A rapid review of interventions to improve medicine self‐management for older people living at home

Abstract Background As people age, they are more likely to develop multiple long‐term conditions that require complicated medicine regimens. Safely self‐managing multiple medicines at home is challenging and how older people can be better supported to do so has not been fully explored. Aim This study aimed to identify interventions to improve medicine self‐management for older people living at home and the aspects of medicine self‐management that they address. Design A rapid review was undertaken of publications up to April 2022. Eight databases were searched. Inclusion criteria were as follows: interventions aimed at people 65 years of age or older and their informal carers, living at home. Interventions needed to include at least one component of medicine self‐management. Study protocols, conference papers, literature reviews and articles not in the English language were not included. The results from the review were reported through narrative synthesis, underpinned by the Resilient Healthcare theory. Results Database searches returned 14,353 results. One hundred and sixty‐seven articles were individually appraised (full‐text screening) and 33 were included in the review. The majority of interventions identified were educational. In most cases, they aimed to improve older people's adherence and increase their knowledge of medicines. Only very few interventions addressed potential issues with medicine supply. Only a minority of interventions specifically targeted older people with either polypharmacy, multimorbidities or frailty. Conclusion To date, the emphasis in supporting older people to manage their medicines has been on the ability to adhere to medicine regimens. Most interventions identify and target deficiencies within the patient, rather than preparing patients for problems inherent in the medicine management system. Medicine self‐management requires a much wider range of skills than taking medicines as prescribed. Interventions supporting older people to anticipate and respond to problems with their medicines may reduce the risk of harm associated with polypharmacy and may contribute to increased resilience in the system. Patient or Public Contribution A patient with lived experience of medicine self‐management in older age contributed towards shaping the research question as well as the inclusion and exclusion criteria for this review. She is also the coauthor of this article. A patient advisory group oversaw the study.

