Measures and tools assessing medication self-management capability in older people across the hospital-to-home transition: a systematic scoping review

Abstract

Background

Adverse drug events from medication-related harm (MRH) can lead to hospital readmissions, compromised quality of life, and even death. Post-hospital discharge is a vulnerable period for older adults, who are often unprepared to resume self-care and medication self-management. Assessing medication self-management capability in older people can guide supportive interventions and improve medication-related outcomes. This review aimed to identify measures and tools used to assess medication self-management capability for older patients during the hospital-to-home transition.

Methods

Medline, EMBASE, PsycINFO, CINAHL and Cochrane Library of Systematic Reviews were comprehensively searched for articles from database inception to December 2023. Eligible studies included participants aged 65 or older across the hospital-to-home transition and measures containing at least one medication self-management component. Data extraction was performed using a standardised form, characteristics of measures tabulated, and a narrative approach used to describe measures. Reporting conforms to the Preferred Reporting Items for Systematic Reviews and Meta-analyses extension for Scoping Reviews checklist (PRISMA-ScR).

Results

Fourteen studies were included, and 12 unique measures identified. Measures predominantly focused on adherence rather than broader medication self-management components. Timing of measure administration and the individual administering the measure varied greatly. Medication self-management capability was determined through assessment of physical and cognitive skills. Number and type of skills assessed varied between measures. No measures considered all medication self-management components.

Conclusions

Current measures for medication self-management capability assessment primarily focus on cognitive and physical skills, with significant emphasis on adherence. Findings emphasise the importance of comprehensive definitions of medication self-management across the hospital-to-home transition. Recommendations are provided for developing future measures.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12877-025-06001-9.

Background

Despite substantial efforts worldwide to reduce preventable medication-related harm, more progress is required [1, 2]. In England alone, it is estimated that 237 million medication errors occur annually, costing over £98 million [3]. Medication-related harm can lead to increased healthcare utilisation, reduced quality of life, serious harm, disability, and even death [4, 5].

Additionally, the global population is ageing [6] and older people are more susceptible to risks associated with medication-related harm [3]. For instance, they are more likely to have multi-morbidities and polypharmacy [7, 8] alongside frailty and poor functional output [8, 9].

Transitions of care are critical junctures in the healthcare journey, where older people are particularly at risk of MRH [7]. In the UK, 37% of adults 65 years and older utilise health services due to MRH during the early weeks post-discharge [10]. Recent evidence shows that most incidents in England and Wales occur in older patients, with almost one-eighth associated with patient harm [11]. Thus, being able to self-manage medication is critical for safe and independent living [12].

Medication self-management has been defined by Bailey et al. [13] 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” [13]. This focus on adherence is echoed in a systematic review [14] of interventions designed to improve medication self-management. However, other aspects of self-management are also important. Previdoli et al. [15] describe medication self-management as consisting of six components: managing supply, monitoring how you feel, ensuring medicines are taken as instructed, communication with healthcare professionals, self-monitoring and involving other people, and adaptability.

Importance of measuring an individual’s ability to manage medication is well-recognised [4, 10, 16] and over recent years, several different measures, tools and instruments have been identified [17–21] often with a focus on the functional and intellectual aspects of a patient’s capacity to manage medication [17–21]. However, the measures vary in terms of their dimensions, and validity and reliability, with little evidence of use in clinical practice [22]. Interestingly, despite the literature investigating the same subject, the terminology used varies between ‘measure’, ‘instrument’, and ‘tool’.

