Medication Management Frameworks in the Context of Self-Management: A Scoping Review

Introduction

The use of medication, prescribed and unprescribed, is common for the management of both acute and chronic health conditions, as well as for the maintenance of overall health.^1,2 There is a high prevalence of prescription medication use worldwide, with some countries reporting an increase in medication use over the last 20 years.^3–5 Among individuals taking medication, there is a large portion who are taking multiple medications (ie, polypharmacy). Polypharmacy is most commonly defined as the concomitant use of five or more medications, which often includes prescription medications, over-the-counter medications and natural health products.⁶ An alternative definition, also known as problematic polypharmacy⁷ or medication overload⁸ describes situations in which more medications are used than are indicated or the harms exceed the benefits.

Based on the high prevalence of medication use and polypharmacy, understanding and improving medication management is essential. Medication management can be defined as, “patient-centred care to optimize safe, effective and appropriate drug therapy”.⁹ Medication management involves a number of services aimed to improve clinical outcomes, such as: completing medication reviews and health assessments, monitoring treatment plans and efficacy and safety of therapy, providing education and promoting self-management.¹⁰ Self-management is a fundamental component of optimal medication management, and is defined as

an individual’s ability to manage the symptoms, treatment, physical and psychosocial consequences and lifestyle changes inherent in living with a chronic condition (p. 178).¹¹

Self-management encompasses three key domains: medical, emotional and role management.¹² Medical management involves tasks and responsibilities related to managing or controlling one’s condition (eg taking medication, eating healthy); emotional management involves learning to manage feelings and emotions related to one’s condition (eg fear, depression, anger); and role management includes adapting or creating new responsibilities as a result of one’s condition (eg scheduling around medication-taking, limiting or finding new hobbies).¹²

Over the past thirty years, the promotion of self-management of chronic conditions through programs and interventions has become increasingly common.^13–17 Self-management programs and interventions commonly focus on the following areas to improve individuals’ clinical and psychosocial outcomes: education, goal setting, decision-making, seeking support and resources, self-advocating, problem solving and relationship building.^12,18 Several systematic and meta reviews have noted positive outcomes of self-management programs and interventions on health outcomes for persons with diabetes,^14,16,17 heart failure,^19,20 asthma^21,22 and depression.²³ These systematic and meta reviews included studies that ranged in quality, according to the authors’ quality assessments, from low quality to very high quality.

Surprisingly, there has been limited integration of medication management and self-management, despite the effectiveness of self-management interventions and it being a fundamental component of medication management. To date, the majority of medication management frameworks are focused on patient adherence to a specific medication regimen, with a limited focus on broader concepts of self-management. Therefore, given the high use of medications and narrow focus of medication management, this scoping review sought to identify what is reported in the literature on medication management frameworks or models within the context of self-management.

Materials and Methods

A scoping review was conducted to examine the extent, range and nature of research on medication self-management frameworks, summarize current research and identify gaps in existing literature on this topic.²⁴ This scoping review followed the methodological approach described by Levac et al²⁵ and aligns with the methodological guidance of the Joanna Briggs Institute on conducting scoping reviews.²⁶ The reporting guidelines of the PRISMA-ScR checklist were also followed (see Supplementary Table 1).²⁷ A protocol was developed and registered on Open Science Framework (https://osf.io/6ckg7).

Stage 1: Identifying the Research Question

The research question guiding this scoping review was: what is reported in the literature on patient- or person-level medication management frameworks or models within the context of self-management? The three main objectives were to: (a) examine the extent to which the principles of self-management were integrated into the frameworks, (b) synthesize the key aspects of self-management principles that were captured in medication management frameworks and (c) identify any gaps in the literature to propose future studies.

Stage 2: Identifying Relevant Studies

Relevant studies were identified by searching four electronic databases (Medline (Ovid Interface), Embase (Ovid Interface), Cumulative Index to Nursing and Allied Health Literature (EBSCO Interface) and Cochrane Library from January 1, 2000, to January 8, 2020. The search strategy was developed in consultation with a librarian at the University of Toronto and adapted for each database. Keywords and medical subject headings, such as medication management, framework, model and self-management, were combined with Boolean and proximity operators (see Supplementary Table 2 for Medline search strategy). The search of electronic databases was conducted on January 8, 2020. In addition to database searches, grey literature was searched on TSpace (University of Toronto research repository) and relevant healthcare and government organization websites (eg World Health Organization, National Institute of Health). Reference lists of included articles were reviewed to ensure the inclusion of all relevant literature.

