health and care practitioners as well as system factors such as access to services or information.

• Mechanisms – the vital but often hidden operational ‘cogs’ that turn in response to the influence of contexts. Mechanisms can include the way people feel, think or act in relation to the resources in an intervention. They are activated by, and sensitive to, contexts. It is the interaction between contexts and mechanisms that leads to outcomes. 51,52,57 In MEMORABLE, mechanisms included older people’s, practitioners’ and informal carers’ diverse responses to what was happening in medication management such as confidence, control, experience and expertise, and mutual trust.

• Outcomes – the pattern of impacts and effects resulting from the way that contexts and mechanisms interact. In MEMORABLE, these included older people’s and informal carers’ coping experiences, achieving appropriate or inappropriate treatment, following adherent or non-adherent behaviours, quality of life as well as health and care relationships and resource use.

The CMO configurations are generated from, and stay close to, the data and are used to explain causation and therefore why certain outcomes happen under certain contexts within the scope of a programme theory.

Substantive theory is pre-existing, higher-level theory about how phenomena are supposed to work, perhaps domain-specific such as theories of behaviour, learning or change. 64 Identifying and applying substantive theory brings a particular lens to the research, helping with the development of initial programme theories, supporting data analysis and providing analogy. In MEMORABLE, normalisation process theory (NPT)65,66 was chosen as the relevant substantive theory. See Chapter 3, Medication management as implementation, for details of NPT and its application in the research.

Having provided an overview of the research methodology, the next section sets out a more detailed account of the way the methodology was followed in the work packages and the steps by which MEMORABLE was delivered.

Work packages

The protocol31 identifies three work packages to be carried out within MEMORABLE, following on from Project Start Up led by the Co-applicant Group, described in Co-applicant Group above. These packages conform with a realist research process,49,51,55 based on the recommended phases for this type of study:50 defining the scope of the review, searching for and appraising the evidence, extracting and synthesising findings, and drawing conclusions and making recommendations. The protocol details each package as follows:

• Work package 1: realist synthesis (understanding context and mechanism) – this package focuses on the realist review and involves developing the search strategy; selecting and screening articles for their relevance/rigour; analysing data; and developing the candidate or initial programme theories.

• Work package 2: realist evaluation (exploring mechanism) – this package is about gathering and analysing experiential, narrative data through realist-informed interviews (29 older people and informal carers, and 21 practitioners from health and care, including formal carers).

• Work package 3: developing framework for intervention(s) and dissemination – this final package involves refining the programme theories; undertaking any additional interviews or searching as required; identifying key mechanisms and related contexts; identifying and developing intervention strategies needed to change contexts and trigger mechanisms; presenting the intervention framework for feedback and further refinement; and dissemination.

Steps associated with work packages in the research process

The MEMORABLE study progressed through seven sequential and partly iterative steps (Figure 1). This figure highlights the non-linear way that the research advanced, with overlaps and iterations, particularly around steps 3, 4 and 5. Such overlaps and iterations are expected in researching a complex intervention such as medication management. 58,59,68

FIGURE 1.

Research process summary: work packages and steps. Reproduced with permission from Maidment et al. 67 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/.

Reproduced with permission from Maidment et al. 67 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/.

Describing this figure, below, steps are grouped by work packages (work package 1: realist synthesis – steps 1–4; work package 2: realist evaluation – step 5; work packages 1 and 2: iterations – steps 3a and b, 4a and 5a and b; and work package 3: developing framework for interventions and dissemination – steps 6–7). The results from this process are presented in Chapter 3.

Work package 1: realist synthesis (understanding context and mechanism)

Step 1: scope of the research

Medication management was scoped during Project Start Up through two strands of work. First, the Co-applicant Group drafted the MEMORABLE protocol,31 setting out the background to the research, highlighting:

• increases in the number of older people and comorbidity in the population1,2

• safety concerns and consequences of susceptibility to poor medication management in terms of harm and hospital admission20,21,27,28

• older people’s experiences of complex care pathways8,28,69

• the pivotal role of informal carers and formal carers24,70

• National Institute for Health and Care Excellence (NICE) guidance identifying the need for a collaborative approach to supporting older people with long-term conditions in managing their medication. 71

Second, just prior to the start of research, a formative Project Group held a day’s facilitated workshop to share knowledge on the principles and practice of realist methodology and to discuss the practicalities of applying them to MEMORABLE. This enabled the group to establish ownership of the planned research and build the team. It also mapped out a simple, tentative medication management process, drawing on the experience and expertise of those present. This mapping process began to theorise about the dynamics and complexity of the intervention and informed subsequent work on the stages of medication management. Although this early informal theorising does not follow the formal progression of the research process set out in Figure 1, it was nonetheless a valuable early contribution to scoping how medication management might work that was gradually refined as data were accumulated; a candidate programme theory in all but name (see Table 5).

Step 2: initial literature search and review

This step involved searching for and reviewing abstracts to broaden the understanding of medication management within MEMORABLE’s aim and objectives. Within a realist review, the aim of the search and synthesis noted by Pawson53 is to identify a body of literature to support ‘explanation building . . . to articulate underlying programme theories’ so that these can be interrogated using the evidence. In MEMORABLE, evidence would be enriched by combining data from the literature and interviews. Accordingly, the literature search was not intended to maximise the number and range of articles on medication management but to be more purposeful in identifying articles with rich data that would best support the generation and exploration of programme theories about what works, for whom, why and in which circumstances. This followed Pawson’s guidance on abstracting policy and theory building, described in detail by Booth et al. 55

The Research Team, which included a qualified information professional, conducted a pilot search to sensitise the team to the literature and to help in determining the scope of the protocol and subsequent review. 52 Details of this pilot search are given in Appendix 1. Once this initial set of literature had been reviewed closely, categorised and coded, the team finalised a search strategy for the initial literature search that would contribute to the review itself.

The initial literature search used terms set out in Table 1, accessing MEDLINE (724 references), CINAHL (Cumulative Index of Nursing and Allied Health Literature) (152 references) and EMBASE (139 references). It included published articles from 1 January 2009 to 31 July 2018. The initial cut-off was chosen because 2009 was the year in which NICE published its clinical guideline on medication adherence. 72 The Research Team had anticipated that systematic reviews published from 2009 onwards would access earlier studies, sufficient to cover the scope of the MEMORABLE realist review.

TABLE 1 - The MEMORABLE study’s search concepts

Reproduced with permission from Maidment et al. 67 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/.

Topic	Search terms
Medication management	Medication adherence – polypharmacy – medication management – medicines optimisation – concordance – compliance – adherence – regimen Health services misuse/inappropriate prescribing – drug prescriptions/practice patterns – physicians/inappropriate prescribing – drug utilisation/medication errors – de-prescribing
Older people	Age/aged
Long-term conditions	Chronic disease
Other limitations	English language

A total of 1018 abstracts were identified, reducing to 974 once duplicates, for example, were removed (Figure 2). 73 Abstracts were then reviewed for inclusion, referral or exclusion by the Research Associate using agreed criteria: population (≥ 60 years, living at home in the community), setting (UK, relevant high-income country), condition (multimorbidity) and interest (medication management). A 10% sample was also reviewed by two members of the Research Team to ensure consistency. A total of 140 articles were agreed for full-text review. The review was piloted by mapping abstracted data using a logic model format74 to identify descriptive accounts of medication management. This contributed to a better understanding of medication management, amplifying and updating the background information in the protocol.

FIGURE 2.

The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram: medication management. Reproduced with permission from Maidment et al. 67 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/.

Reproduced with permission from Maidment et al. 67 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/.

On reviewing the initial data set, the researchers concluded that the breadth and volume of the data explored a range of diverse issues that extended beyond the scope of the immediate needs of the research. More importantly, as a data set, the 140 articles generally lacked explanatory accounts to support the analysis of programme theories. As the research priority was to understand more clearly how medication management might work, it was agreed to focus more closely on a subset of articles containing causal accounts of this intervention. This approach was endorsed by the Project Group.

Step 3: further literature searches and review

Following this decision, the 140 articles were then searched for those containing the terms ‘concept, framework, model or theory’ to identify those articles that might include a logic model, programme theory or a theory of action or change. 74 A total of 29 articles were identified for full-text review and 24 were reviewed following the exclusion of duplicate articles that contained two of the search terms, such as concept and theory.

Articles were analysed and the data were coded into NVivo Pro 11 (QSR International, Warrington, UK). Two approaches were followed. The first used a predetermined coding framework based on the five factors and associated characteristics listed in a World Health Organization (WHO)12 report on adherence to long-term therapies. These factors and characteristics were frequently cited in the literature. The data were structured and reported using the five factors as headings with 27 associated subheadings, listed below to highlight their relevance to MEMORABLE:

• patient factors – relevant to older people and informal carers: demographics (age, diagnoses-multimorbidity, education, gender, living situation, nationality-culture), independence-dependence-functional ability (dexterity, hearing and sight, mobility, self-care, self-management), knowledge-beliefs-attitudes (about health, about medication, health burden, health literacy), mental health (anxiety, depression, cognitive capacity), physical health

• medication factors – aligned with polypharmacy: adverse events (including consequences of taking or missing doses, drug–disease interactions, drug–drug interactions, drug–food interactions), medication changes, medication information and instructions, medication name and type, medication number and regimen, packaging, storing, time horizon to benefit

• health-care provider factors – addressing practitioner and practice issues: challenges, communication and shared decision-making, co-ordination and consistency of care, person-centred care, practitioner roles, trust and confidence

• health-care system factors – covering health and care organisations: access to dispensers of medication, access to information, access to prescribers, access to review

• socioeconomic factors – the infrastructural underpinning of health and care organisations: evidence, finance, policy, technology.