Badawoud et al. [17] provided a recent review of instruments which assess patients’ ability to manage medication. This scoping review identified existing validated instruments assessing ability to independently self-manage medication in adults, adding to findings from those previously conducted [18, 19]. However, Baby et al. [23] further demonstrated that measures fail to consider motivational and environmental factors as well as cognitive, physical, and sensory skills. Furthermore, existing reviews have not studied measures designed specifically for older people at the hospital-to-home transition. Tomlinson et al. [24] propose the importance of considering behaviour change theory when designing interventions to support older people to manage medication post-discharge. It therefore, seems appropriate to apply a behavioural perspective in this review, which has not previously been considered, specifically through use of the Behaviour Change Wheel (BCW) [25]

The BCW holds the Capability, Opportunity, and Motivation Model of Behaviour (COM-B model) [25] at its core, providing a simple framework; capability, opportunity and motivation are three essential conditions required for a behaviour to occur. This is expanded by use of the Theoretical Domain Framework (TDF) [26], which subdivides COM-B components into 14 domains, providing a more granular understanding of behaviours (See Fig. 1). Application of the COM-B model and TDF to existing measures of medication self-management capability will enable a more comprehensive understanding of the behaviours currently being assessed, providing further insights into what is currently being measured and what is missing. A behavioural approach was therefore be employed, as we hypothesised that measures which are underpinned by the behaviours required to enact medication self-management, may be more accurate and informative.

Fig. 1.

The relationship between the COM-B model and TDF

Successful assessment of medication self-management capability could potentially guide the supportive interventions an older person is offered on discharge. Involving and engaging older people with their medication management during hospital admission and post-discharge is essential [15, 27]. A first step to achieving this is understanding whether measures exist that assess medication self-management capability in older people across transitions of care.

Therefore, the primary aim of this review was to use a systematic approach to identify and characterise these measures, mapping out existing literature, hence a scoping review approach was chosen. The review also seeks to: understand how measures are used in research or clinical practice; identify the components of self-management that are assessed when mapped to domains of the COM-B model [25] and TDF [26]; and describe the theoretical foundations of measures identified. This review adds to existing knowledge, providing up-to-date information on measures that healthcare professionals can use to understand an older person’s medication self-management capability across the hospital-to-home transition.

Methods

Search strategy and selection criteria

The reporting of this systematic scoping review conforms to the Preferred Reporting Items for Systematic Reviews and Meta-analyses extension for Scoping Reviews checklist (PRISMA-ScR) (see Supplemental File 1). A protocol was developed a priori and is available upon request from the corresponding author.

A comprehensive search for published articles was conducted across electronic databases (Medline, EMBASE, PsycINFO, CINAHL and Cochrane Library of Systematic Reviews). Searches included articles from database inception to December 2023. A forward and backward citation search was conducted on included articles. Additionally, a Google Scholar search was limited to the first 100 results and the register, PROSPERO, was also searched. Grey literature was searched via the Royal Pharmaceutical Society, Royal College of Physicians, Department of Health and Social Care, National Institute for Health and Care Excellence and World Health Organisation.

Medical Subject Headings (MeSH) and search terms were determined through a combination of clinical and research experience and support from subject librarians. The terms included were based on four key concepts: ‘measurement’, ‘medicines’, ‘older people’, and ‘hospital discharge’ (see Supplemental File 2 for detail). No limits were applied to searches and all study designs were considered. Eligible studies had participants aged 65 years or older, across the hospital-to-home transition (i.e., before, during or after hospital discharge), and a measure which contained at least one medication self-management component. The Previdoli et al. [15] definition was used throughout this review, providing a broader interpretation of medication self-management. Studies were excluded if all participants were aged less than 65 years, and if the measure had no medication self-management component. Conference abstracts or oral presentations were also excluded where full text was unavailable. Articles were downloaded to bibliographic software (EndNote®) [28] and a research software tool (Rayyan®) [29] and duplicates were removed. Screening was conducted by seven reviewers (HM, JT, JS, AMB, AS, AJ, EA) and discrepancies resolved through discussion. All included full-text articles were independently screened by two reviewers (See Fig. 2).

Fig. 2.

Flow diagram detailing review selection process

Data extraction

Data extraction was performed independently by three reviewers (HM, JT, EA) relating to study, patient, and intervention characteristics utilising a standardised data extraction table for each study. Key characteristics of measures were tabulated to allow comparison, as well as identification of patterns (See Table 1). Furthermore, the types of skills assessed in each measure were extracted in a separate table (See Table 2). Additional details were sourced as needed from primary studies and authors.

Table 1.