Stage 3: Study Selection

For inclusion in the scoping review, articles were required to meet the following criteria: (a) described a model or framework originally developed or adapted by authors for medication management; (b) included components of self-management; (c) peer-reviewed literature or grey literature; and (d) published between January 1, 2000, and January 8, 2020. The search was limited to the year 2000 because the majority of research on self-management occurred after this date.¹² Articles were excluded if any of the following criteria were met: (a) described models of care (interventions, programs, etc.) that were not guided by a concrete model/framework related to medication self-management; (b) models/frameworks that had not been adapted to include components of medication management or self-management; (c) the model/framework was lacking details on what changes/adaptations were made to the model/framework for medication self-management; (d) models/frameworks that only included self-efficacy for medication-taking behaviour as the self-management component; (e) opinion pieces; and (f) conference abstracts or articles in which the full-text was not accessible.

Study selection was conducted using EndNote X8 for de-duplication (reference manager software) and Covidence for screening (online review management software). Following Bramer’s method, EndNote X8 was used to de-duplicate the exported articles from the database searches.²⁸ The core study team (SJTG, LC, KAJ, TR) used a spreadsheet in Microsoft Excel 2016 to facilitate an interrater screen of a subset of titles and abstracts. All disagreements were discussed until consensus was reached and revisions were made to the eligibility criteria, as needed. Once good agreement (defined as >0.80)²⁹ was achieved, the remainder of the titles and abstracts were imported into Covidence and divided among three reviewers (LC, KAJ, TR) to be screened by a single reviewer. After the completion of title and abstract screen, the core study team (SJTG, LC, TR, SRC) completed an interrater screen of 100 full-text articles (in subsets of 25) to test their agreement. Following each subset of 25 full-text articles, percent agreement was calculated and clarifications were made to the eligibility criteria as needed. After achieving good agreement (>0.80), all full-text articles were double-screened (two separate reviewers screened each article). Disagreements were reviewed by a third screener (senior author) to make the final decision on inclusion.

Stage 4: Charting the Data

A data extraction sheet was created in Microsoft Excel 2016, which was piloted by the extractors (LC, TR, SRC) to ensure consistency in the data extracted from the articles. Questions and disagreements were discussed until clarification or consensus was achieved. A spot check (review of extracted data) was completed (LC, TR) on 15% of the articles to ensure the accuracy of extracted information. The following data were extracted from each article: general information, study characteristics, population characteristics, framework characteristics, intervention characteristics, study outcomes and key findings.

Stage 5: Collating, Summarizing and Reporting the Results

Extracted data were compared across variables. Descriptive numerical analysis and thematic analysis were conducted and are presented below. The data were also critically analyzed to identify current gaps in the literature and develop suggestions for future research, practice and policy.

Results

Search Results

The database searches identified 8683 articles and no additional articles were found from the grey literature searches (see Figure 1). Following the removal of duplicates, 5668 articles remained for title and abstract screening. The full-text of 426 articles were reviewed. Following full-text review, 39 articles met all of the criteria and were included in the scoping review (see Table 1 for study characteristics of included articles).

Table 1 Characteristics of Included Articles (n=39)

Figure 1 PRISMA flow diagram of included articles.

Study Characteristics

There was a fairly even distribution between quantitative (n=15)^30–44 and qualitative (n=12)^45–56 study designs, with few mixed method studies (n=3)^57–59 and several other article types (eg reviews; n=9).^60–68 From 2013 to 2019, there was an increase in the number of articles published yearly (total=24),^{30–32,36–42,47–51,54,55,57–61,64,67} when compared to the 11 years prior (2002 to 2012; total=15).^{33–35,43–46,52,53,56,62,63,65,66,68} The studies identified were conducted across 12 different countries: the United States (n=24),^{30,31,33,34,36,39,40,43–46,48,50,51,53,56,59–63,65,67,68} China (n=2),^38,42 the Netherlands (n=2),^35,55 Singapore (n=2),^49,64 South Africa (n=2),^37,47 South Africa and Kenya (n=1),⁵⁷ Australia (n=1),⁶⁶ Canada (n=1),⁴¹ Israel (n=1),⁵² Malaysia (n=1),⁵⁸ Philippines (n=1)³² and the United Kingdom (n=1).⁵⁴ Of the 39 articles, only one study involved more than one country.⁵⁷