European cohort study found that 32% of citizens aged 65 years or older were taking five or more medicines. 8 A cohort study in Sweden found that 44% of those aged 65+ took at least 5 medicines and 12% took 10 medicines or more 9 , and a recent US cross-sectional study of patients 65 years of age and older found that 37% were prescribed five or more medicines. 10 In the United Kingdom, a Scottish study found that 35% of those aged 85 years and older take more than 10 medicines, 11 while the most recent data on medicines prescribed within the National Healthcare Service (NHS) in England indicated that, by the age of 80, a third of the population takes eight or more medicines. 12 Polypharmacy has been linked to a range of negative outcomes in older people, including drug-related problems, adverse drug events, impact on physical and cognitive function, hospitalization and mortality. 13 Managing medicines poses multiple challenges to older patients and those who support them. Older patients, for example, have been found to face difficulties in filling prescriptions, especially after an unplanned hospitalization, in reading and understanding medicines' instructions, in opening containers and retrieving medications, in swallowing pills, in taking the right medicines at the right time and in detecting and reporting adverse reactions. 14 Moreover, older patients prescribed many medicines are likely to have to manage a complex medicine regimen, often with no or limited support. According to George and colleagues, 15 regimens are complex when medicines come in different forms (e.g., pills, injections, drops), have different dosing frequencies and come with additional instructions that guide administration. Evidence from a Swedish cohort study has shown that, in patients aged 60 years and older, medication regimen complexity is associated with increased mortality. 16 Among the older population, patients living with frailty have been found to be particularly vulnerable to adverse events associated with complex medicine regimens. 17 Described as an ageing-related process in which multiple body systems gradually lose their in-built reserves and become increasingly vulnerable to relatively minor stressor events, 18 frailty is estimated to affect around 10% of people over 65 years. 19 Thus, there is a clear need to improve older patients' abilities to self-manage their medicines as well as their abilities to detect and respond to problems, to prevent deterioration. Bailey and colleagues define medicine self-management as 'the extent to which a patient takes medication as prescribed, including not only the correct dose, frequency and spacing, but also its continued, safe use over time'. 20 This requires a range of knowledge, skills and behaviours and patients must follow six steps to safely and effectively use their medication in primary care: (1) filling (getting hold of the medicines prescribed at the right time), (2) understanding, (3) organizing, (4) taking, (5) monitoring and (6) sustaining. More recently, Howell and colleagues 21 focused on the functional skills required by patients to manage medicine preparation and administration. In their realist synthesis, Maidment and colleagues 22 looked at how older people, family carers and healthcare professionals (HCPs) engage in medicine management at five different stages: (1) identifying problems, (2) getting diagnosis/medications, (3) starting/ stopping medications, (4) continuing to take medications and (5) reviewing medications. Finally, Cheraghi Sohi, Schafheutle and their colleagues 23,24 looked at medicine self-management as a workload carried out by patients, often with support from their family and their informal network, 25 in four main areas: medication articulation work, surveillance work, emotional work and informational work.
While managing many medicines poses multiple challenges, older patients and patients with long-term chronic conditions are able to overcome difficulties and prevent harm caused by medicines. 26 Adopting a Safety II (Resilient Healthcare) 27 approach, Fylan and colleagues 28 looked at how cardiology patients managed their medicines after discharge from hospital. They found that some patients contributed to medicine management system resilience through monitoring and responding to supply problems, anticipating discrepancies and notifying HCPs of errors in documentation. Some anticipated problems with their own adherence and put mitigating strategies in place. Tomlinson and colleagues found that some patients and their support networks were proactive, for example, in seeking additional information on new medicines introduced or facilitating communications between hospital and pharmacy to mitigate disruptions in their supply. 26 Whilst such observational studies have identified some people's abilities to implement safety strategies, there is a lack of evidence concerning the range of interventions that support safe medicine self-management activities, such as adherence to the regimen, knowledge of medicines and condition, supply management and monitoring effects. 29 Whilst there is a wealth of evidence about how healthcare staff can reduce the risk of adverse drug events, for example, through deprescribing, medication review and staff training, 30,31 there is less focus on developing interventions to improve the role that patients themselves can play in the safety of medicines.
Interventions that build on enhancing the way some older people safely self-manage their medicines at home may have a positive impact on their health outcomes. This rapid review, therefore, aims to explore the range and nature of interventions that support medicine self-management for older people. We wish to understand the selfmanagement components that they target and the abilities they aim to support patients to develop, so that gaps in patient support can be identified and addressed.

| METHODS
A rapid review 32 was undertaken to inform intervention development. 33 A rapid process was chosen as a pragmatic approach to generate evidence to inform empirical research and codesign activities in a timely way. Only published papers in English were included and no searches were conducted for grey literature or citations. 32 Measures were adopted to mitigate the risk of excluding relevant papers and to maintain consistency in applying the inclusion criteria among the team. The measures included testing and refining the screening checklist using a limited number of papers and multiple researchers screening a sample of excluded papers at each stage. 32 Articles were excluded if they were an editorial, a research protocol, a thesis, a conference paper or a literature review. Only articles written in English were considered. Articles were excluded if they focused on improving medicine management at care transitions because there is already a rich literature around interventions to support older people at hospital discharge or moving between different care settings. 35 Our focus in conducting this review was on interventions addressing the less frequently explored topic of selfmanagement of medicines and its components (e.g., adherence, knowledge, supply management and monitoring effects) 29 in the community.
This review was conducted as the first stage of an intervention development process for people with polypharmacy aged 65 years and over who are not living with dementia or cognitive impairment. It was therefore necessary to exclude papers focussing on people with dementia. Also, a systematic review of evidence assessing interventions to support medicines management in this population has recently been published. 29