However, this initial framework proved unwieldy because coding was constrained to fit with the list of existing factors and terminology, rather than more accurately reflecting the accounts in each text and the causal links between factors identified within and between texts. A different approach was needed. A second analysis coded data directly from the articles. Individual nodes and the eventual node structure emerged, emphasising the meaning and links attributed in each text and across multiple texts. The node summary is set out in Appendix 2. Nodes were then grouped under headings such as diagnoses, medication, health and care relationships, and an analysis framework established (Table 2). This framework combines the headings or factors and the levels at which they operated to ensure that coding was causally structured. In the table, factors are set out as potential contexts, mechanisms or outcomes, by levels, and a working definition formulated for the research. This work made an important contribution to sense-making at this stage and to later analysis.

TABLE 2 - Factors in the initial analysis framework

Factor in analysis framework		Working definitions in MEMORABLE
Potential contexts
Individual level	Diagnoses: multimorbidity	Two or more long-term conditions
	Medication: polypharmacy and treatment regimen	Five or more medications or, if fewer, a complex regimen that is proving difficult to manage
	Identity and day-to-day life	Older people aged ≥ 60 years, encompassing the individuality of older people and the unique bundle of personal influences on their decisions and actions
Interpersonal level	Health and care relationships and communication	Formal relationships and communication with practitioners
	Family and social relationships and communication	Informal relationships and communication with significant people in older people’s lives, including informal carers
Institutional/infrastructural level	Health and care system	The provision of services by organisations
	Support systems and information	Access to support and information services including via third sector or online, and published resources (e.g. patient information leaflet)
Potential mechanisms
Individual level	Workload	Can be ‘high or increasing’ (e.g. ‘high’: a large number of medications and a complex regimen, or ‘increasing’: the addition of a new medication) or ‘low or decreasing’ (e.g. ‘low’: a few medications with a simple, stable regimen, or ‘decreasing’: further simplification of the regimen)
	Capacity	Can be ‘high or increasing’ (e.g. ‘high’: knowledge and motivation about medication management or ‘increasing’: gaining knowledge and confidence by experience) or ‘low or decreasing’ (e.g. ‘low’: unable or unwilling to make sense of their medication and how it should be taken, or ‘decreasing’ as a result of an acute exacerbation or loss of informal support)
	Burden	The combination of workload and capacity as it impacts on people’s health and well-being. Burden can come from people’s diagnoses, their medication at the individual level, health and care relationships at the interpersonal and the heath and care system at the institutional or infrastructural level
Potential outcomes
Individual level	Coping experience	The way people are able to manage the burden of medication management in their day-to-day lives
	Health literacy	People’s knowledge of their health, diagnoses and medication
	Appropriate/inappropriate treatment	Whether or not medication is prescribed to meet people’s treatment needs, optimising safety, benefit and cost
	Adherence/non-adherence	The degree to which people take their medications as recommended by and agreed with the prescriber
	Performance/clinical outcomes	The impact of the medication in terms of the diagnosis or symptoms for which it was prescribed
	Quality of life	The impact of taking medication on the older person that accords with their values, goals and preferences
Interpersonal level	Health and care relationships	The interaction between older people or informal carers and practitioners
	Family and social relationships	The interaction between older people and informal carers as well as a wider network of family and friends
Institutional/infrastructural level	Resources	The operational and strategic goods, services and other outputs provided by organisations

This analysis was summarised in a discussion paper for a Research Team meeting to inform their work on drafting initial programme theories and reported to the Project Group.

Step 4: candidate programme theories and substantive theory

In this step, the Research Team continued to refine the tentative medication management process developed in step 1 as a theory of action or process. In addition, they drafted three candidate programme theories about medication management; one from the perspective of older people, one from the perspective of informal carers and one from the perspective of practitioners. These three candidate theories were developed from the literature searches and preliminary review, and drew on the Research Team’s own wider expertise and experience. It appeared that the ‘dissonance’ of agendas between different participants might have some traction:

• Older people – if an older person, living in the community with multiple diagnoses and complex medication regimens, feels that there is a dissonance/difference between their own priorities and those of the practitioners dealing with their medication management, then positive outcomes will not be achieved for and with them.

• Informal carers – if an informal carer who is providing help with medication management for an older person living in the community with multiple diagnoses and complex medication regimens feels that there is a dissonance/difference between the priorities of the person they care for, themselves or the practitioner dealing with their medication management, then positive outcomes will not be achieved for and with the older person.

• Practitioners – if a practitioner feels that they have to focus on structural/practice outcomes, then they are not able to accommodate other agendas/priorities, leading to a lack of influence on the knowledge, attitudes, behaviours and outcomes of older people who are living with multiple diagnoses and complex medication regimens in the community.

These theories extended the Research Team’s attempts to explain what works in medication management, for whom, why and in which circumstances to achieve the outcomes that mattered to those involved in the process.

Informed by the data, the simple medication management process that was mapped out informally in step 1 was also revised at this time. In its original format, it appeared to suggest that medication management was a linear process. However, it became increasingly clear that the process model needed to accommodate more of the complexity of potential loops and iterations that people might experience in their day-to-day lives (see Tables 5 and 6). These concepts were then available to discuss with interview participants in work package 2.

Several substantive theories of interest were considered at this stage but none were sufficiently well evidenced from the literature, pending clarification from the interview data.

Medication management as good practice

Good practice recommendations from medication management policy and guidelines informs practitioners’ work, and the management and commissioning of the services in which they operate. Policy and guidelines are generally aligned with preferred outcomes: adherence,12,72 in which the older person’s actions matches the prescriber’s recommendations, informed by their involvement in and agreement with decisions about their treatment, and optimisation33 that aims to ensure that older people get the best possible outcome from any treatment. Reports from two organisations are noted below, framing medication management practice.

First, a formative WHO12 report endorses a behavioural approach to medication adherence, expressed in an information–motivation–behavioural skills model. Behaviours are characterised by type, frequency, consistency, intensity and/or accuracy. A broad level explanation of behavioural change and adherence is contained in Annex 1 in the WHO report. Good practice is delivered through the practitioner’s clinical and communication competencies, such as providing emotional support (empathy, validation, relationship building); giving information (empowering, tailoring, supporting health literacy); asking adherence-specific questions (monitoring, feedback) and; discussing therapeutic options, negotiation of regimen and discussion of adherence (collaboration, shared decision-making). By pursuing this approach, the practitioner secures the preferred change in the individual. Adherence is however subject to the caveat:

. . . almost everyone has difficulty adhering to medical recommendation, especially when the advice entails self-administered care.

Reproduced from Sabate. 12 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY-NC 3.0 IGO) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: https://creativecommons.org/licenses/by-nc/3.0/igo/

Second, turning to the evolving policy framework for medication management in the UK, 2009–2018, NICE has published relevant guidance on adherence, optimisation and multimorbidity. 33,41,72 It has also issued several linked documents including a guideline and pathway for older people with social care needs,71,83 a flow chart for optimisation,84 and a key therapeutic topic document on multimorbidity and polypharmacy. 29

The NICE guidance on multimorbidity and polypharmacy29,41 acknowledges the complexity of managing future morbidity and mortality. NICE highlight two associated risks: first, in relation to multimorbidity, for which multiple treatments may be recommended although the evidence base is specific to the treatment of single conditions and, second, in relation to polypharmacy, for which benefits decrease and the risk of harm increases with the addition of each medication. NICE also underscore the responsibility of health-care professionals to make decisions appropriate to the circumstances of the individual patient, in consultation with the patient. Good practice interventions accordingly include the need for:

• an individualised approach to accommodate needs, treatment preferences, lifestyle and concerns through individualised and planned care, assessment and treatment;

• partnership with shared decision-making to ensure that there are informed discussions and decisions in which the patient experience is understood; and

• outcomes that include not only clinical aspects of medication safety and effectiveness but also the outcomes that matter to the individual, including improving quality of life and reducing treatment burden.

Despite an extensive, supporting evidence base, the NICE documents do not explain how these good practice interventions might work to enable practitioners to adapt them to local or individual circumstances. 85 Such an approach may appear to support the view that the act of ‘doing’ good practice might suffice. In such circumstances, practitioners could struggle to make informed, individualised decisions and assure good practice within the evidence available, noted by Hughes et al. 86

To summarise, key elements of good practice from these strategic reports on medication management have been mapped out in Figure 3, as a logic model: individual, interpersonal and system-level inputs, individual and interpersonal activities, and expected impacts.

FIGURE 3.

Mapping generic factors informed by WHO12 and NICE. 33,41,72

Medication management as a complex intervention

Medication management as implementation

Having established the stages of medication management as a complex intervention, this subsection addresses how medication management can be understood as a form of implementation involving ‘purposive and directed’66 work. As MEMORABLE is using realist methodology, the researchers have drawn on Carl May’s65,66,117 work on NPT and particularly the article by May and Finch. 66 This substantive theory was identified by the researchers during step 4a.