Measure characteristics

Measure	Purpose	Administration			Scoring Scale	No. of Items/ questions	Validity		Reliability			PPI	Theoretical underpinning
		When	By whom	Time (min)			Content	Construct	Inter-ratera	Test–retesta	Internal consistencyb
Medi-Cog (Anderson. K. 2010) [30]	Cognitive screens and pillbox organization assessment tools. Used to provide timely interventions to facilitate safer care transitions	Before discharge	Pharmacist	7 to 8	0–10	3	✔	✔	Not recorded	Not recorded	Not recorded	Not recorded	Not recorded
Anderson & Birge (Anderson et al. 2016) [31]	Assesses medicines management ability to understand if cognitive dysfunction is a risk factor for early readmission in older adults independently managing medication	Before discharge	Research staff	Not recorded	None	2	Not recorded	Not recorded	Not recorded	Not recorded	Not recorded	Not recorded	Not recorded
H3TQ (Arbaje et al. 2023) [32]	To assess the quality of Hospital-to-Home-Health transitions and proactively identify transition issues	After discharge	Home nurse	< 10	Yes/No	12	✔	✔	> 0.90	0.84	Not recorded	Older adult, caregiver & Home Health provider input at each stage (observations, interviews, focus groups)	HFE Systems Engineering Initiative for Patient Safety (SEIPS) 2.0 model as the conceptual framework guiding study design, data collection, and analysis
BMQ (Barnason et al. 2010) [33]	To measure medication adherence and medication use barriers, to evaluate impact of a hospital transition intervention for older adults promoting medication self-management	After discharge	Research nurse	Not recorded	0–11	13	✔	✔	Not recorded	Not recorded	Not recorded	Not recorded	Not recorded
CTM-15 (Boockvar et al. 2022) [34]	To assess patient understanding of their condition(s) and medication(s) after discharge, which was a secondary outcome measure of the RCT, testing effectiveness of a Care Transitions Intervention	After discharge	Medical team	8 to 12	0–100	15	✔	✔	Not recorded	Not recorded	0.93	Developed with substantial input from patients and family caregivers	Not recorded
CTM-15 (Coleman et al. 2007) [35]	To assess the quality of care transitions and identify deficiencies in hospital care transition process	After discharge	Research staff	8 to 12	0–100	15	✔	✔	Not recorded	Not recorded	0.93	Developed with substantial input from patients and family caregivers	Not recorded
Lawton IADL scale (Dumur et al. 2019) [36]	To assess older adults’ functional ability on admission to hospital to provide a baseline characteristic to help evaluate the effectiveness of the intervention (SMRF)	Before discharge	Medical team	10 to 15	0–8 (women) 0–5 (men)	8	✔	✔	0.85	0.96 (men) 0.93 (women)	Not recorded	Not recorded	Not recorded
DRUGS (Barnason et al. 2010) [33]	To assess medication use skills as part of a group of assessments, to evaluate impact of a hospital transition intervention for older adults with HF which promotes medication use self-management	After discharge	Research nurse	5 to 15	0–100	5	✔	✔	> 0.90	> 0.90	Not recorded	Not recorded	Not recorded
DRUGS (Gianotti et al. 2019) [37]	To assess medicine management ability of older patients for specific formulations (blister packs, tablets, child resistant drops, insulin pens, inhaler devices) to identify main clinical factors associated with potential inability to manage drugs	Before discharge	Not recorded	5 to 15	0–100	5	✔	✔	> 0.90	> 0.90	Not recorded	Not recorded	Not recorded
Home Functioning Questionnaire (Kneebone et al. 1996) [38]	To assess functional ability of older patients prior to discharge. Any discrepancies with healthcare professional observations are addressed in hospital before patient goes home	Before discharge	Occupational therapists	Not recorded	None	12	✔	✔	Not recorded	0.67	0.89	Not recorded	Not recorded
MedMaIDE (Mortelmans et al. 2021) [39]	To assess medication self-management capacity and identify medication management deficiencies to evaluate post-discharge medication self-management problems in older patients	After discharge	Research nurse	30	0–13	16	✔	✔	0.74	0.93	0.71	Developed by a panel of experts who have experience with and knowledge of the elderly population in community settings	Not recorded
PACT-M (Oikonomou et al. 2019 and 2020) [40, 41]	PACT-M 1 to capture the immediate post discharge period and PACT-M 2 to assess the experience of managing care at home	After discharge	Research staff	10 to 25	None	8 (× 2)	✔	✔	Not recorded	Not recorded	> 0.80	PPI panel in the measure conceptualisation. Incorporated feedback from PPI panel into measure development process	Not recorded
CTM-15 (Toles et al. 2023) [42]	To assess patient and caregiver preparedness for care at home, alongside other measures, to test the efficacy of the Connect-Home transitional care intervention	After discharge	Research staff	8 to 12	0–100	15	✔	✔	Not recorded	Not recorded	0.93	Developed with substantial input from patients and family caregivers via focus groups	Not recorded
Adapted Red Flag Instrument (Verweij et al. 2018) [43]	To assess problems with medication use post-discharge to help investigate if unplanned hospital readmission and mortality can be reduced by the Cardiac Care Bridge transitional care program	After discharge	Home nurse	Not recorded	Yes/No	16	✔	✔	Not recorded	Not recorded	Not recorded	Not recorded	Not recorded
Mortelmans Self-Developed Survey (Mortelmans et al. 2021) [39]	Assesses how older people manage medication as inpatient, preparation on discharge, post-discharge management of medication and side effects, and correct intake of medicines post-discharge	After discharge	Research nurse	Not recorded	Likert	17	Not recorded	Not recorded	Not recorded	Not recorded	Not recorded	Not recorded	Not recorded