The majority of frameworks or models were developed, adapted or applied to adults and older adults, but two were for a younger population (adolescents and young adults).^36,39 The most common target populations were adults with HIV (n=10)^{30,36,37,43–45,53,57,63} and cardiovascular disease (n=5).^{46,49–51,65} Of the included articles that had a study population (n=30), the majority reported the age and sex or gender of participants.^{30–38,41–46,51,56} There was a fairly even distribution of males and females, with mean ages ranging from 21 to 81 years. Other demographic characteristics reported in about half of the studies were ethnicity/race^{30,31,33,34,36,39–46,48–51,56,59,67} and education level.^{30–38,41–46,51,55,56} The most common ethnicities/races studied were Caucasian and African American among individuals who had at least some college education. Demographic characteristics that were not consistently reported included: income level, marital status, household composition, employment status and comorbidities. The overall sample sizes ranged from 6 to 2213 (mean=265; median=136).

Framework Characteristics

Approximately half of the included articles developed and presented a new framework (n=20), while the other half adapted previously existing frameworks, or applied them to their findings (n=19; see Table 2 for framework characteristics). New frameworks and models were developed by research teams through the use of statistical modeling (structural equation models, confirmatory factor models, measurement models, path analysis; n=8), data collected from interviews and focus groups (n=6) and literature syntheses (n=4). One new framework was developed from literature syntheses and focus groups and one framework lacked sufficient details on how it was developed. Adapted frameworks were most commonly based on social cognitive theory, behaviour change models and medication adherence models. Similar to newly developed frameworks, adapted frameworks were often based on quantitative surveys (n=7), literature syntheses (n=4), qualitative data from interviews and focus groups (n=3) or a combination (n=3). No frameworks were developed or adapted through co-design sessions with participants.

Table 2 Framework Characteristics

The majority of self-management domains (eg adherence; self-efficacy; depression; and seeking advice, support and relationships) included in the frameworks were categorized as medical management (eg medication/treatment adherence, medication/adherence self-efficacy, medication-taking behaviour), followed by emotional management (eg depression, emotional impact) and role management (eg problem solving, day-to-day management, spirituality). Across all frameworks, developed, adapted and applied, the most common domains (not limited to self-management) were related to adherence, self-efficacy, depression, social support and regimen complexity/factors (eg number and size of pills, frequency of refills). There were no major differences between the framework domains from quantitative and qualitative studies.

Discussion

This scoping review analyzed the range and nature of research on medication self-management frameworks. Within the three self-management domains (ie, medical, emotional and role management), the vast majority focused on medical management (eg adherence, medication-taking behaviour, medication self-efficacy, relationships with healthcare providers), with a limited focus on emotional and role management. Further, within medical management, the majority of studies focused specifically on medication adherence and lacked a comprehensive, holistic conceptualization of self-management. Given the prevalence of individuals taking medications and the potential for problematic polypharmacy,^7,69 it is important that research and clinical practice conceptualizes medication management beyond adherence to promote person-centred care.

While limited, there were a few frameworks that included a more holistic approach to medication self-management.^50,56,62 For example, Blalock used behavioural science theories (eg social cognitive theory, theory of social support, Health Belief Model) to describe how to improve medication use among individuals requiring pharmacotherapy.⁶² An ecological model was used to present factors impacting medication use at five levels – characteristics of the patient; services from providers; support from family, friends and small groups; characteristics of the community and health care system; and social policy and government regulations. Blalock then proposed a framework specific to medication self-management that built on the ecological model and Fisher’s self-management model⁷⁰ to incorporate specific intervention goals at each level. The author highlighted the value of using this as a guiding framework for medication self-management, but also identified the need to further develop it as knowledge evolves.

A more holistic framework for medication self-management was also presented by Mickelson et al.⁵⁰ A qualitative study was conducted among older adults with heart failure to describe the medication management work process (effort and time to produce or accomplish something) and adapt Unertl et al’s Workflow Elements Model.⁷¹ The adapted model contained five main process, each containing subprocesses: sensemaking (information gathering, adapting mental models, story building), planning (generating action plans, adapting plans, anticipatory thinking), monitoring (problem detection, tracking), decision making (applying rules, pattern matching, mental stimulation, making trade-offs) and coordinating (reconciling information, managing interdependencies, negotiating). Furthermore, these frameworks highlight the intrinsic complexities of medication management that are critical to consider when developing programs, interventions and initiatives targeting medication self-management.