Normalisation process theory articulates the way new practices or activities are introduced and are made routine or sustained through the work that is done by those involved. It identifies four linked areas of work: coherence work to clarify what the new work is, relational work to build and sustain a network of people to carry the work out, operational work to undertake it, and appraisal work to understand the impact of what has been done. Normalisation process theory views this work as purposive, in other words, outcome focused. As a substantive theory on implementation, NPT aligns with medication management as a complex intervention, set out in Structuring complexity. It has been adopted here as an evaluative lens because of the way that it provides a structure and process for understanding the work associated with medication management, integrated across individual, interpersonal and system levels.

May et al. ’s65,66,117 work initially focused on change implementation in organisations. Gallacher et al. ,118 among others, have extended its application, using this theory to explain the workload and capacity issues associated with the experience of stroke. They locate burden work in health-care settings and people’s day-to-day lives. Gallacher et al. ’s article118 is, therefore, referenced here for the way the authors apply NPT to people’s experiences of health care. They also simplify NPT terminology. Accordingly, May et al. ’s terms have been revised as follows: coherence work to ‘sense-making’; relational work to ‘relationships’; operational work to ‘action’; and appraisal work to ‘reflecting’ (associated with older people and informal carers) and ‘monitoring’ (associated with practitioners).

Table 7 builds on Tables 5 and 6. Following the stages of medication management and focusing on the older person and the core relationship between them and the practitioner, this version of the table introduces NPT steps, distinguishing what is happening within and between stages:

• Stage 1: identifying problem – involving the older person (1) in individual sense-making about the importance of the problem and what to do about it. Drawing on both knowledge and confidence, choosing to act enables the older person to seek help, or not, while reflecting on the appropriateness of that decision and action.

• Stage 2: getting diagnosis and/or medications – within the relationship between the older person and the practitioner (1 : 1), shared decision-making is located here as an action. This is because shared decision-making is a type of work facilitated by the practitioner and contingent on them providing the time and space for it to occur. The action extends as far as the plan to do something that is subsequently initiated in Stage 3 and continued in Stage 4. The older person may reflect on their experience of the encounter and the practitioner relationship, in that moment and as a longer-term, trusted collaboration. Practitioner monitoring includes the performance elements of the work and the relationship, as well as the anticipated clinical outcomes.

• Stages 3 and 4: starting, changing or stopping medications and continuing to take medications – these stages are where the older person (1) makes sense of what they are doing in terms of the information that they have been given, what they are currently doing and what needs to be done to start and sustain a new routine. From this, the older person initiates (Stage 3) and follows that routine (Stage 4) and reflects on it, for example thinking about the way the medication makes them feel or the ease with which they can fit the new routine into what they do day to day.

• Stage 5: reviewing/reconciling medications – as with Stage 2, and the relationship between the older person and the practitioner (1 : 1), shared decision-making located here as an action. Building on Table 6, Table 7 also includes disruption, medication and diagnosis loops discussed previously.

The table highlights the way medication management can be understood using substantive theory as a framework. Within the apparent complexity in this table, patterns of work can be identified that contribute to explaining medication management including:

• The importance of sense-making and individual and shared decision-making from which actions are generated within medication management. These actions include the way shared decision-making might establish the choice architecture119 for older people’s day-to-day individual decision-making at home.

• The amount of work older people are doing to start and to continue to take medications as they make them routine, and fit that routine into their day-to-day lives. In this way, they experience and value the control over what they do with medication, as with all other aspects of their lives.

• The limited but significant opportunities for practitioners to influence older people’s motivation, knowledge and behaviours in the overall medication management process. Generally in NPT, implementation networks and interactions are task-recruited and specific to the introduction of new ways of working after which they may be disbanded. In medication management, although the interactions between older people and practitioners align with Stages 2 and 5, the relationship tends to be more enduring. It is valued for its consistency and qualities such as mutual trust. This can be identified in the table where the final step involves reflection/monitoring. Here, the question in Stages 2 and 5, ‘how did it go’ indicates the work being done to determine satisfaction with the current or recent interaction and also whether or not to confirm continuing trust in the relationship.

With medication management understood as a complex intervention and as implementation, structured and detailed across Tables 5–7, the next subsection explores medication management as burden.

Medication management as burden

Burden was identified as an issue of concern during the research process in step 3a: further literature searches and review (iteration). The Research Team were aware of Gallacher et al. ’s118 research, identifying treatment burden through the workload or demands of health care and people’s capacity or personal resources to ‘manage health problems and follow treatments’, a workload-capacity dynamic that is central to the concept of burden experienced by older people and informal carers. 13–15 A search for articles on burden was, therefore, carried out, linking it with multimorbidity, polypharmacy and chronic disease. Three articles were initially identified: Mohammed et al. ,78 Ridgeway et al. 77 and Sav et al. 14 Several other papers were subsequently identified from included references: Demain et al. ,13 Eton et al. ,82,120 Katusiime et al. ,121 Krska et al. ,122 Shippee et al. 123 and Tran et al. ,124 informing this subsection.

Illustrative examples of burden for each stage of medication management are set out in Table 8: Examples of burden: workload and capacity in medication management. A more detailed analysis of burden and coping is set out in Appendix 5, based on Mohammed et al. ,78 Sav et al. ,14 May et al. ,39 Eton et al. 120 and Demain et al.,13 and in Appendix 6, drawing primarily on the work of Katusiime et al. 121 and Krska et al. 122

Multiple factors inform workload and capacity: individually, these include the stability or instability of diagnoses, side effects from medications or the total medication regimen; interpersonally, having a consistent relationship with a practitioner such as a doctor, pharmacist or nurse that is founded on mutual trust in which the older person feels confident; or at an institutional or infrastructural level, accessing local services (e.g. being able to make convenient or timely appointments, transport or parking) and information.

However, burden is an individual experience, noted by Mohammed et al. :78 ‘patients on the same number of medicines may experience different levels and aspects of medication related burden’. In addition, in as much as workload and capacity change, the experience of burden is not static and people’s response in terms of coping also fluctuates, as set out in Table 9. This table highlights the relationship between increasing/high workload or capacity, and decreasing/low workload or capacity. Decreasing or low workload and increasing or high capacity are likely to be experienced as no burden. Where both workload and capacity are increasing or high, or both are decreasing or low, burden is experienced. In these two burden situations, increasing or high capacity is likely to indicate coping whereas decreasing or low capacity suggests not coping. Finally, increasing or high workload and decreasing or low capacity is associated with high burden and not coping. Here, there are both workload and capacity risks. The experience of low capacity was summarised in a practitioner’s interview:

Generally coping mechanisms, as in psychological, emotional coping. And we get people who, again it’s not necessarily a diagnosis, but personality. They don’t sort of cope with things well and might need a lot of support with relatively minor things. And that’s not a medical diagnosis. But they’re just sort of quite dependent on professionals or support.

P53 – general practitioner

The way people experience burden-control is described by Mohammed et al. :78 ‘Patients tolerate, control and adapt all events of a medication related burden until the cumulative effect of the burden interference in day-to-day life exceeds their coping threshold and overcomes their controlling capacity’. Older people and informal carers, who are not coping or at risk of not coping, need to be identifiable to practitioners (see Proposed interventions).

Finally, Shippee et al. 123 provide a compelling account of burden as a mechanism through which ‘complicating’ factors impact on outcomes: ‘a functional mechanism by which medical, social, and personal factors translate into the patient’s experience’. Although this use of the term ‘mechanism’ may not align with the way it is understood in realist terms, it does highlight people’s experiences, linking a range of causal influences and outcomes. The Research Team therefore adopted it in this study by conceptualising burden as a potential mechanism of interest. It has been mapped in to the analysis framework in Appendix 7, enhancing the initial framework set out in Table 2. Nonetheless, the Research Team also acknowledge that burden can also be a context or an outcome in the way medication management is explained and experienced, day to day.

Findings set out above have confirmed medication management as a complex intervention, as an implementation process and as burden. The final subsection below addresses the way it is experienced, drawing on the interviews with older people, informal carers and practitioners from health and care.

Medication management as experience

Adherence and optimisation are emphasised as key outcomes in both policy and the literature reported thus far, underscoring a practice-performance perspective. However, participant interviews provided rich and more divergent causal accounts through descriptions of diagnoses, treatments and experiences of managing medicines day to day and, finally, explanations of the way medication management works, or not, for them.

Reviewing/reconciling medication as good practice

Reviewing/reconciling medications aims to optimise medication management and improve quality of life in older people with long-term conditions requiring complex medication use. 129 This has been identified as Stage 5 in the medication management process, above.

The need for effective interventions to review or reconcile complex medication regimens and their use by older people has increased because of the growth of polypharmacy, the variety of alternative medications per diagnosis and the need to avoid the risks and costs associated with inappropriate treatments:130

The growing number of older adults who are prescribed multiple medications is at increased risk of receiving PIMs (potentially inappropriate medications) and PIRs (potentially inappropriate regimens). A major challenge in their care is the provision of medication reviews to identify these PIMs and PIRs.