^aReliability coefficient

^bCronbach's Alpha coefficient

Table 2.

Types of skills assessed within identified measures [32, 34–43, 45]

F2F Face-to-Face, HCP Healthcare professional

Quality assessment

Methodological quality assessment was carried out independently by HM and EA using the Quality Assessment for Diverse Studies (QuADS) tool [44] (See Supplemental File 3). The QuADS tool [44] was designed for use in health services research reviews when included studies are of multiple methods, hence most relevant for this systematic scoping review which included studies of varying designs. The QuADS criteria consist of 13-items, with a maximum score of 3 per item (i.e. maximum total score of 39). A higher score indicates better quality. One of the valuable aspects of this tool is that it considers a broad range of matters when assessing research quality, such as the involvement of stakeholders in research design. Patient and public involvement (PPI) ensures research is relevant and ultimately improves the quality of research. For this review, where the end-user of identified measures are patients, PPI has an essential role in ensuring the measure is more likely to be effectively utilised and implemented.

To ensure consistency across reviewers, 3 included studies were chosen and reviewers (EA and HM) independently assessed them using QuADS criteria. Scores of the 3 papers were discussed between HM and EA allowing a consensus to be reached on how they used the tool thereafter when assessing all included studies. Once both reviewers independently scored included studies, discrepancies were discussed and agreement reached. The methodological quality of included studies informed evaluation but was not part of the eligibility criteria for this review.

Data synthesis and analysis

A narrative approach was used to describe the measures. All 12 measures were mapped to the six medication self-management components in the Previdoli et al. [15] definition (See Fig. 3). The skills assessed in the 12 measures were mapped to the COM-B model [25] components, sub-components, and the TDF [26] domains (See Table 3). The TDF was first published in 2005 [46], with a revised version published in 2012 [26]. For this review, we utilised the 14-domain 2012 TDF version, which we will refer to as TDF(v2) [26].

Fig. 3.

Measures mapped to medication self-management components

Table 3.

Skills mapped to the behaviour change wheel (COM-B model and TDF domains)

HCP Healthcare Professional

Data synthesis was conducted by HM and one other reviewer independently. Both reviewers mapped each skill to the COM-B and TDF domains they believed determined the performance of that skill. Following discussion, consensus was reached between reviewers before finalising the medication self-management components and the relevant constructs of the COM-B system [25] and TDF(v2) [26] domains.