| Review process
A team of 10 researchers conducted the review. For consistency, one researcher (G. P.) was involved in all the stages. Title screening was conducted by one reviewer only (G. P.), adopting a conservative PREVIDOLI ET AL. approach, and at least a third of the excluded papers were screened by another reviewer 32 (either R. C. or R. H.). Abstract screening was conducted by eight different researchers (G. P., B. F., V. C., D. A., D. O., C. P., J. S., G. W. P.); references were split among them, who all used the same screening checklist. Before that, each item in the checklist was discussed with the team in a meeting, using practical examples to reach a shared interpretation. At least 50% of abstracts were screened by a second reviewer (either R. C. or R. H.), to ensure consistency and avoid bias.
A screening checklist was designed for full-text screening and differed from the abstract checklist, with the second being more restrictive. Exclusion criteria relative to type of publication (no protocol study paper, no conference abstract) were added. G. P. and B. F. tested the checklist on five articles. After team discussion, the checklist template was improved and then used by all reviewers involved in full-text screening (G. P., B. F., V. L., D. O., G. W. P.). At least 50% of papers were screened for full text by more than one reviewer.
Rayyan ® , a web app to conduct collaborative literature reviews online, 36 supported the work of the team. Both during abstract screening and during full-text screening, disagreements were resolved via discussion. On Rayyan ® , disagreements between researchers become visible in real time. This allowed G. P. to promptly address them and facilitate discussion between researchers.
Resolved disagreements were also discussed as examples at team meetings to improve consistency. If researchers suspended their judgement, the final decision was made by BF, the principal investigator of the funded study of which this review is part. 29 A table was created by G. P. and B. F. in Microsoft Excel to extract the data. G. P. identified and selected information relevant to the research question to be extracted and rephrased. The form included author (s), year of publication, country, aims of intervention, study design, length, participants, outcomes and measures, description of intervention, aspects of medicine self-management addressed and delivery. The form was tested on 20 papers by G. P. and then refined after discussion with B. F. Subsequently, additional columns were added: 'type of intervention' (e.g., medicines review and coaching) and 'effect of intervention' and the column 'delivery' was split into two ('where' and 'by whom'). Data extraction was conducted by G. P. and assessed for accuracy by B. F. Data were also extracted about patient capabilities that the intervention sought to enhance, based on Hollnagel's Safety II resilient abilities: the ability to monitor: to respond; to anticipate; and to learn. 37

| Data synthesis
Heterogeneity of data sources (e.g., different study design, different outcome measures) as well as the aim and the time scale of this review made quality appraisal and statistical meta-analysis not suitable or necessary. 32,38 Where feasible, indication of the outcome measures and the effect of the intervention were reported ( Table 2). A narrative synthesis approach was applied to summarize the interventions identified and their components (e.g., grouping together studies with similar characteristics, categorizing, describing outcomes and context) 32 and to identify gaps. Data synthesis was undertaken by G. P. reviewed by B. F. for completeness. B. F. and J. T. independently checked a subset of articles where aspects of the intervention addressed one or more resilient abilities. 37

| RESULTS
A to tal of 14,353 records were identified and 14,320 were excluded.
The process followed for the screening is reported in the PRISMA diagram ( Figure 1).
Thirty-three studies were included in the review (listed in Table 1). Two articles 39,40 are about the same international trial; however, the first 40 only presents the results from Northern Ireland (United Kingdom) and was found through database search, and the second 39 presents the overall results from seven countries and was found cited in the first paper ( Figure 1, PRISMA diagram). The rational to include an additional article that was not found through database searches was to provide a better understanding of the intervention.
Details related to geographic area of intervention, intervention type, sample and setting characteristics are reported in Table 1.

| Country of intervention
The majority of studies (n = 17) were conducted in the United States, two in the United Kingdom (one in England and one in Northern Ireland), two in South Korea and one each in Sweden, Japan, Australia, New Zealand, Hong Kong, Italy, Canada, China, Thailand, Croatia and Taiwan. One intervention 39,40 was developed and implemented by an international network of seven European countries.