Niehoff et al. 131

Reviewing/reconciling medications and burden

Analysis and format of findings

The findings for this stage of medication management have emerged from the continued application of a causal approach to the analysis of subsets of data, in this case centred on burden. Figure 4 sets out the key concepts in the analysis, informed by NPT. The aim is to understand burden in this stage and to identify potential contributions to its mitigation through implementation processes: sense-making, action, reflection/monitoring and relationships.

FIGURE 4.

Structure of the analysis of Stage 5: reviewing/reconciling medications – understanding burden.

Explaining reviewing/reconciling medications and burden

The subsections that follow set out the findings on reviewing/reconciling medications and burden, highlighting key themes that have explanatory value in understanding the complexity of burden in this stage and contributing to the framework in Figure 5. To aid navigation, the ‘dashboard’ (Table 14) highlights themes and the subsections in which they are set out, burden types and key steps.

FIGURE 5.

Summary of findings for Stage 5: reviewing/reconciling medications, mitigating burden.

TABLE 14 - Key themes: reviewing/reconciling medications and burden

Key theme	Burden type	Key step
Knowing about medication reviews and reconciliations: purpose and content, timing, practitioners’ experience and expertise, home-based reviews, reconciliations	Ambiguity	Sense-making
Having personalised, meaningful information	Concealment	Sense-making
Continuity and mutual trust	Unfamiliarity	Relationships
Addressing complex needs as a whole	Fragmentation	Action
Recognition and engagement	Exclusion	Action

Subsections explain each theme, setting out summary CMO configurations evidenced with extracts from the literature, as well as from interviews. CMO configurations are prefixed with RR as an abbreviation for reviews/reconciliations in this stage of medication management. This causal analysis supports the stratification of burden into five types: ambiguity, concealment, unfamiliarity, fragmentation and exclusion, locating these burdens as mechanisms. The researchers acknowledge that if this stratification was applied more widely to the analysis of data, as opposed to identifying the way that the identification of these burdens evolved, burden in general as well as specific burden types could also be contexts or outcomes in more generative explanations of the way medication management stages work.

Illustrative quotations have been selected to provide a rich and insightful example of interviewees’ experiences or because they make a unique but important contribution to understanding the topic. They add to the understanding of each subsection through subtle nuances, often lacking in the literature, that explain the interviewee’s experience of burden, directly or indirectly. The number of data extracts under any configuration does not imply a greater weight of evidence but illustrates its range, highlighting, for example, the difference between the perspectives of older people and practitioners, or the particular experiences of informal carers. CMO configurations contribute to explaining the stage and the burden within it.

Despite focusing on reviewing/reconciling medication (Stage 5), theorising about this stage contributes to confirming, refining or refuting the candidate programme theories on medication management as a whole that were set out at the start of the research (see Step 4: candidate programme theories and substantive theory). Figure 5 summarises the implementation steps within reviewing/reconciling medication (Stage 5) and, for each step, CMO configurations that describe and explain the researchers’ causal claims. Figure 5, therefore, represents a detailed subsection of the research’s emerging programme theory, focused specifically on Stage 5. This will be considered in Chapter 4.

Knowing about medication reviews and reconciliations: ambiguity

Having personalised, meaningful information: concealment

Practitioners draw on evidence from published sources as well as that gained through continuing professional development opportunities to inform and substantiate their practice. However, there can be tension between guidelines and experience, and the degree to which more powerful practitioner groups or individual practitioners feel confident to adapt them to meet the individual needs of an older person (Box 27), as well as limitations in the scope of some guidelines, such as those only relating to single rather than multiple pathologies.

Older people’s and informal carers’ contact with practitioners provides them with an opportunity to receive individualised information that addresses their own combination of multimorbidities and polypharmacy. The exchange of information enables them to be more equally involved in shared decision-making in this stage and potentially influences and informs day-to-day, individual decision-making and actions at home, supporting self-management and coping. The knowledge, understanding and confidence that come from information sharing contributes to increasing control and collaborative capacity, whether sourced from evidence or experience. Conflicting information is disruptive and has an impact on their confidence in practitioners (Box 28).

A few older people and informal carers seek information from other sources such as newspapers and websites despite not being sure about quality. In some cases, it is sufficient to feel that the information they find is useful whether or not it is derived from a safe or trustworthy source.

This subsection highlights concealment of information. Findings are summarised in the following CMO:

When practitioners do not provide information in ways that are personalised, meaningful and consistent (C), they do not engage older people and informal carers effectively in this key intervention that supports their health and wellbeing, including self- management (O), and waste time and other resources (O), through concealment (M).

RR CMO 02

Concealment can be addressed through the exchange of information, generic and individualised, ensuring that it is understood and useful, to support sense-making.

Continuity and mutual trust: unfamiliarity

Mutual trust is an emergent property of the relationship between the older person, the informal carer and the practitioner. It is central to practice. 148 Once established, continuity enables trust to be sustained. Furthermore, it engenders loyalty in older people and informal carers towards individual practitioners, especially GPs and their decisions. However, there are also risks that it can also be frustrated or irreconcilably lost.

Trust develops over time, from seeing the same practitioner in whom they are confident, underscoring the importance of continuity of care to practice and practitioner relationships:153–157

The patient–provider relationship is the fundamental starting point for patient centred care delivery. Central to the patient–provider relationship is trust.

Murray and McCrone157

Across diverse clinical settings, patients reported to be more satisfied with treatment, to show more beneficial health behaviours, less symptoms and higher quality of life when they had higher trust in their health care professional.

Birkhäuer et al. 112

Continuity of care is a long-standing feature of healthcare, especially of general practice. It is associated with increased patient satisfaction, increased take-up of health promotion, greater adherence to medical advice and decreased use of hospital services.

Pereira Gray et al. 114

We know that hospital admissions are often undesirable for patients, and there is some evidence that if GP care is managed by the same practitioner over a long period of time, patients have fewer emergency admissions.

Stafford et al. 6

Reproduced with permission from The Health Foundation

The experience of mutual trust and consistency is described in Box 29.

Disruption to services and practitioners is experienced negatively by older people and informal carers, even generating avoidance. This is illustrated in Box 30.

Disruptions in staff and services are also problematic for practitioners, arising from reorganisations, financial cuts and new initiatives, occurring in succession and, increasingly, together (Box 31).

This subsection explores unfamiliarity. Findings are explained in the following CMO:

When there is change and inconsistency in services and practitioners (C), older people, informal carers and practitioners do not establish and sustain mutual trust and confidence that comes from enduring relationships (O), and the foundations for shared decision making (O), as well as the co-ordinated management of complexity and risk (O), because of unfamiliarity (M).

RR CMO 03

Unfamiliarity can be addressed through interpersonal work to build and sustain relationships, supported by systems that ensure space, time and stability in which this work can occur.

Addressing complex needs as a whole: fragmentation

The health and care needs of older people and informal carers living with multimorbidity and polypharmacy are best addressed by a co-ordinated network of practitioners. Responsibility for carrying out medication reviews is increasingly being shared by doctors, nurses, pharmacists and pharmacy technicians across a number of organisations and sectors. Effective medication review is reliant on good interprofessional communication to bridge gaps and reduce any tension between this multiplicity of teams, staff, the older person and any informal carers involved:

Medication reviews, generally performed by physicians, are increasingly involving pharmacist and nurses as well.

Kallio et al. 151

Benefits of face-to-face ‘case conference’ communication between pharmacists and GPs as part of medication reviews.

Jokanovic et al. 152

Value of a multidisciplinary approach to preventing adverse drug events is widely acknowledged. Collaboration and teamwork between patients, general medical practitioners (GPs) and pharmacists is important to achieving the safe and effective use of medications.

Jokanovic et al. 152

Delivery models for health largely support specialisation and centralisation. Models for community teams are emerging that augment the range of knowledge, skills and practices in ways that are more local, integrated and complementary. Nonetheless, there needs to be one practitioner who retains responsibility for oversight. For reasons legal accountability, the primary health role is vested in the GP, although there can be positional ‘deference’ or uncertainty, leading to inertia between primary and secondary care. In addition, older people and informal carers feel that they are not understood nor that their complex needs are addressed as a whole, when they are seen by separate practitioners and across diffuse services114 (Box 32).

Practitioners adopt ways of working together to overcome diffusion and address complexity (Box 33).

Access to quality information supports practitioners in reviews and reconciliations, although there are issues with accuracy, delays and the completeness of the picture, partly as a result of disconnected care pathways and practitioners working in silos. This impacts adversely on co-ordination within and across services (Box 34).

In the UK, there is a political appetite for closer working across health and care being enacted through policy on more integrated delivery systems. However, in practice it appears that economic drivers for efficiency, opportunistic and transient initiatives, and clinical agendas around specialisation may be undermining these aspirations. Practitioners compensate for a lack of integrated health and care by working in formal and informal networks. Older people and informal carers also make significant efforts to navigate and bind disparate practitioners and services together to reflect their needs.

This subsection addresses fragmentation. Findings are explained in the following CMO:

Increasing specialisation and centralisation (C), and extending responsibility for reviewing and reconciling medications to multiple practitioner groups (C), results in older people and informal carers having multiple, discrete contacts with services (O), limiting the way they are understood and their needs addressed as a whole (O), including the co-ordinated management of complexity and risk (O), because of fragmentation (M).