Results

The final search yielded 15,508 distinct records and 15,301 records were excluded after title and abstract screening, leaving 192 full text records for retrieval. Subsequently, 14 records were found to meet the eligibility criteria and included in the review (see Fig. 2).

Study characteristics

The 14 identified studies were conducted in seven countries, representing a variety of healthcare systems and funding models (see Supplemental File 4): USA [30, 31, 32–35, 42], Australia [38], France [36], Italy [37], Belgium [39], UK [40, 41] and the Netherlands [43]. Studies had varying numbers of participants, ranging from one [30] to 500 [43] and the study design also varied: RCT [34, 42, 43], prospective observational cohort [31, 32, 35, 37], cross-sectional [38, 39, 41], experimental study [33, 36], case report [30], and pilot study [40].

Quality assessment of included studies was performed utilising the QuADS tool [44] and demonstrated that with the exception of Anderson et al.’s [30] case report, studies were of good quality. The mean score of 13 included studies (excluding Anderson et al.’s [30]) was 27 out of 39, with scores ranging from 22 to 37. Anderson et al.’s [30] case report scored significantly lower (10), and this is relating to criteria assessing study design. For each of the 13 items in the QuADS tool [44], scores were totalled across the 14 studies. This highlighted that amongst all 14 papers, lowest performance was for theoretical underpinning and evidence of stakeholder involvement in research design (See Supplemental File 5 QuADS assessment of included papers).

There were 12 unique measures identified and their purpose included assessing the quality of care transitions [32, 35, 40–42] testing the effectiveness of an intervention [31, 33, 34, 36, 37, 39], or to enable safer discharge home by detecting and correcting medication self-management discrepancies [30, 38, 43] (i.e. low medication self-management capability). Of the 12 measures, 3 (DRUGS [33, 37], MedMaIDE [39] and Medi-Cog [30]) have been reported in previous reviews [17–21] and 9 were new.

Measure characteristics

(For more details on measure characteristics, see Tables 1, 2 and 3).

Administration of measures

Studies were administered prior to hospital discharge in the hospital setting or post discharge at the patient’s home. For measures administered prior to hospital discharge, timing ranged from the point of admission to 48–72 h before discharge, and for measures administered post-discharge timepoints ranged from 3–30 days post-discharge. Some studies did not record this information [30, 38].

The individual administering the measure also varied, and the number of items in a measure ranged from 2 to 17, with different scoring systems (see Table 1).

Use of measures

In this review, all measures were being utilised in clinical research, with the exceptions of Medi-Cog [30] (which is described in case reports taken from clinical practice) and the Lawton IADL scale [36] (which is widely used in clinical practice to assess older adults’ functional ability). Administration time for measures ranged from 5 to 30 minutes and was not recorded for six measures [30, 31, 33, 38, 39, 43].

Validity

Content validity was demonstrated across 11 measures. No information was available on the validity of the ‘Anderson & Birge’ [31] measure, which was part of an intervention at care transitions rather than a standalone measure. Content validity was mostly via expert review confirming face validity and some piloting with older people.

Construct validity was determined for all measures, except Anderson & Birge [31]. Most measures fit into one of two categories: a broader assessment of care transition quality; or a more precise assessment of older people’s ability to self-manage their medication across care transitions. Convergent validity was reported for two measures [45, 47].

Reliability

A degree of reliability was reported for more than half of the measures (seven). Comprehensive reliability data was only reported for one measure: MedMaIDE [39]. The studies including PACT-M [40, 41] and CTM-15 [34, 35, 42] measures only reported internal consistency.

Patient and Public Involvement

Of the 12 unique measures identified, only four had varying degrees of PPI reported in their development: CTM-15 [34, 35, 42], H3TQ [48], MedMaIDE [39] and PACT-M [40, 41]. Arbaje et al. 2023 [48] describe PPI in the most detail with different stakeholders involved at different stages and using a range of methods to develop the H3TQ [48].

Skills assessed using the measures

Medication self-management capability was determined through physical and cognitive skills (see Table 2). There were 23 skills identified across the 12 measures; all measures assessed cognitive skills, with three measures assessing both cognitive and physical skills (DRUGS [33, 37], BMQ [33] and MedMaIDE [39]). Most frequently, patient knowledge of dose and administration time of their medication was assessed through patient recall (verbally or in writing).

