Defining, identifying and addressing problematic polypharmacy within multimorbidity in primary care: a scoping review

Numerous polypharmacy definitions

There is no consensus on a definition for polypharmacy, with significant variations in approaches to targeting problematic polypharmacy.2 3 17 Over 100 definitions of polypharmacy have been used, reflecting the discordance of approaches.18 19 Two main approaches to defining polypharmacy can be grouped into quantitative (using a form of medication count) and qualitative definitions (using descriptive notions of prescribing quality), with some studies using a combination of these definitions. Table 3 gives illustrative examples of these definitions.

Quantitative definitions of polypharmacy were more frequent, with over 90% of publications using some form of medication count.2 18–21 For example, the WHO defines polypharmacy as four or more medicines, academic studies most commonly use 5 or more.1 2 Other quantitative definitions included categorisations rather than cut offs of medication count. These were frequently labelled levels (eg, mild, moderate and severe) or attributes (eg, excessive, extreme), yet counts within these categories were also inconsistent.12 18 19 22–24 Generally, quantitative definitions were easier to operationalise and more reproducible, with a focus on medication count, regardless of whether polypharmacy is problematic. In contrast, qualitative definitions largely required clinical judgement to evaluate prescribing quality, carrying a focus on when polypharmacy becomes problematic. This frequently highlighted the overuse or overprescribing of medications. But definitions also covered aspects of misprescribing, often through applying a list of defined prescribing criteria, and also underprescribing, though only a few studies emphasised this aspect. The terms ‘appropriate’, ‘inappropriate’ and ‘problematic’ polypharmacy were also commonly used to describe when multiple medications were justified compared with when the clinical indication was unclear.3 18 19 25 These definitions have now been expanded to cover further dimensions of polypharmacy, such as the increasing recognition of the importance of patient and carer input in determining the appropriateness of medications.26 27

The challenges of defining when polypharmacy is ‘problematic’

The understanding of polypharmacy has progressed over time, with an increasing shift to more clinically applicable definitions. This reflects the increasing complexity of decision-making for combinations of medicines tailored to individual needs. There is also recognition that it is not possible to account for clinical appropriateness through simple medication counts.18 19 Commonly people with multiple health needs may well be appropriately prescribed more than 10 medications for therapeutic and symptomatic benefit, which would be termed extreme polypharmacy in some studies and guidelines.28 29 Yet there is some validity to numeric approaches as increasing medications are strongly associated with drug-related problems, and very high counts of medication are usually questionable.30 There is also a need to improve the consistency of reporting medication exposure characteristics.18 19 31–33 Various definitions have been used to define temporality and ‘long-term’ use, with some publications including ‘acute’ and ‘as required’ medications as opposed to chronic medications, with varied definitions of time periods (ranging from 1 to 240 days).18 19 Terms such as problematic polypharmacy and inappropriate polypharmacy have been increasingly favoured, as they consider appropriateness and clinical decision-making.34 Yet qualitative research suggests that these labels were still insufficient to reflect the complexity of medicines management, with practitioners juggling terms such as ‘potentially inappropriate’ and ‘specifically appropriate’ and others considering them ‘judgemental’ and even ‘accusatory’.35

Targeting potentially high-risk populations

Various strategies target higher risk populations to try and identify problematic polypharmacy. One common approach uses simple cut offs of age (commonly ≥65 years) combined with cut offs of medications (frequently ≥5) and this was the main inclusion criterion for the majority of trials.13 Another approach adopted by multiple national recommendations advocate case finding through high-risk groups.36–38 For instance, both NICE guidelines and the Australian Commission on Safety and Quality in Healthcare recommend greater attention for older people with frailty, and complex multimorbidity and co-existing mental and physical health problems.2 29 36 Accordingly, several national indicators, initiatives and studies also use combinations of these approaches.36–43 Other approaches include risk scores to identify patients at high risk of particular outcomes (eg, hospitalisations or adverse drug reactions) but these require further development.44 45 Overall, strategies to identify potentially high-risk populations currently demonstrate variable validity in polypharmacy and are seldom comprehensive or holistic, as they are specific to the needs of particular groups.2 36 46