RR CMO 04

Fragmentation can be addressed through relationship and action work at inter- and intra-agency levels to build and sustain operational and strategic networks.

Recognition and engagement: exclusion

As established previously in this report, older people and informal carers develop expertise in the day-to-day management of their health, diagnoses and medications (see Medication management). Their experience and expertise can make a useful contribution to decisions and actions about their health and care. This includes when reviewing and reconciling medications. In this type of interpersonal work, shared decision-making is a well-established part of collaborative working. It is underpinned by a substantial body of evidence42,158–164 and support for its use in practice:33,41,165

Achieving shared decision making depends on building a good relationship in the clinical encounter so that information is shared and patients are supported to deliberate and express their preferences and views during the decision making process.

Elwyn et al. 164

Face to face interactions, permission and space to discuss options, and continuity of patient-professional relationships are key in supporting older people with complex needs to engage in shared decision making.

Bunn et al. 159

Shared decisions based on what is important to each person in terms of treatments, health priorities, lifestyle and goals.

NICE. 41

© NICE 2016. Multimorbidity: Clinical Assessment and Management. Available from www.nice.org.uk/guidance/ng56. All rights reserved. Subject to Notice of rights. NICE guidance is prepared for the National Health Service in England. All NICE guidance is subject to regular review and may be updated or withdrawn. NICE accepts no responsibility for the use of its content in this product/publication

Older people and informal carers value recognition and engagement in decisions and actions affecting their health and medication (Box 35).

Practitioners also value their input and ways of securing the involvement of older people and informal carers (Box 36).

Box 37 draws on more general evidence about the need for time and space for this interpersonal work to evolve that can be translated to reviews and reconciliations.

This subsection addresses exclusion. Findings are summarised in the following CMO:

When older people and informal carers are not recognised and engaged with practitioners in shared decisions about their health, diagnoses and medications (C), the individualised basis for decisions is limited (O), older people and informal carers are less likely to apply what has been agreed (O), and their relationship with practitioners is undermined (O), because of exclusion (M).

RR CMO 05

Exclusion can be addressed through action work using shared decision-making to benefit from the individualised expertise and experience of all those most involved.

Summary: burdens in reviewing/reconciling medications

Reviewing and reconciling medications is a key, complex interpersonal stage in medication management. Using a realist approach, MEMORABLE has stratified this complexity by identifying five burdens that could be addressed through mitigation work in this stage, influencing what is done in other stages:

• The ambiguity burden about reviewing/reconciling medications within medication management – this involves sense-making by clarifying the purpose and content of medication reviews and reconciliations.

• The concealment burden due to a lack of information-giving that prevents older people and informal carers from understanding, personalising and using what they want or need to know – this is also about sense-making and establishing meaning through information to increase personal efficacy, agency and control.

• The unfamiliarity burden arising from not seeing the same practitioner consistently – this is about establishing continuity and mutual trust in relationships as the foundations for interpersonal work.

• The fragmentation burden limits the way older people and informal carers are understood and their needs addressed as a whole when they are seen across a range of services – this is about the importance of inter- and intra-agency collaboration in operational and strategic networks.

• The exclusion burden when older people and informal carers are not recognised for their experience and expertise, nor fully or effectively engaged in decisions that affect their health and care – this is about action and enacting collaboration through shared decision-making.

These findings are summarised in Figure 5, which highlights several key elements:

• Identification of those involved in Stage 5: reviewing/reconciling medications – older people, informal carers and practitioners, and their association through multimorbidity and polypharmacy.

• Complexity is acknowledged and accommodated.

• Burden issues are highlighted – ambiguity, concealment, unfamiliarity, fragmentation and exclusion.

• Work to mitigate burden is associated with sense-making, relationships and action as linked implementation steps (NPT) in this and each of the five medication management stages. This includes the way steps interact across stages in the medication management process. These generative links are suggested by arrows that overlap the perimeter of the box.

• Enduring relationships are emphasised as foundational to the work that is done.

• Important elements of sense-making and action are highlighted.

• Reflection/monitoring are included as part of the implementation process, although it was only implied in the data rather than specified.

• Indicative but important outcomes of burden and risk management, and sustaining coping and control are included.

Based on these findings, the figure also identifies proposed interventions generated within this stage:

• Risk identification as a point of access from Stage 4: continuing to take medications, signifying its contribution to identifying those who are not coping and should be offered further evaluation or review rather than waiting for a planned review.

• Individualised information is located centrally to emphasise its importance to sense-making, action and relationships.

Details of the proposed interventions are set out below.

Risk identification (linking Stages 4 and 5)

Older people and informal carers who are not coping and are at risk may not be readily visible to practitioners. There needs to be a simple and easily implementable way of identifying older people and informal carers who are not coping and then, having done so, to enable them to access appropriate help and support through a more detailed follow-up, such as a review. Many older people and informal carers have multiple contact points with health and care and therefore, one way of identifying those at risk would be to have simple ‘Whooley-type’ questions [URL: https://whooleyquestions.ucsf.edu (accessed 10 December 2019)] that could be used by any practitioner. These screening questions would be burden focused, one on workload and one on capacity with two ‘no’ answers indicating no significant burden at this time and one or two ‘yes’ answers requiring further evaluation.

Risk identification could be undertaken at any contact point or time in Stage 4, while older people and informal carers are continuing to take medications as a way of fast tracking them to further evaluation, help and support, or to a medication review (Stage 5).

Individualised information (linking Stages 5 and 2 to Stages 1, 3 and 4)

McHorney et al. 101 state: ‘people can only do what they understand’. Despite good information on single diagnoses and single medications, there is limited information, if any, that encompasses the unique combination of multimorbidities that increasing numbers of older people live with, as well as the complex polypharmacy regimens they are prescribed. Older people and informal carers stated that current information does not reflect the uniqueness and complexities that each individual experiences. 167

Learning how to manage this ‘by doing’ is inadequate and fraught with risks, including following routines that could be mistaken in their underlying assumptions about when and how to take certain medications. Making individual decisions on a day-to-day basis, without adequate information, adds to that risk. In addition, the way older people and informal carers have to repeat the same information about themselves to practitioners is burdensome, time-consuming and may inadvertently lead to them not divulging the way that they are actually managing or mismanaging their medications.

In this case, the proposed intervention is for older people and informal carers to have individualised information that they can use and share, tailored to their particular preferences, needs and experiences of living with several long-term conditions and taking many medications. The format, content and ways of using this individualised information would be developed collaboratively by working with those groups most directly involved, as with this research, as well as with others with expertise to turn proposals into practice.

Both proposed interventions would be pursued using realist methodology.

These findings are discussed in Chapter 4.

Complexity in medication management: stages and steps

Much of the existing literature identifies medication management as a single intervention. Policy and guidelines are often based on good practice. Evaluation is based on describing cause and effect168 rather than explaining real-world complexity. Policy and practice are typically aligned with issues of medication adherence and optimisation, with a behavioural underpinning to achieving both of these aims. The message is that these ‘betterment’ goals, optimisation and adherence, can be secured through various forms of behavioural incentives, disincentives and evidence-based advice.

Advancing this understanding, MEMORABLE views medication management as a complex intervention, aligning with the Medical Research Council58,59 definition, engaging older people, informal carers and diverse practitioners across health and care. Medication management is also an implementation process, operating across five linked stages with four implementation steps and three loops in each stage. This process structure offers a framework for understanding the generative and serial dynamics of real-world complexity within which individual and interpersonal work is carried out (NPT). 65,66

Stages 1, 3 and 4 involve older people doing work on their own: identifying problems (Stage 1); starting, changing or stopping medications (Stage 3); or continuing to take them (Stage 4). In some instances these stages may be supported by an informal carer, typically a family member, or a formal carer, assuring adherence and minimising risk.

Routines are fundamental to what older people and informal carers do, emerging from sense-making, action and reflection, and founded on key relationships, from which they exert agency and control. Routines are flexed to minimise burden and accommodate short-term and longer-term change. Examples of routines include linking medication taking with mealtimes and keeping pills in the kitchen as a prompt, organising regular pharmacy pick-ups as part of trips to the shops or arranging doctor’s appointments well in advance to ensure that older people see their preferred practitioner. Thus, older people and informal carers cope by fitting the practicalities of medication management into their day-to-day lives to achieve the outcomes that matter to them. Medication management plays a part in this, but is not an end in itself; adherence is only part of their narrative. Accordingly, practitioners need to address routines, outcomes, coping and burden in their contacts with older people and informal carers to ensure a purposeful and productive alignment between their agendas and to ‘nudge’ adherence within those discussions.

Difficulties in Stages 3 and 4 can provoke a ‘disruption loop’ when a problem occurs (Stage 1). Sustaining stable medication management, coping and optimisation in Stage 4 appears to be desirable. Frequent jumps from Stage 4 to Stage 1 may indicate a lack of coping and risk, particularly if red flag issues arise in Stage 1 and the older person or informal carer does not contact a practitioner for help and support. Red flags might include consistently not keeping enough tablets at home or stockpiling; muddling up tablets that look similar but which should be taken for different reasons; or avoiding taking particular tablets because of attributing undesirable or inconvenient side effects to them, but not seeking help to identify an alternative. Practitioners need to be able to identify those at risk and provide older people and informal carers with accessible, personalised information that enables them to be confident in determining the significance of problems and how best to resolve them, on their own or by accessing appropriate and timely support.