Measures were categorised according to method of administration (see Table 2): (1) Face-to-face capability assessment via observation [30, 33, 37]); (2) Face-to face interview questionnaire [31, 32, 33, 36, 38, 39]; (3) Telephone interview questionnaire [34, 40, 41, 42]); (4) Face-to-face capability assessment via observation combined with interview questionnaire [39, 43]). Interestingly, the skills that a measure assessed were linked to the method of assessment. Most measures using interview questionnaire (face-to-face or telephone) did not assess physical skills. However, the BMQ [33], assesses difficulties patients face with medication supply and use of medication bottles.

In addition, most measures administered via interview questionnaire were crude, with items that were broad, for example the question “Who manages your medicines?” [31], or statement “Able to remember to take my tablets at home” [38]. Due to the generality of such statements, these measures assessed a lower number of skills (see Table 2).

In contrast, measures that involved direct observation of skills were more specific in their questioning, for example “Does the patient experience problems with reading and/or understanding the instructions for use? (e.g., due to functional illiteracy or vision problems)” [43]. The number of skills assessed ranged from one to eleven in the different measures (see Table 2).

Behaviours assessed through identified measures

The 23 skills assessed across the measures were mapped against the conditions for behaviours in the COM-B [25] and TDF(v2) (See Table 3). Notably, every skill is mapped to the COM-B capability domain. Most skills assessed were cognitive so, in the behaviours, the psychological sub-domain maps to knowledge-based skills assessment, but the physical sub-domain does not. The COM-B opportunity domain was involved across eight of the skills assessed and the motivation domain in ten. Both opportunity sub-domains (physical and social) were addressed, as well as the corresponding TDF(v2) domains (environmental context and resources, and social influences). For the motivation domain, some skills covered the reflective sub-domain, but none mapped to the automatic motivation subdomain. Additionally, all three COM-B domains were addressed across varying cognitive skills, but no physical skills mapped to the motivation domain. Lastly, three TDF(v2) domains were not covered by any assessed skills: professional/social role and identity; reinforcement; and emotion.

Medication self-management components

(See Supplemental File 6 for more detailed description of the Previdoli et al. [15] definition).

The measures had different objectives, influencing the extent medication self-management was assessed by each measure. Figure 3 shows that no measure mapped to all six medication self-management components [15], although four measures addressed three components. Component 3: Ensuring medication is taken as instructed, is addressed by every measure but the other five components are implemented less frequently. Importantly, the extent to which component 3 is addressed throughout measures differs. For example, PACT-M [40, 41] addresses this through two statements, whereas MedMaIDE [39] comprises two sections that seek detailed understanding of both cognitive and physical aspects of how a person manages their medication. In addition, the depth by which the medication self-management component is addressed depends on the detail of individual items and the range of skills that they assess. Table 2 shows differences in the number of skills assessed by BMQ [33] (nine), PACT-M [40, 41] (three), MedMaIDE [39] (eleven) and the Adapted Red Flag Instrument [43] (eight).

Discussion

This systematic scoping review identified 12 distinctive measures that assess medication self-management capability of older people across the hospital-to-home transition. Medication self-management capability was determined through assessment of physical and cognitive skills, with 23 individual skills identified within the measures. The number and type of skills assessed varied greatly, affecting the behaviours addressed and depth of assessment. The COM-B model [25] capability domain was consistently covered in all 12 measures, as was Previdoli et al.’s [15] medication self-management component 3: Ensuring medication is taken as instructed, indicating measures were particularly adherence-centred. The measure assessing the highest number of skills and medication self-management components was MedMaIDE [39]. For most measures reviewed, there was evidence of validity, but this was limited to content and construct validity. Additionally, comprehensive reliability data was only available for one measure: MedMaIDE [39].

