A systematic review of adherence in Indigenous Australians: an opportunity to improve chronic condition management

Background Indigenous Australians experience high rates of chronic conditions. It is often asserted Indigenous Australians have low adherence to medication; however there has not been a comprehensive examination of the evidence. This systematic literature review presents data from studies of Indigenous Australians on adherence rates and identifies supporting factors and impediments from the perspective of health professionals and patients. Methods Search strategies were used to identify literature in electronic databases and websites. The following databases were searched: Scopus, Medline, CINAHL Plus, PsycINFO, Academic Search Premier, Cochrane Library, Trove, Indigenous Health infonet and Grey Lit.org. Articles in English, reporting original data on adherence to long-term, self-administered medicines in Australia’s Indigenous populations were included. Data were extracted into a standard template and a quality assessment was undertaken. Results Forty-seven articles met inclusion criteria. Varied study methodologies prevented the use of meta-analysis. Key findings: health professionals believe adherence is a significant problem for Indigenous Australians; however, adherence rates are rarely measured. Health professionals and patients often reported the same barriers and facilitators, providing a framework for improvement. Conclusions There is no evidence that medication adherence amongst Indigenous Australians is lower than for the general population. Nevertheless, the heavy burden of morbidity and mortality faced by Indigenous Australians with chronic conditions could be alleviated by enhancing medication adherence. Some evidence supports strategies to improve adherence, including the use of dose administration aids. This evidence should be used by clinicians when prescribing, and to implement and evaluate programs using standard measures to quantify adherence, to drive improvement in health outcomes. Electronic supplementary material The online version of this article (10.1186/s12913-017-2794-y) contains supplementary material, which is available to authorized users.


Background
Chronic conditions such as type 2 diabetes mellitus, hypertension and kidney disease impair the health of many Aboriginal Australian and/or Torres Strait Islander people (hereafter the term 'Indigenous' is used). In 2012-13, an estimated 20% of Indigenous Australian adults had high blood pressure, 25% had elevated cholesterol and 18% had indicators of chronic kidney disease [1]. Chronic conditions lead to considerable morbidity and premature mortality, contributing to 80% of the difference in life expectancy between Indigenous Australians and non-Indigenous Australians [2].
Management of chronic conditions usually requires a combination of behaviour modification and taking prescription medicines; accordingly effective management relies in part on medication adherence. The determinants of adherence to medicines are complex and incorporate human behaviour, health literacy and adequate access to resources to support adherence. Not surprisingly, numerous studies in the 'general' population indicate that when adherence is suboptimal health outcomes are poorer [3]. Indigenous Australians continue to die almost ten years earlier than non-Indigenous Australians [1]. The Pharmaceutical Society of Australia recently stated that inadequate medication adherence will impede improvements in the life expectancy for Indigenous Australians living with chronic conditions [4] and many health professionals and researchers assert that adherence in this population is especially challenging. Two thirds of Indigenous Australians have at least one chronic condition [1] but there has been limited objective examination of the contribution of adherence to medicines to chronic condition management in this population. Very few studies have quantified adherence for this population or examined the association between adherence and clinical outcomes for Indigenous Australians.
Adherence in the context of Indigenous health in Australia has been reviewed previously, though not systematically nor focusing on chronic conditions, and no attempts to assess the quality of the existing literature have been made. Davidson et al. identified cost of medications, patient mobility and 'culturally alienating' health services as barriers to adherence, and recommended adherence support strategies including regimen simplification, use of dose administration aids (DAAs) and building the cultural competence of health professionals to strengthen relationships with patients [5]. A review in the Medical Student Journal of Australia noted that inadequate family support and culturally inappropriate service delivery could all impair adherence [6]. Strategies suggested for supporting adherence included training health professionals in 'cultural values and healthcare beliefs of Aboriginal and Torres Strait Islander communities' , increased use of interpreters and the subsidisation of medications for Indigenous Australians [6]. While many Indigenous Australians (35%) live in major city areas, the majority (44%) live in regional areas and 21% live in remote areas [7], so all health services in Australia need to provide culturally appropriate care.
In this era of escalating prevalence of chronic conditions, especially in populations of Indigenous Australians, clinicians and researchers require high-quality evidence to guide approaches to optimising adherence. The data on adherence rates for this population are often based on anecdote and information on appropriate strategies requires collation and analysis so that recommendations can be made.
Therefore we aimed to undertake a systematic literature review to provide a comprehensive compilation and examination of the literature on adherence to long-term medicines by Indigenous Australians living with chronic conditions. The primary objectives were: to synthesise data on the rates of adherence to medicines; explore health professionals' attitudes towards adherence, and examine the impediments to and supporting factors of adherence as reported by health professionals and patients. The secondary objective was to identify and collate data on the health outcomes associated with adherence.