Stages 2 and 5 are where older people and informal carers engage with practitioners, typically in a clinical or formal setting. These two interpersonal stages are where they are getting a diagnosis and/or medications (Stage 2) and reviewing/reconciling medications (Stage 5). Stage 5 may trigger a ‘medication loop’ to Stage 4 where older people simply continue to take medications as before or to Stage 3 if they are advised either to stop, adjust doses or start something new, which then becomes part of their regimen in Stage 4. In addition, Stage 5 may trigger a ‘diagnosis loop’ where concerns are raised about their health, addressed in Stage 2. Shared decision-making is a key aspect of both stages.

Time-limited, interpersonal contact points are where information about older people’s and informal carers’ experiences of making medication management work day to day is exchanged, shared decisions are made and actions taken that influence what happens in the other three stages. Based on the research findings, shared decision-making provides an opportunity to establish the details and acknowledge the importance of the routines that older people and informal carers follow to feel that they have control over medication management in their lives and to achieve meaningful outcomes. Quality collaboration, with practitioners, optimises older people’s and informal carers’ intrinsic motivation and better informs the individual decisions they make at home. Alongside assuring safety and effectiveness, addressing these issues of ‘fit’ may have greater resonance and leverage in the decisions and actions of older people and informal carers about what medications they take, how and when. Supporting them to make medication management one of the many activities that they choose to engage in, day to day, may prove more enabling than emphasising practice-defined aims of adherence or optimisation.

The MEMORABLE Research Team identified that there were advantages in understanding medication management as a complex, dynamic stage, step and loop implementation process. First, it enabled them to work with its complexity, at a level that was manageable. Second, it provided a structure to identify and characterise key stages, steps and loops in the process, framing the data. Having set out this process early in the research and refining it as the research progressed, it proved to be useful in discussions about the direction of the literature search, in focusing the interviews and in interpreting and applying findings from the data from a consistent perspective.

The Research Team acknowledge that one of the tensions in medication management as an intervention is how to view it as a discrete aspect of the wider and complex processes of care involving older people with multimorbidity and polypharmacy,9 to underscore its importance and ensure that appropriate practitioners such as pharmacists are involved and their work co-ordinated. However, it also needs to be integrated within the wider health and care process, acknowledging that medication is increasingly important in the treatment of chronic diseases. To contribute to this, the Research Team recommend that further research is undertaken to extend the application of this stage, step and loop approach to complexity in medication management, underpinning it with a more substantive evidence base and causal account of each stage and extending across stages, to support the development of policy and practice in this field.

Outcomes in medication management

Medication management guidelines emphasise adherence and optimisation as outcomes. The literature on medication management provides an extensive and growing evidence base for both.

The MEMORABLE study’s interviews identify an extended range of outcomes that interviewees attach meaning to or want to achieve. Older people’s outcomes generally appear more diverse, purposeful, functional and qualitative than the role-informed outcomes of practitioners. As the key beneficiaries of medication, there is an embedded friction between minimising the number of medications older people have to take while achieving the effects that enables them to do things that matter to them. There is a reciprocal tension for practitioners as they mediate between criteria in guidelines that inform good practice on which their performance is assessed and the older person’s expressed wishes.

There is some degree of convergence in the outcomes identified by informal carers and practitioners attached to their roles in medication management, perhaps reflecting the way informal carers feel responsible for ‘good practice’ and adherence. There is also divergence in the outcomes informal carers and practitioners identify for others, highlighting the tension between personal and role-based agendas.

Early in the research, the researchers agreed that adherence and optimisation were important practice outcomes in MEMORABLE, acknowledging the importance of taking medications as prescribed and agreed (adherence) to achieve the best outcomes (optimisation). However, the researchers extended the range of outcomes by sourcing them from the literature, set out in the analysis framework (see Table 2 and Appendix 7), and populating them from the research interviews. By asking interviewees about the outcomes that they derived from medication management and then what had brought these about, the researchers identified outcomes occurring at different stages or levels, and evolving over time, as well as identifying the contexts and mechanisms involved. These causal accounts resolved explanatory gaps in the literature and informed the writing of CMO configurations.

The Research Team recommend undertaking further research on outcomes in medication management, involving older people, informal carers, and health and care practitioners. This could contribute to developing a vocabulary and possibly a decision support tool for practitioners to better engage older people and informal carers in discussions about what matters to them and the routines they might use to ensure that these outcomes are achieved. These outcomes would be documented as part of their individualised information.

Informal carers in medication management

Although the main focus in MEMORABLE has been on older people and medication management, the role and contribution of informal carers has emerged as an important topic. MEMORABLE, in line with other research, found that the depth and breadth of support informal carers provide to older people with multimorbidity and polypharmacy is increasing, many taking on a pivotal role and making significant personal sacrifices as a result. 21–25 The outcomes informal carers want from medication management highlight the tension that exists between the responsibility of being a care provider, in which they seek to adhere to the practitioners’ directions, and their identification with the wishes of and for that older person, based on familial ties and personal knowledge.

Informal carers experience role-creep as they become more engaged in the older person’s medication management and a time lag in acquiring the information and the increasingly specialist knowledge and skills that they need, which are largely gained by experience. With many taking on more responsibility and more complex tasks than paid carers, they nonetheless lack equivalent training or support.

Informal carers can also lack recognition from health and care practitioners or within health or care systems, arising from legalities around the older person’s ‘capacity’ as defined by mental health legislation. In circumstances when an assessment of ‘capacity’ may be borderline, an older person may nonetheless be failing to cope with the complexities of medication management. As a result, informal carers may be taking on a significant part of the workload, yet be excluded from key shared decision-making opportunities.

The Research Team recommend that further research is needed on the role of informal carers in supporting medication management. MEMORABLE will publish a paper on the research findings for informal carers to contribute additional evidence to the debate on this issue.

To conclude, medication management is a complex intervention and, therefore, improvement proposals need to engage with that complexity. MEMORABLE has developed a five-stage implementation process, causally focused and evidence and experience based. However, MEMORABLE remains a formative study. Accordingly, the researchers believe that framing medication management as a five-stage, four-step, three-loop process merits further exploration and critique to advance its contribution to research and practice.

Complexity in reviewing/reconciling medications

Reviewing/reconciling medications are the interventions associated with Stage 5 of the medication management process. It is an interpersonal stage where older people and practitioners collaborate, as with Stage 2: getting diagnosis and/or medications. Informal carers may also be involved. It involves regularly reviewing the way in which older people are continuing to take medications (Stage 4) or reconciliation at a health or care transition. As a result of what is decided in Stage 5, older people may continue to take their existing medications (medication loop to Stage 4); they may start, change or stop medications (medication loop to Stage 3); or they may need an assessment for a new diagnosis and medications (diagnosis loop to Stage 2). This process is illustrated in Table 6.

Reviewing/reconciling medications (Stage 5) was selected for more detailed analysis, as it was judged to be a key stage and an exemplar within medication management. This decision was made during the iterations of work packages 1 and 2, steps 3b/5b, based on:

• the emerging significance of this stage to the overall process

• its immediate links to Stage 4, where older people are continuing to take their medications and where non-adherence may emerge as an issue of concern

• the way it is understood through policy and guidance that could clarify standards of good practice and their intended impact, causally explained or not

• its importance in the way it focuses on interpersonal decisions and actions, and, therefore, its potential relevance to Stage 2, where the other interpersonal work occurs when getting a diagnosis and/or medications.

There are, however, important distinctions between the two interventions that make up this stage: reconciling and reviewing medications.

The first intervention is reconciling medications. It is associated with transitions between settings, the accuracy of medication lists and risk minimisation, such as avoiding drug–drug interactions or prescribing two medications with similar actions for the same or similar symptoms. Medication reconciliations are perceived as more straightforward than medication reviews. In the interviews, prescribers commented on the importance of having accurate information about the medications older people are taking, who prescribed them and why. This information is essential in treatment planning, including rehabilitative work to support older people to regain their independence in managing their medications when they return home from hospital. However, practitioners commented on problems with the quality and speed of communication needed to carry out reconciliations or act on results.

The second intervention is reviewing medications. MEMORABLE has identified three issues in the way medication review is understood and practised.

The first reviewing issue is about ‘what is done and when’. The researchers’ interpretations of the data indicate variation in older people’s and informal carers’ experiences of the content and timing of reviews. There are three accepted levels of medication review based on patient need. Policy, and associated funding, require MURs to be carried out by community pharmacists, with the older person present. Elsewhere, practitioners currently determine which level is most appropriate for any individual, acknowledging the resource limitations in the system.

It appears from the interview data that older people who are coping and whose condition is stable may have 6-monthly or annual reviews reflecting the practitioner’s perception of relatively low risk. Older people with less stable conditions and who are not coping appear to be having reviews as part of regular contact with practitioners. Their formal reviews are then often carried out as level 1 or 2 at which they are not present. 132,169 However, MEMORABLE has established that, despite contact with practitioners as part of ongoing medication management, older people and informal carers who are not coping and are at risk are not being identified or their needs addressed. Risk may be exacerbated by waiting for a planned review. MEMORABLE proposes that older people and informal carers whose day-to-day medication routines are breaking down and who are not coping need to be identified at any point of contact with a health or social care practitioner so that they can be offered further evaluation and support. Risk identification is discussed in Proposed interventions/future research.