Badawoud et al. [17] argue that most measures aim to identify cognitive and physical skills but vary in the specific skills they evaluate. The authors identified ideal characteristics of a medication self-management capacity assessment for adults, concluding that the DRUGS [33, 37] and MedMaIDE [39] measures met these criteria (with MedMaIDE [39] the most comprehensive). While Baby et al. [23] also noted MedMaIDE [39] has a good degree of assessment, it’s focus on cognitive and physical skills led to their recommendation that a combination of tools is necessary to measure the different components of medication self-management capability [23]. Unsurprisingly, this review confirms this as we found MedMaIDE [39] assessed the most skills (cognitive and physical), included half of the Previdoli et al. [15] medication self-management components, and was also one of the few measures where PPI was reported. MedMaIDE [39] appears to be the most comprehensive existing tool, albeit with some components missing for a complete measurement of medication self-management capability.

It became evident that there is an important link between measure purpose and depth of information extracted. Badawoud et al. [17] explain the importance of considering the purpose of a measure for use in research or clinical practice. Some measures (PACT-M [40, 41], H3TQ [32], Home Functioning Questionnaire [38], CTM-15 [34, 35, 42], Mortelmans’ Self-Developed Survey [39]) assess a patient’s confidence and preparedness to go home, manage medicines on discharge, and seek help post-discharge. However, having the confidence to do a skill may not mean you can correctly perform that skill. Thus, the measure captures a perceived ability, which is less useful for understanding a patient’s capability to self-manage their medication.

This review aimed to identify existing measures that assess medication self-management capability in older people across the hospital-to-home transition. Taking a simplistic view of the meaning of the word ‘measures’, it refers to “a way of judging something” [49]. In this sense, the measures make a judgement, to varying degrees, on whether an older person can manage their medication across care transitions. However, some measures also perform as tools or instruments: “something that helps you to do a particular activity” [50]. Differentiating between a measure and a tool is important; a measure on its own does not automatically lead to another action and a desired outcome.

Three measures identified in this review were also being utilised as tools: Medi-Cog [30], the Home Functioning Questionnaire [38] and the Adapted Red Flag Instrument [43]. This was explicitly linked to the purpose of the studies, which was to enable a safer discharge home by detecting and correcting medication self-management discrepancies (i.e. low medication self-management capability) before discharge. This is what makes them tools: the information obtained through the measure is used to support safer use of medication on discharge to differing extents.

Our findings emphasise the importance of recognising how medication self-management is not one behaviour. There are several components that need to be fulfilled to enable effective medication self-management: a complex set of behaviours. Thus far, most interventions are task rather than behaviour driven and many lack any foundation in behaviour change theory [24]. A significant finding when mapping skills assessed to the COM-B [25] and TDF(v2) [26] was that no skills mapped to the COM-B automatic motivation subdomain or TDF(v2) reinforcement and emotion domains (See Table 3). These domains relate to the habitual nature of behaviours and how people feel enacting those behaviours [25]. None of the current measures explore this, which raises the question: is it difficult to assess, or has this been overlooked due to measures not being underpinned by behaviour change theory? One way of addressing this could be asking whether a patient has a routine for taking their medication, or how they feel about managing their medication.

Previdoli et al. [15] describe a more holistic approach to medication self-management components than previously observed in the literature. However, they focused on patients managing medication at home, and not across care transitions. Thus, there were gaps when trying to map the components against the measures and skills identified in this review. This definition does not account for the dynamic nature of care transitions and the change in abilities that older people can experience, for example, physical aspects of medicines management are not explicitly tackled. At the hospital-to-home transition, physical handling of medication is key, so must also be assessed.

Developing an effective measure may not translate to its use outside the research setting. The implementation gap described in the literature is, therefore, a great concern when considering the impact of healthcare research [51]. Implementation is often inadequate and complicated by issues around sustainability of healthcare systems, making introduction of an assessment tool challenging. Our findings suggest the current underutilisation in clinical practice of medication self-management assessment could be the effect of inconsistent terminology. Researchers must be clear on whether they are developing a measure or a tool, and understand what, if anything, will be done with the information obtained.