Methods
The review was conducted and reported according to the PRISMA guidelines [8].

Eligibility criteria
Studies and evaluations of any design reporting original data on adherence to self-administered medicines for chronic conditions were included. The authors were aware that very few randomised controlled trials had been conducted to examine this issue, so a deliberately broad inclusion criterion was applied regarding study methodology. The study population had to include Aboriginal and/or Torres Strait Islander Australians and results needed to be disaggregated by ethnicity. Eligible clinical trials and intervention studies were included in the facilitators and barriers section; only baseline adherence rates from these studies (where reported) were included in the compilation of adherence rate data since adherence during clinical trials is often higher than in real-world settings [9]. To maximise information available on health professionals' perspectives commentary pieces by experts in the field were also included. Results were limited to articles published in English.
Most people with chronic conditions rely on self administered medications to manage their health, so studies reporting adherence to dialysis, chemotherapy and radiation therapy and directly administered medications (such as Benzathine Penicillin G injections) were excluded. Treatment and prophylaxis for tuberculosis were included as these treatment regimens are prolonged.

Information sources and search strategy
Articles were identified through database searches, citation searching and review by experts. The electronic databases searched were: Scopus, Medline, CINAHL Plus, PsycINFO, Academic Search Premier, Cochrane Library, National Library of Australia (Trove) (limited to books and theses), Indigenous Health infonet, and Grey Lit.org, from inception until 23 February 2015. In addition the Charles Darwin University library catalogue was searched and a Google search was conducted (limited to sites ending in .gov.au or org.au, the first ten pages of results reviewed). A shortlist of relevant references was distributed to experts in the field who were asked to identify any missing publications. In addition the reference lists of included articles were searched.
The search strategy was kept intentionally broad to ensure all data relevant to chronic conditions were included. The search strategy for Medline was: (((*adheren* OR *complian* OR concord*) and (treatment* or medicine* OR medication* OR drug*)) or (MM "Patient Compliance +" OR MM "Medication Adherence"and (indigenous or aborigin* or "torres strait" or MH "Oceanic Ancestry Group"). Limited to English language. See Additional file 1 for full details of the search strategy.

Study selection, data extraction and quality assessment
Data elements extracted were: study aim, methodology, site, dates; participant eligibility criteria; recruitment strategy; sample size; data collection method; data analysis strategy; results. Study quality was assessed using a tool adapted from McInnes and Chambers [10] (see Additional file 2). The study quality criteria were not applied to evaluation reports, correspondence or commentary pieces. Due to the limited data available on the subject, study quality score was not used to exclude articles or weight study findings. Data extraction and quality assessment were completed by the first author. Findings from qualitative studies were compiled in a narrative synthesis. See Fig. 1 below for a flowchart of the article selection process.

Results
The 47 included articles reported on people living with: HIV, diabetes, epilepsy, kidney disease, mental health issues, tuberculosis, cancer, cardiovascular disease and those who had experienced a seizure. Five sources reported rates of adherence and 14 reported barriers, facilitators or strategies, either from the patient or health professional perspective. Participants from all states and territories except the Australian Capital Territory and Tasmania were represented. The characteristics, results and quality scores for the journal articles are included in Table 1 below (when multiple articles reported on the same study the information has been combined). Summaries of the evaluation reports and letter in reply can be found at Additional file 3; commentary pieces were not summarised.
Study quality scores ranged from −4 to +96% (median: 54; IQR: 39). The criteria could not be applied to one study due to the absence of any methodological information on the assessment of adherence [11]. The most common issue was incomplete methodology reporting.

Adherence rates
Six articles included quantified adherence rates and these studies reported that approximately two thirds of Indigenous Australians take their regular medications at least some of the time (rates and definitions from individual studies are listed in Table 1) [12][13][14][15][16][17]. The methods used to quantify adherence were rarely reported in detail and studies differed in the way they quantified or categorised adherence; consequently a meta-analysis was not possible and comparison of findings was difficult. Two of these studies found that Indigenous Australians were less adherent than non-Indigenous Australians [12,13]. These findings were based on self-reported adherence. and we did not identify any studies validating self-reported adherence in this population so the accuracy of the data is unclear.