| Target and population
The intervention was offered to informal carers/medicines supporters as well or on behalf of patients in only 7 out of 33 articles. Where reported (in 26 out of 33 articles), the mean age of the participants varied from 67.8 years (SD = 7.8) 47 to 82.7 (SD = 6.4). 46 In a minority of studies, the intervention explicitly targeted patients taking multiple medicines 39,52,55,57,61,70 or living with multiple chronic conditions. 42,55,66 Some interventions targeted people with specific conditions such as congestive heart failure, 44 hypertension, 49,59 diabetes 69 and COPD 62 or taking specific types of medicines, e.g., anticholinergic, 47 anticoagulants 71 or antacids. 58 Only one of the studies identified explicitly targeted older people living with frailty. 63

| Who delivered the intervention and where
The majority of interventions (n = 20) were delivered in people's homes.
Details related to study design, intervention length, aims and outcomes of intervention, aspects of medicines management addressed and indication effect of intervention are reported in Table 2.

| Intervention aims and effect
In the majority of the interventions (n = 19), improving patients' adherence was one of the aims (n = 13) or the only aim One-to-one warfarin therapy review + education.

Not specified
Abbreviations: IQR, interquartile range; Medi-Cog, medicines cognition and ability to safely manage medicines.
T A B L E 2 Author, year, study design, intervention length, measures, description of intervention, aspects of medicine management addressed, aims, outcomes and effect of intervention.
Authors Measures were taken before and after the intervention.
Self-efficacy: General Selfefficacy Scale. Quality of Life (n = 6). 43,46,48,49,57,65 Of all interventions aiming to improve adherence, four did not report complete outcome measures 41,42,50,65 and one 45 reported that changes in adherence were not significant. All the remaining studies (n = 14) reported significant differences in adherence scores, either compared with baseline, or between the intervention and control group(s) in measurements taken after the intervention (see Table 2). One of those 49 found that although in the intervention group the adherence score increased significantly at 4 weeks, the increase was not sustained at the 5-month follow-up.
Of the four studies that aimed to increase communication with HCPs, 42,47,54,56 one reported that after the intervention, participants started conversations with physicians, 47 one registered increased interactions with pharmacists, 54 one found that more patients planned to bring an updated list of medicines at their next visit to a pharmacist 56 and one did not report outcome measures. 42 All interventions assessed in this review aimed to achieve a change either in participants' knowledge, attitudes, behaviours, selfefficacy and quality of life or a combination of them, but some also reported systemic goals, 39,52,63,68 such as reducing admissions to care homes or hospital. Significant changes reported by those included decrease in cost of care after a home medicines review 40 and decrease in healthcare use after an education programme. 68 Finally, a prospective cohort study 63 found that participants using medication calendar cards were 66% less likely to be admitted to a nursing home than nonusers.

| Intervention characteristics
The six interventions that solely aimed to increase adherence were either based on new dosage packaging, 57 dispensing systems with reminders, 46,63,65 new ways to provide administration instructions 43 or coaching programmes to enable patients to develop medicinetaking routines. 48,49 Most interventions that sought to improve knowledge 39,41,42,45,50,51,60,64,66 (n = 9) aimed to increase the general understanding of medicines and medicine regimens, for example, reasons for taking them, dosage, timing and additional instructions.
Some also aimed to enhance medicines' safety knowledge. 42,60,66 Patients, for example, learned how to recognize side effects, 60 avoid mistakes in dosing and timing and how to repeatedly check their medicines, 66 including the medicines received from the pharmacy.
A number of interventions focused on increasing participants' knowledge of specific topics (n = 7) such as self-management and therapy of specific conditions, 59,62,69 risk associated with medicine poisoning, 67 risk of medicines and alcohol interactions and 58 anticholinergic medicineassociated risks. 47 Finally, one intervention aimed at improving knowledge of the role of pharmacists in medicine management. 54 Two studies compared the effectiveness of different educational materials, such as playing a game 67 or watching a video tutorial 62 versus reading leaflets on the same topic, with the more interactive mode showing better results.