Furthermore, measures must be quick and simple for staff to administer in busy clinical areas [52]. This factor is why such measures may not be widely adopted in clinical practice [17, 23], one suggestion being that a measure should be administered in less than 5 min [17]. Of the 12 identified measures in this review, only two demonstrated use in clinical practice: Medi-Cog [30] and the Lawton IADL scale [36]. The administration time of these measures was 7 to 8 minutes and 10 to 15 minutes, respectively. While a measure needs to be as concise as possible, it may not be possible to address the complexity of medication self-management capability in 5 minutes, whereas 15 minutes could be a good compromise between retrieving adequate information and having a measure which is implementable.

A practical outcome from our review is a set of recommendations to consider when developing a measure to assess medication self-management capability in clinical practice:

1. Use a behavioural theoretical underpinning so actions which are required to enable an individual to enact medication self-management are addressed.

2. Be clear on purpose and objective as this impacts quality, method utilised, and ultimately what information is extracted.

3. Incorporate clear questions or statements which achieve depth to attain precise assessment of skills, rather than perceived abilities.

4. Ensure the measure is also a tool, which is short and practical to aid implementation in clinical practice.

5. Utilise a comprehensive definition of medication self-management, encapsulating the complexity of this phenomenon to avoid generating an adherence-focused measure.

6. Develop the measure with patient and public involvement through co-design.

If such a measure was developed, one final consideration we suggest is ensuring steps are taken to evidence validation and reliability. It is clear from this review that medication self-management capability measures which demonstrate thorough validity and reliability are lacking. If the intention of a measure’s development is to embed utilisation into clinical practice, accuracy and consistency are fundamental characteristics, so prioritising this in future measures is key.

Strengths and limitations

The main strength of this review was the exhaustive and systematic search strategy; we searched five databases, with no date limitation, and all literature was included regardless of language, study design, validity, as well as relevant grey literature. Findings from this review have added to the current understanding of measures which assess medication self-management of older people, specifically across care transitions, providing guidance on how we can address this gap through co-design of a comprehensive assessment tool. Included papers underwent quality assessment which catered for the range of study designs across studies. Largely, the included papers were of high-quality, meeting most of the QuADS tool [44] criteria, for example: clear research aims, strong study design and a well-defined population. The areas studies performed less well in were theoretical underpinning and evidence of including stakeholders in study design. Overall, the good methodological quality of included studies supports confidence in the findings. Despite best efforts to run a thorough search of literature across five databases, it is still possible that important studies could have been missed.

Conclusions

This systematic scoping review has identified and evaluated 12 measures that assess medication self-management capability in older people across the hospital-to-home transition. Current measures predominantly focus on cognitive and physical skills, with a significant emphasis on medication adherence. The complex and multifaceted nature of medication self-management is highlighted in this review. Medication self-management goes beyond adherence, and our findings further emphasise the importance of comprehensive definitions when considering this phenomenon.

The relationship between purpose of a measure and its ability to provide detailed information was evident, stressing the importance of aligning a measure’s objectives with its intended use. Despite identification of several useful measures, there remains a notable gap in implementation in clinical practice, possibly due to administration time. We suggest aiming for an administration time of around 15 minutes could balance thorough assessment with practicality, facilitating wider adoption in busy healthcare settings. Further research in this area is required to determine if there are other factors at play. Lastly, to bridge the gap between research and implementation, a set of recommendations have been provided for developing future measures.

Abbreviations

COM-B

Capability, Opportunity, Motivation – Behaviour system

MESH

Medical Subject Headings

TDF

Theoretical Domain Framework

QuADS

Quality Assessment for Diverse Studies

PRISMA-ScR

Preferred Reporting Items for Systematic Reviews and Meta-analyses extension for Scoping Reviews checklist

Authors’ contributions

HM conceptualised and designed the study with input from JT, HS, BF and PHG. Screening was conducted by HM, JT, EA, AMB, AJ, JS and AS. Data extraction and quality assessment was conducted by HM, JT and EA. The manuscript was drafted by HM and all authors provided intellectual content to the manuscript, critical feedback and approved the final version.

Funding

Not applicable.

Data availability

The datasets used and analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

None required as systematic scoping review.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.