The second reviewing issue is about ‘who’. The legal accountability for health care is held by the GP in primary care. However, the development of professional roles and multidisciplinary team working has resulted in pharmacists and nurses taking on additional responsibilities for prescribing and reviewing, building on their specialist knowledge, skills and supplementary training. Specialist nurses contribute to the management of specific long-term conditions, such as respiratory or cardiac, and community matrons support older people and informal carers living with complex long-term conditions in the community.

Pharmacists increasingly work with GPs in reviewing medications. Both the literature and interview data, describe a variety of effective pharmacy service developments in primary care where the specific expertise and experience of pharmacists is valued. Developments do, however, appear to be patchy and sometimes transient, subject to local sign-up and funding uncertainties. With clarification of roles and responsibilities that integrate pharmacists within the primary care team, improved information that gives them access to patient notes, and improved communication systems that would come from co-location, it appears that pharmacists could contribute to more effective service delivery, quality and impact, increasing capacity and reducing workload for GPs. 170,171 Thus, potentially, MEMORABLE found some support for the developing role of practice-based pharmacists.

The third reviewing issue is about ‘where’. There is no need for home-based reviews for all older people with multimorbidity and polypharmacy. However, the potential advantages of reviewing older people and informal carers who are not coping, at home, are clearly made when the focus of the review encompasses the fit between the demands of medication management and older people’s day-to-day lives. Practitioner interviews confirmed the success they have in identifying and resolving risks in the home setting where older people’s and informal carers’ routines can be demonstrated and discussed. It is harder for older people and informal carers to indicate that they are coping, perhaps out of a misplaced loyalty to the practitioner, when there are boxes of pills randomly scattered about the house, a litter of old prescriptions or a muddled store of medications, some of which are long out of date. Support to regain control of the situation and to re-establish routines that fit with day-to-day lives is readily available from practitioners once the issue becomes known. As another aspect of the risk issue, risk identification is discussed in Proposed interventions/future research.

Causality in reviewing/reconciling medications and burden

By focusing in detail on this stage, MEMORABLE has extended the understanding of reviewing/reconciling medications within medication management. It has scoped and identified tasks associated with good practice. Using a realist approach has enabled the researchers to explore how implementation work is done and with what effect.

From the many lenses the researchers could bring to this analysis, they chose to conceptualise burden as a potential mechanism of interest, based on an initial assumption and subsequent validation of the need to mitigate burden in medication management. To do this, they drew on the evidence, experiential accounts and theorising processes within realist methodology, basing analysis on subsets of CMO configurations, and used the additional analytic lens afforded by NPT. The findings on reviewing/reconciling medications therefore have a robust evidence and experience base to inform intervention proposals.

Conceptualising burden as a mechanism of interest has assisted data analysis thus far. In doing this, the Research Team acknowledge that it is also possible to locate burden as an outcome from the work done in any stage, as well as a context to what is being done. The value of this generative, terminological fluidity is acknowledged by Manzano,75 enabling analysis to reflect individual experiences of medication management and changes to those experiences from, for example, acute exacerbations of a particular diagnosis or combination of diagnoses, disease progression or age-related issues such as increasing frailty or cognitive decline.

Realist analysis of the complexity inherent in reviewing/reconciling medications has enabled the researchers to theorise about this stage in ways that are potentially transferable to other stages and circumstances based on the assumption that the same mechanisms may be in operation. In addition, the use of a realist approach has increased the understanding of the influence of contexts–mechanisms on outcomes, potentially contributing to better tailoring of medication management in different contexts.

Understanding burden mitigation in reviewing/reconciling medications

The MEMORABLE study has identified five burdens associated with reviewing/reconciling medications that are experienced primarily by older people but also by their informal carers: burdens of ambiguity, concealment, unfamiliarity, fragmentation and exclusion (see Chapter 3, Reviewing/reconciling medications and burden). Mitigation of these burdens occurs through implementation work, primarily involving sense-making, action and relationships. Reflection/monitoring were not identified directly from the interviews or the literature but are acknowledged as important aspects of implementation in this and all other stages. An absence of data in this aspect of the programme theory would be addressed in further research. Ways of mitigating specific burdens are discussed in the following paragraphs.

Risk identification

The MEMORABLE study has established the need to identify those who are not coping with medication management and who are at risk. From the literature and the interviews, this intervention should be a brief, simple and targeted risk assessment at any of the multiple points of contact with different practitioners that older people have as part of the medication management process. MEMORABLE found that this screening assessment should focus on identifying who is or is not coping with the burden of medication management, rather than assuming that all older people and informal carers are at risk. This needs to be a simple tool, with questions focused on the workload experienced by older people and informal carers and their capacity to cope with that workload. Such a tool would enable scarce health-care resources to be directed to the older people and informal carers most in need of support.

As it is only possible to outline this proposed intervention in MEMORABLE, it is recommended that scoping the detail and developing, implementing and evaluating a trial of risk identification is progressed through collaborative research to follow this study (MEMORABLE2). Involving all those with an interest, knowledge and skills in this area in realist evaluation would ensure the generation of an evidence and experience base on which development and roll out could be implemented.

Individualised information

The MEMORABLE study found that older people, informal carers and professionals need access to individualised information in an accessible format that can be shared by these key groups. It would provide a reference point to support older people’s and informal carers’ day-to-day decisions and actions, and increase the sense of control that they have over their medications. As a personalised working document, it could include, for example, a short summary of the older person’s values, preferences and goals; integrated and individualised information about their diagnoses and medications, individually and together, including links to trusted resources; their medication routine; key practitioners and contact information. It could be held in paper and electronic format, to be shared and updated as and when required. However, deciding on the most useful content and format would need to be co-created by those involved, with the input of others with expertise in how to make such systems practicable, such as through smart phone app (application) design.

As with the risk identification discussed above, it is recommended that, following co-production principles, future research should scope the detail of the required individual information, and then develop, implement and evaluate the tool, as part of follow-on research. This would be a significant amount of work, including determining lines of responsibility and accountability that would need to be completed before any trial.

Combining review and evaluation

The MEMORABLE study is a realist synthesis, integrating a review of the literature with an evaluation of the primary interviews, described in Chapter 2 and Figure 1. This combined approach is gaining popularity among researchers. It enables secondary and primary evidence to be brought together, facilitating the search for and the resolution of resonance and dissonance in their causal explanations of complex interventions. In MEMORABLE, the richness of the interview data was invaluable in supporting explanatory reasoning about the literature.

The MEMORABLE study benefited from the way the synthesis supported understanding of real-world complexity, being theory driven and purposefully evidence generating. In the research, preliminary analysis of the literature was used to inform discussion and debate by the Research Team. This guided further work on the literature review, including extending some of the search activity to address emerging themes and flexing interview schedules to gather interviewees’ opinions on the way medication management was beginning to be understood in the study.

Involvement of the Project Group and Stakeholder Group brought scrutiny and expertise to discussions of progress and the consideration of findings. In this way, the research progressed deductively by considering how the researchers thought medication management might work and then refining those ideas through data, as well as inductively by looking at emerging patterns in the data and generating theory from them. These processes were augmented by accessing the experience and expertise in the Research Team and in both groups to gain consensus on data-informed, causal explanations of patterns in the data. The iterations and associated opportunities for combinations of reasoning across a broad spectrum of experience and expertise provide a robust platform for researching real-world complexity.

Evidence and experiential data

An additional benefit of this combined approach to the research was the added value of interrogating secondary and primary data, moving backwards and forwards between them to ensure that one informed the other. In MEMORABLE, secondary sources provided access to broader, relevant and substantive data whereas primary sources illuminated the subject with finer-grained experiential perspectives.

Interviewing older people, informal carers and practitioners brought real voices directly into the research, informed by years of personal experience and expertise engendered by those experiences. This supported the co-creation of evidence. However, these interviewees were not a sample in research terms but contributed to theorising through their explanatory narratives. Their value was in what they described and explained, rather than who they might represent. In this type of real-world research, gaps can be identified between the way interventions are supposed to work, or are described as working when generic good practice is adhered to, and the way in which individuals actually experience them working in terms of the uniqueness of their own situation. Diverse data contribute directly to the analysis, its structure as well as its content. They also contribute indirectly by constantly challenging the researchers’ assumptions and methodological practices in the steps and iterations within the research process.

Qualitative data have strengthened MEMORABLE’s findings. A range of interviewees were invited to take part, reflecting the breadth of those involved in day-to-day medication management: older people, informal carers, practitioners from all sectors of health and care, including doctors, nurses, pharmacists, pharmacy technicians, social workers, and managers to front-line staff. The structure of the interview schedule enabled interviewees to describe their own experiences of multimorbidity and polypharmacy and to account for the way in which medication management worked for them at that time, and how it might work better in the future.

Approximately 60 hours of audio-recordings were collected and analysed in combination with the literature. From these, 59 initial CMO configurations were developed as well as five additional CMOs explaining burdens in reviewing/reconciling medications, informing a framework for burden mitigation (programme theory). Data saturation was achieved and no new themes were generated from the last two or three interviews. The realist analysis, subsequently conducted, confirmed that data saturation had been obtained.