Targeting potentially inappropriate medicines

Evaluating the appropriateness of individual medications is a common approach both as a case-finding approach and as a surrogate measure of prescribing quality across polypharmacy. Various tools have been developed to identify potentially inappropriate medicines and these can be split into explicit and implicit tools, with some tools combining both (examples in table 4).47–49 The majority have been developed using expert opinion and consensus methodology, and originally were designed for evaluating individual medications, rather than polypharmacy as a whole.47 50 Explicit tools contain specific criteria or scenarios leading to potential adverse drug events and carry advantages of reproducibility and ease of automation.51–55 Implicit tools require judgement, which means they can be subjective and demand more time and clinical expertise. Nevertheless, explicit tools are limited to specific drugs and diseases, but implicit tools can be applied to any medication. This perhaps allows implicit tools greater applicability in polypharmacy, as explicit tools will miss out any medicines outside criteria.56

Several systematic reviews have revealed a high level of variability of included criteria within explicit tools.47–50 54 57 A review of 36 explicit tools reported criteria spanning 907 medications and medication classes, but only 44 medications and 4 classes were reported by the majority.48 This was despite over 85% of these tools being developed based on either the Beers or the Screening Tool of Older Person’s Prescriptions/Screening Tool to Alert doctors to Right Treatment (STOPP/START) criteria.48 Due to this, many studies combine several explicit criteria to complement the list of medications included.47 48 50 58–62 Only about a third of tools suggested alternative treatments to potentially inappropriate medicines, yet nearly 70% of suggested alternatives were deemed inappropriate by other tools.47 Implicit tools are also diverse in nature, with reviews identifying over 16 different tools incorporating implicit criteria.54 63 These ranged from risk scores to lists of questions specifying appropriate use or criteria to evaluate the administrative burden to patients.54 63–65 Several tools combine implicit and explicit indicators, including documents used for national guidance (eg, Australian Prescribing Indicators Tool).63 66 67

Key limitations in identifying problematic polypharmacy in practice

Current strategies to identify problematic polypharmacy demonstrate inadequate performance. At present, risk stratification tools remain too broad, and seldom consider the clinical appropriateness of individual medications.34 68 Though comprehensive explicit criteria are helpful in identifying potentially inappropriate medications, translation into everyday care remains elusive due to challenges in clinical application, and the omission of medications not included in criteria.48 69 For instance, previous studies have found that less than 25% of adverse drug reactions are caused by drugs listed by Beers criteria.70 71 Additionally, apart from STOPP/START, most widely used tools were not designed to also cover underprescribing (table 4), with some studies also choosing to omit many of the underprescribing criteria in its application.47–50 54 Furthermore, there have been questions as to the utility of long lists of medications as studies have shown a high prevalence of potentially inappropriate medications (over 30% of patients) but low variability within many criteria, potentially leaving little room for improvement.72 Studies also mention usability issues with such long lists, even with computerised integration, and the difficulties of making treatment decisions without prioritisation of criteria, particularly as their predictive validity is unknown.47 59 68 73 74 Still, as the majority of instruments were developed focussing on patients over 65 years old, the suitability for middle-aged adults is unknown, yet both polypharmacy and multimorbidity are increasing in this age group.20 30 75 Only a handful of criteria have been developed and validated (eg, Prescribing Optimally in Middle-aged People’s Treatments criteria), all including significantly fewer criteria for individual medications and medication classes.54 56 63 76 77 Again, this further limits applicability in problematic polypharmacy, where the whole of the medication regimen should be considered.

Large variability in interventions addressing polypharmacy

Interventions to address problematic polypharmacy have covered a wide range of aims, such as reducing adverse drug reactions, increasing the appropriateness of medicines use, reducing falls, improving patient adherence and maintaining quality of life.13 78–81 To combat overprescribing specifically, deprescribing interventions have also received significant attention, though interventions that focus on underprescribing are much less.82–86 Several large reviews highlight good evidence of improving prescribing patterns, yet mixed and low certainty of evidence in improving patient-relevant outcome measures.2 13–15 80 87–90 Reviews covering over 150 primary studies reported no differences in all-cause mortality and no clear evidence of benefit in reduced hospitalisations, when comparing interventions to usual care.13 80 88 91–94 There were also no differences in quality of life, adverse drug reactions, readmission rates, primary care visits and emergency department visits.13 80 92–94 Two reviews have highlighted some economic benefits in reducing healthcare expenditure, but others highlight inconsistencies due to low-quality evidence.92 93 95 Overall, there is evidence that these interventions are safe and do not lead to harm, but may still be time and resource intensive for both patients and practitioners, as many require continuing input.13 80 82 Likewise, mixed evidence of improved clinical outcomes, such as falls, is also observed even in more focused populations, such as those with frailty and in long-term care facilities.84–86 96

Multiple intervention components to address polypharmacy, with unclear effectiveness