| Mapping resilience abilities
To determine which resilience abilities 36 Table 3.
Most studies (n = 26) supported participants in learning about medicines and conditions. Some focused on preventing risks associated with medicine use. This included learning how to use adherence aids, to avoid errors, how to keep updated lists of medicines, how to store and dispose of medicines and how to communicate effectively with HCPs about medicines. Other studies (n = 5) supported patients' ability to monitor medicine supply, 66 problems with medicines through a list shared with HCPs, 41 adherence through self-monitoring tools 48,49 and tracking logs. 55 Two studies supported patients' abilities to respond to medicine concerns or risks through starting a conversation with a GP, 54 including about potential de-prescribing. 47 Details of the topics covered and resilience abilities addressed are shown in We have identified clear gaps, based on the evidence about patient experiences with their medicines and on the proactive strategies that they implement. 28 These are as follows: − Interventions that support people to navigate the complexity of the medicine management system (e.g., where there are multiple prescribers and frequent changes). − Interventions that prepare people to anticipate and respond to errors in the system (e.g., receiving incorrect medicines).
− Interventions that support people to detect problems in their medicine self-management abilities, recognize when they may be deteriorating and ask for help. Our review found that medicine self-management interventions do not generally attempt to augment patients' proactive safety roles.
Interventions aimed at other areas of healthcare have attempted to enhance patient roles. Lawton et al., 72 for example, evaluated an intervention to support patients to report safety concerns in hospital.
They concluded that, while patients were willing and able to report concerns, staff also needed support to respond to patient feedback.
More recent work has indicated that people play important roles in the care that they receive for chronic healthcare conditions. Systematic review has demonstrated that a lack of involvement in care is not a desired state for patients; rather, it is one that is forced on them by the system and the context in which care is provided. 73 Patient activation as a concept has gained traction since Hibbard et al's measure was first introduced in 2004. 74 Our review, however, shows that little has been explored to support patient activation to be proactive partners in the safe use of medicines.
Previous work has highlighted how the safety-critical role of patients and carers is overlooked 75

| Informal carers
Only in 7 out 33 articles targeted informal carers as well as patients; only one was delivered to the family network of supporters, instead of to patients. 69  | 983 support network has not been explored in any of the articles found in this review. Research conducted with patients with long-term conditions, however, indicates that such relations are complex and ever changing, with the distribution of tasks constantly re-negotiated.
Interventions are needed that take into account the multiple and ever-changing roles that patient support networks play in medicine management.

| Frailty, multimorbidity and polypharmacy
Only one intervention was found that specifically targeted older people with frailty. People living with frailty have been found to be particularly vulnerable to the consequences of inappropriate polypharmacy and medicine errors, and interventions can support patient roles in how medicines are managed in this population. 79 More interventions are required to support this population to manage polypharmacy, recognize when their symptoms might be deteriorating and avoid unnecessary stressors that can cause further deterioration. 79,80 Complex medicine regimens, polypharmacy and multimorbidity in older populations have been associated with negative health outcomes. 13,16 Conversely, it was surprising to find that only 7 interventions targeted people either taking three or more medicines or living with multiple long-term conditions. We recommend that in the future, interventions to support self-management of medicines prioritize frail and/or multimorbid populations dealing with complex medicine regimens.