The MEMORABLE study has accumulated a rich, qualitative data resource. In addition to writing up the research, its methodology and findings, the researchers intend to revisit it with a view to publishing additional articles that explore multiple themes of interest such as perspectives on the emerging role of pharmacists and the experiences of informal carers.

Theorising about causality

Theorising about how interventions work is central to realist methodology. In MEMORABLE, this involved:

• setting out candidate theories (see Chapter 1, Step 4: candidate programme theories and substantive theory)

• using substantive theory (NPT) to inform the analysis (see Chapter 3, Medication management as implementation)

• generating CMO configurations to explicitly and transparently account for the multiplicity of ways interventions work for some people under certain conditions (see Chapter 3, Medication management as experience and Reviewing/reconciling medications and burden and Appendix 8)

• developing a more narrowly focused final programme theory from a subset of those CMO configurations to explain Stage 5: reviewing/reconciling medications (see Figure 5).

Although substantive theory brings an additional analytic ‘lens’ to the synthesis, theorising in realist methodology is undertaken flexibly. In this way, it supports rather than controls analysis. Programme theory is set out to be confirmed, refined or refuted through the research process. CMO configurations are modifiable as meaning becomes clearer; one person’s outcome can be another person’s context and, even for one individual, contexts, mechanisms and outcomes can shift as the intervention process advances. This can feel messy and daunting.

In MEMORABLE, having a clear process, the work packages and steps by which the research progresses, was important. Process iterations occurred in response to the data with the Research Team mindful of the overall structure and the place of theorising within it. Having a methodology workshop at the start of the research was key to developing a shared knowledge base and an appreciation of realist methodology in the Research Team and Project Group. However, for theorising, having access to high-level realist knowledge and skill in the team and group, and regular meetings where data interpretations and analyses were discussed and debated, has been invaluable and are processes that should be considered by all realist research teams.

Strengths

As far as the Research Team is aware, this is the first realist synthesis of medication management combining realist review and realist evaluation to understanding this key clinical area. As part of the research process, MEMORABLE has developed two key outputs:

1. a five-stage, four-step, three-loop medication management process that provides a structure to address intervention complexity, highlighting the cycles of interactions between individual and interpersonal stages (see Table 5, further developed in Tables 6 and 7)

2. a more detailed and focused programme theory of Stage 5: reviewing/reconciling medications that sets out the way that this key stage can be understood as implementation as a burden, explained by the evidence and informing proposed interventions (see Figure 5).

In addition to these outputs, the researchers have also obtained rich, qualitative data from diverse interest groups: older people, informal carers and health and care practitioners. They have set out an evidence-informed analysis framework to support their use of a realist logic of analysis and integration of data from the literature and interviews. They have also demonstrated the way substantive theory (NPT) has added explanatory value to the analysis. Analysis has generated proposed interventions that are evidence and experience based, and theory informed.

The MEMORABLE study has demonstrated the value of following a clear research process that enacts the research methodology, while maintaining the flexibility to respond to emerging topics of interest. It has also shown how the structures, processes and broad membership of Research Teams and groups supports accountability and transparency in decisions and in the scrutiny of outputs at key stages. This study has also added to practitioner capacity in realist methodology and is already being disseminated as an example of the application of that methodology.

Limitations

Few studies directly address the complexity of medication management as a process and how it works. Therefore, the researchers had to draw more heavily on primary data to validate causal attributions being made in the analysis and in the generation of programme theory. As a formative piece of research, grappling with that complexity proved time-consuming. As a result, the researchers decided that addressing causality across the whole medication management process was likely to be beyond the scope and resources available to them. The researchers had to make judgements on which of the stages of this process were more important to focus on. Analysis was, therefore, not extended or integrated across all five stages of the medication management process. Extending the analysis in the future would build useful knowledge of medication management and the development of evidence- and experience-based interventions that might bring about improvements. In addition, as acknowledged earlier, the framework for the interventions requires further development.

The decision to focus on Stage 5: reviewing/reconciling medications was not taken until part way through the research. This was because the researchers needed to have worked through what data were available to them before identifying that reviewing/reconciling medications (Stage 5) was the key part of the process to focus on. This ‘delay’, caused by a need to refocus a realist review, or evaluation, is a potential challenge for realist researchers and can have an impact on the conduct of a project. In this project, the impact was to delay the search for literature on these interventions and limit the opportunity to focus some of the interview questions on this stage.

Other limitations include that the search was limited to English-language literature, the research focused on non-institutionalised populations and the search had a broad definition of ‘older people’. Furthermore, only a relatively limited number of people with mild dementia were interviewed.

Relevant articles: by date and content

Additional terms identified from the remaining records

The 35 abstracts were then mapped onto an Excel database (AB/SL Search Database, tabbed Pilot Search), identifying the record number from the search, author, title, topic, design, number of subjects and country.

The contents of each abstract were then analysed for additional terms that added to the scope of the original four search terms noted above or because they appeared to have explanatory or causal significance in relation to the research topic. These were identified in the content of the abstract and the key word listing. As a result, the four combined search terms were expanded to 56, and grouped in three ways: those associated with individual experiences, whether older people and carers, or practitioners, and those associated with health and care practice and systems that were often highlighted in the articles’ recommendations (as below).

The following bullet points provide a sense of the data, setting out the top three clustered terms occurring in each group:

• Individual experiences (older people and carers)

  • High risk patients/Health status and complications/Chronic diseases/Chronic conditions/Long-term conditions/Co-morbidity/Multimorbidity/Progressive/Diagnoses/Other health conditions/Disability/Chronic illness trajectory/Restrictive health care conditions (n = 32)

  • Medication adherence – nonadherence: (n = 22)

  • Complex medication regimens/Dosing frequency/Schedules of medication-taking events and Pill burden/Frequent changes/New medications/Inconvenience/Taking and timing adherence/Adjusting routines (n = 21).

• Individual experiences (practitioners)

  • Drug therapy (unnecessary, ineffective, incomplete)/Dosage/Pharmaceutical preparations/Medication therapy management (n = 6)

  • Discussion/Shared decision-making/Effective communication/Decision-making/Time/Priority reconciliation/Health professions interest-disinterest in balancing diagnosis, regimen and life (n = 6)

  • Collaborative approaches/Patient centred care (n = 5).

• Health and care practice and systems

  • Person centred approach/Decision support systems/Patient care/Personal support/Patient–provider relationship/Teach-back methods/Coaching (n = 6)

  • General practitioners/Physicians practice patterns/Socio-economic influences/Doctor-patient/Physicians role (n = 6)

  • Community pharmacist/Professional role/Pharmacist (n = 4).

Sift terms and a data extraction form

Based on this analysis, an indicative (rather than definitive) list of terms is proposed below, identifying those associated with inclusion as well as exclusion.

Under inclusion, these are grouped under core terms and supplementary terms. The core terms reflect the scope of the research to ensure that abstract content aligns with the research topic. This frames the search. The supplementary terms encompass the different levels, interests and perspectives within the research: individual older people and carers, individual practitioners, and health and care practice and systems. Applying these permutations enabled further searches to explore different aspects of the literature, following separate but integrated lines of thinking about this complex topic (e.g. core terms + combined older person/carer perspective terms to explore their perspective and experiences; or core terms + shared decision making and decision aids/decision support systems; or core terms + deprescribing to focus on particular interventions that might be included in the final research output, i.e. the framework to support improved medication management).

Core terms

• aged; aged, 80 and over; independent living; home/community

• chronic disease; comorbidity; long-term conditions; multimorbidity

• polypharmacy; multiple prescribed medications

• complex medication regimens; schedules of medication-taking events

• medication management; medication compliance; medication adherence; medication concordance; medicines optimisation; patient compliance.

Supplementary terms: individual – older person/carer perspective

• patient acceptance of health care; health literacy; health belief model; belief about medications

• self-care; self-management; self-administration; medication knowledge

• physical/cognitive/emotional capacity; motivation; intentional/unintentional nonadherence

• social support (family/friends); interpersonal relationships

• treatment burden/pill burden; medication related health regimens; lifestyle changes

• patient reported adverse reactions/events/intolerance; drug/treatment related side effects; safety

• patient goals/outcomes; lived experience; life style; sense of normality; goal conflict with clinicians

• medication-related A&E visits/hospital admissions or re-admissions/length of stay.

Supplementary terms: individual – practitioner perspective

• evidence based medicine; practice guidelines/guideline directed interventions; cumulative treatment recommendations; risk benefit comparison

• patient centred care; shared decision-making; decision aids/decision support systems; patient information/education; coaching; practitioner/professional role/relationships

• medication errors; inappropriate prescribing/dosing

• Home Medication Review; home visits; telephone interviews

• deprescribing

• health related quality of life; treatment outcome/adverse health outcomes; outcome data

• general practitioner; practitioner practice patterns

• nursing staff/nurse practitioners; community matron

• community pharmacist; pharmacist

• home care/aides.

Supplementary terms: health and care practice and systems perspective

• continuity of care (records, plans, medication, team approach); transitions; medication reconciliation

• telehealth; personal health applications.