The majority of polypharmacy interventions were multimodal with a review revealing 14 different elements from 80 studies and an average of 2.5 elements per intervention.13 97 The most common elements included medication reviews, training for professionals and using tools, such as clinical decision support, checklists or audit and feedback.13 43 74 97–100 Other components strengthened interprofessional or multidisciplinary collaborations by involving clinical pharmacists, nurses or geriatricians.13 92 94 97 100–107 There were also patient-facing components, such as education and training for patients and patient interviews to seek their understanding and lived experiences with their medicines.108–113 Despite the growing literature on the importance of patient-centred care in medicines management, current literature highlights that patient priorities are seldom fully integrated into polypharmacy interventions.13 82 91 97 114–121 Patient-centred approaches also appear to be key to improving adherence, as a frequent discordance between practitioner and patient views is reported.13 15 97 122–128 More recent interventions that do adopt a patient-centred model show some mixed improvements in appropriate prescribing, but limited improvements in outcomes, reflecting some of the challenges of integrating patient priorities into routine medication reviews.99 108–113 129 130

In terms of effectiveness of individual intervention elements, similar effect sizes have been observed in reducing the number of potentially inappropriate medicines, with no particular components showing particular superiority.13 80 97 However, generalised professional education programmes appear to be less effective than individualised interventions.13 131 Medication reviews are also the most commonly adopted component, but as a single intervention, there remains insufficient evidence of medication reviews alone improve clinical outcomes.84 132 133 Despite the advantages of automation, electronic tools in trials demonstrate high variability in implementation within large pan-European and global trials, and no clear positive advantages on relevant patient outcomes have been reported.13 134 135 Pharmacists show promise as an extra resource for managing polypharmacy in individual studies, but two recent reviews revealed uncertain effects on optimising medicines.92 94 102–106 136 137 Community pharmacists can contribute to medication safety, but more in-depth management such as polypharmacy medication reviews was seen as outside the scope of community pharmacy.105 138 139

Key challenges in addressing problematic polypharmacy

In spite of the breadth of interventions targeting polypharmacy, it remains unclear which intervention components are more important.13 Theory-informed interventions are few and there are opportunities for improvements in intervention design through stronger foundations on theoretical frameworks and behaviour change techniques.128 140–144 Widespread variation exists in the everyday management of medicines and polypharmacy.2 3 145–147 These variations occur at patient, prescriber, regional and international levels, and indicate links between problematic polypharmacy and health inequalities.1–3 39 145 146 148–150 As such, multiple challenges to addressing problematic polypharmacy need to be overcome, going beyond the identification of individual barriers and facilitators and translating these into practice within the complexity of interlinked systems of care.2 39 151 152 The failure of the implementation of interventions is commonly down to the lack of consideration of integration into an already high-demand system in everyday primary care.152–155

For patients with polypharmacy and multimorbidity, prioritisation and decision-making are a challenge, given that they can receive 10 times the amount of information during consultations due to compounding health issues, interacting medications and complex social issues.156 Yet patient priorities and shared decision-making are vital to deciding the appropriateness of medications, so improvements need to be made to both the clarity of information provided and the integration of patient views into polypharmacy decisions.2 26 27 114 118 121 128 130 156 157 The majority of patients appear willing to discuss deprescribing medications, particularly if they have a good relationship with their doctor.82 105 118 135 155 However, they also have strong beliefs and attitudes of the value of their medicines, with inertia generated when feeling well on their current medication regimen.82 118 120 152 158–160

For health professionals, work and effort are required to even consider deprescribing, particularly as prescribing is so embedded in routine practice and finding an appropriate time to initiate the discussion is often difficult given competing priorities.153 154 161–163 A comprehensive polypharmacy medication review is described as ‘impossible’ to complete in 10 minutes, leading to practitioners defaulting to a swifter review and degrading medication reviews to being ‘mundane’ tasks.158 This is combined with the work to gain awareness (of new policies, guidelines and tools), overcome significant uncertainty in evidence (with ‘unmeasurable’ risk-benefit) and increase self-efficacy with limited resources and alternatives.149 154 162–167 On an organisational and systems level, fragmentation of care and poor coordination between healthcare teams and specialists often lead to deferring ownership of deprescribing, and miscommunication to patients, leading to medication-related problems.149 151 161 166 168 169 More comprehensive approaches and better resources are needed to support practitioners and organisations in pushing for improved polypharmacy decisions in a patient-centred manner, rather than simply maintaining the ‘status quo’.35 82 148 162 164