| Defining medicine self-management
A new, more comprehensive understanding of medicine selfmanagement components, co-designed with patients and their network of supporters, is needed to design interventions that are grounded in the knowledge of the processes involved and embody the patient perspective. 29 Informed by the literature in this review and by our patient advisory group, we propose the following (Table 4) as components for safe self-management of medicines: Managing many medicines can be challenging and requires a wide range of skills. Each time a new condition is diagnosed, or a symptom deteriorates, patients need to adapt the way they cope.
Living with multiple long-term conditions demands increasing efforts to organize and remember multiple medications, to manage healthcare appointments, to perform self-care and change lifestyles and routines. [81][82][83] Interventions that can be adapted over time are therefore preferable, as they will be able to provide support both when people's needs change (e.g., loss of dexterity or vision) and when new challenges or opportunities arise in the healthcare system (e.g., medical practices accepting only medicine orders online). An adaptive design could be explored to enhance responsiveness. 84 In the interventions identified by this review, education on medicines and condition self-management was delivered in a variety of ways, including using software, video tutorials and Apps. Future interventions need to carefully consider the impact of digital exclusion on the most vulnerable in the older population (e.g., over 70, living alone, on lower income), 85 to avoid making health T A B L E 4 Self-management of medicines-Key components.

Knowledge of medicines
To understand what medicines have been prescribed and why, including how and when to take them

Managing supply
To understand how to access supply (prescription journey, repeat prescriptions, impact of changes in medicines, knowing when is the right time to order, dealing with inputs from different prescribers -e.g., GP and specialists) To monitor medicine supply, anticipating problems and knowing how to respond to unexpected events (delays, errors, changes not actioned in prescription).
Monitoring how you feel To monitor medicines' effects and side effects, especially when changes are introduced and knowing how to respond (e.g., knowing what to do in case of an adverse event).
Ensuring that medicines are taken as instructed To develop ways to monitor that medicines are taken as instructed (self-monitoring of adherence).
To develop ways to overcome barriers to adherence (e.g., special containers, calendar reminders, daily routines), anticipating where adhering might be difficult, for example, when routines are disrupted.

Communication with healthcare professionals
To communicate effectively with healthcare professionals about medicines, including identifying and targeting the relevant professional, feeling confident about asking questions and feeling empowered to make informed decisions.
Self-monitoring and involving other people To monitor their own self-management skills, learning to know where help will be needed and delegating tasks to others, if self-managing becomes difficult.
To be able to share instructions and information related to their medicines with people (family, friends, professional carers) willing to help.

Adaptability
To be able to cope with ever-changing circumstances (change in symptoms, in doses, brand, timing) and constantly adjust.
inequalities worse. The use of participatory methods would, in our view, ensure that the content addressed by the intervention is relevant 86 and the mode of delivery is appropriate to people's abilities and circumstances.
Finally, an intervention to support older patients around their medicines, in the first place, needs to avoid placing additional burden on their shoulders, starting from considering what patients and their networks already do and respecting them for 'what they do'. 87 Medicine regimens should be person-centred to take into account patient values and preferences as well as the effects of multiple comorbidities and social circumstances. 88

| Strengths and limitations
To our knowledge, this is the first rapid review to characterize interventions aimed at enhancing how older patients self-manage their medicines from a Resilient Healthcare perspective. A recent systematic review looked at the resilient abilities addressed by interventions to support medicine self-management at home, but only included studies targeting patients living with dementia or cognitive impairment. 29 A previous review investigated RCT interventions to support older people's ability to take medicines and adherence, when more than 4 medicines were prescribed. 89 This is the first review to include both a variety of study designs and interventions addressing a range of aspects related to medicine self-management.
This review did not include a citations search or grey literature. A quality assessment and risk of bias assessment were not performed because they were not necessary to meet the study aim. A narrative approach was chosen to describe the results and no quality assessment was conducted on the interventions included. Nevertheless, this review offers useful insight into the aspects of medicine self-management targeted by interventions aimed at an older population. Adopting a Resilient Healthcare perspective, finally, this review offers a new definition of medicine self-management that we hope could inform interventions that enhance older people's resilience capabilities.

| CONCLUSION
Few interventions were found that address the full range of challenges that older people face in self-managing medicines and only one that specifically targets older people living with frailty. Most identify deficiencies within the patient, rather than preparing them for problems inherent in the medicine management system, and address their knowledge, attitudes or behaviour around medicines to