Importantly, medication-taking has a substantial impact on an individual’s life.⁷² For example, a systematic review was conducted by Mohammed et al to explore medication burden among patients with experience taking medications.⁷² This review identified patients’ experiences with medication-related burden (ie, burden caused by medication routines, medication characteristics, adverse events, healthcare system, social life), medication-related beliefs (individual attitudes, coping skills, outside influences) and medication-taking practices (following instructions, accepting medication use, modifying care plans). Further, Mohammed et al’s review highlighted the impact of medications and medication burden on the day-to-day lives of individuals taking medications, as it interfered with their daily activities and influenced their beliefs, attitudes and overall quality of life. The impact of medication management on an individual’s life extends beyond medical management to include emotional and role management. Therefore, it is critical for medication self-management frameworks to be all-encompassing of the self-management domains.

Frameworks that took into account the participants’ sociodemographic and clinical characteristics are lacking based on our scoping review. Medication self-management may be impacted by micro, meso and macro level factors. Micro, or individual, level factors can impact medication use and the potential for polypharmacy, as age, gender, cognitive and physical ability and multimorbidity have been noted in the literature to affect medication-taking behaviour.^73–78 For example, a systematic review was conducted by Smaje et al (2018) to identify factors associated with medication adherence in older patients and found that older age, multimorbidity, cognitive impairment and being male were negatively associated with adherence.⁷⁴ Additionally, characteristics of a medication regimen, including complexity, dose, frequency and side effects can also impact medication adherence.^79–83 Increased regimen complexity results in reduced medication adherence, as identified in a systematic review by Pantuzza et al (2017) that examined evidence on the association between medication regimen complexity and adherence.⁷⁹ Similarly, a qualitative study conducted by O’Donovan et al (2019) in the United Kingdom explored the impact and management of medication side effects among individuals 18 or older who had experienced side effects.⁸⁴ It was identified that participants used non-adherent behaviours as a method of coping with side effects. Further, many of these individual level factors have also been identified as having an impact on general self-management,^85–87 which adds a layer of complexity to medication self-management.

Macro or system level factors can also have an impact on medication self-management. This review identified that, to date, the majority of frameworks for medication self-management were from the United States (n=24), with an overall lack of global representation. Medication self-management may look different in different regions or countries based on healthcare system and funding structures, insurance coverage and support from and access to healthcare providers.^88,89 In an analysis of health data from 1980 to 2016, Sarnak et al compared drug spending and trends in the United States, Australia, Canada, France, Germany, the Netherlands, Norway, Sweden, Switzerland and the United Kingdom.⁸⁸ The United States had considerably higher prescription drug spending and drug prices than the other high-income countries. Americans were also more likely to experience high out-of-pocket costs for medications. High out-of-pocket medication costs can impact medication self-management, as an inability to afford medications may lead to cost-related nonadherence.⁹⁰ It is important to acknowledge contextual factors in the development and application of medication self-management frameworks. In addition, more work is needed to explore what sociodemographic and clinical characteristics, as well as what system level factors impact medication self-management in order to understand if, and how, frameworks can be applied to different populations in different contexts.

Future Work

This scoping review highlighted several gaps in the literature and areas for future work. First, it is important for future research to develop a framework for medication self-management that directly incorporates patients’ and providers’ voices (ie, through co-design) and can be applied to a broader population of individuals taking medications. Integrating the thoughts, experiences, beliefs and concerns of persons with lived experience and providers is important, as patients are the ones self-managing and impacted by these processes on a day-to-day basis and providers are supporting patients with their health conditions and medications. More specifically, co-design would allow for the collaborative development of a framework that meets the needs of patients and providers. Second, a framework that encompasses all components of medication and self-management should be developed and validated. Current frameworks mostly focus on adherence, rather than incorporating the impact of experiences, beliefs and preferences on self-management behaviour. Frameworks guide research and consequently clinical practice recommendations.^91,92 Advancing the collective knowledge on medication self-management can help patients and providers navigate self-management and related self-management support.

Limitations

There are a few limitations of this scoping review that should be noted. First, it is possible that relevant articles were missed based on the databases and grey literature searched. Second, our search was conducted in English, so it is possible that articles published in other languages were not identified. Third, some may note that a critical appraisal of included articles was not undertaken. However, critical appraisal is a less common and optional component for scoping reviews.²⁷

Conclusion

Medication self-management is complex and has the potential to impact multiple aspects of an individual’s life, including mental and physical well-being, as well as day-to-day activities. The majority of frameworks included in this review focus on medical management, with few incorporating components of emotional and role management. It is important to acknowledge the impact medication self-management can have on all aspects of one’s life and focus future work on developing and validating holistic frameworks for medication self-management that can be applied to a broad population.