Outcomes of adherence
Few studies have explored associations between adherence and clinical outcome, or sought Indigenous Australians' views on this relationship. Some authors attributed poor clinical outcomes (including relapse of tuberculosis; failure of a kidney graft; recurrence of angina after heart surgery; and death after heart surgery) to inadequate adherence [18][19][20] and 82.9% of health professionals working in mental health services in the NT reported inadequate adherence as a common cause of relapse [21]. Predictors of clinical outcome are multifactorial, for example, poor clinical outcomes in Indigenous Australians with diabetes have been attributed to clinical inertia and inadequate dose titration rather than adherence problems [15].
In a study of cultural beliefs of disease causation (specifically, cultural factors affecting renal transplant outcome), just one patient connected their inadequate adherence to a negative health outcome (graft rejection); other patients instead nominated factors such as alcohol, poor nutrition and separation from kin and country as causes of poor health outcomes [22].

Attitudes about adherence rates
In the majority of studies, health professionals expressed the view that Indigenous Australians have inadequate adherence to medications [23,24] and this is having a negative impact on Indigenous health in Australia [22,[24][25][26][27]. In one sample 97% of health professionals believed it was a 'major or significant' problem [28]. Just one study reported a positive view of adherence, with an Aboriginal Health Practitioner saying that 'a good percentage are compliant' [25].
Many health professionals acknowledged the challenges faced by Indigenous Australians taking long-term medicines, but there were mixed views on patients' attitudes to medicines and adherence [29][30][31]. No studies reported patients' own perceptions of their adherence.
An issue reported by patients only was forgetting to take doses [31,40,45,46] and a few participants stated that their belief in God meant they did not need to take medicines [44][45][46]. In addition, health professionals reported that inadequate safe storage for medicines at home impaired adherence [30,47].

Enablers of adherence
Dose administration aids (DAAs) and other simple strategies have been associated with improved adherence for Indigenous Australians. Patients in two studies reported that DAAs were useful, [33,48] and in a small (n = 11) Central Australian study, the provision of DAAs increased patient adherence from 70% to 87% (p = 0.019). [16] Patients with HIV indicated that adjusting quantity/ timing of alcohol consumption assisted with adherence, [49] and patients with diabetes indicated once-daily dosing assisted [48]. Health professionals reported that a good understanding of Western medicine and establishing good rapport with patients was associated with good adherence [28].
Approaches which engaged communities, and involved Aboriginal Health Practitioners and family, appeared to improve adherence. For example, increased availability of Aboriginal Health Practitioners was associated with higher adherence to tuberculosis treatment in northern Queensland [19] and was part of a successful strategy implemented at an Indigenous Health Service in Brisbane [36]. The involvement of community members in medication dispensing in a remote NT community was also associated with increased adherence [43]. Both health professionals and patients reported that family support enhanced adherence [28,33].
Medication cost reductions via the Australian Pharmaceutical Benefits Scheme co-payment, have alleviated the financial burden and reportedly improved adherence [43,50].

Suggested strategies to improve adherence
The health professionals and participants interviewed in the eligible studies suggested a variety of approaches which they felt would enhance adherence; the proposed strategies aligned well with the barriers and facilitators described. Both groups felt that culturally appropriate resources designed to enhance the provision of patient education about medicines would increase adherence [17,24,33,40,51,52]. Patients also suggested the establishment of a support group [45,46]. Health professionals proffered a variety of adherence support strategies: increased involvement of Aboriginal Health Practitioners in medication management; [24,[52][53][54] simplification of dose regimens, including once-daily dosing and long acting formulations; [17,35,36,52,[54][55][56] provision of DAAs, [27,37,52,55] and the use of home medicines reviews [24,42,51] (although adaptation of this program is required to suit the needs of Indigenous Australians [25]). Some health professionals also emphasised the need to address social determinants of health [17,27,37,54].
A summary of the opportunities to improve adherence is provided in Table 2.

Discussion
We identified a range of publications addressing adherence among Indigenous Australians. The articles varied widely in research methodology and quality, preventing a meta-analysis, and some of the sample sizes were small (two studies each with 11 participants). However, despite differences in study design and quality, many findings were remarkably similar, indicating the value of including all eligible studies for this review.

Adherence rates
Health professionals were essentially unanimous in seeing inadequate adherence as an important issue in Indigenous health in Australia, but just two studies reported results which support the assertion that adherence for Indigenous Australians is significantly worse than for non-Indigenous Australians [12,13]. A number of studies found that two thirds of Indigenous Australians were adherent (based on varying definitions of 'adherence') [14,16,17], which aligns with findings of international studies across a range of populations and health conditions [57]. The potential discrepancy between provider beliefs and reality requires clarification through further studies using validated methods to measure adherence.

Barriers and facilitators of adherence
There was considerable overlap in the barriers and facilitators reported by health professionals and Indigenous Australians living with chronic conditions, indicating that many health professionals are aware of the challenges faced by their patients. Some patients reported that traditional or religious beliefs could cause or prevent ill health [22,[44][45][46] supporting the call for improved health literacy made by consumers and practitioners working in Indigenous health in Australia [17,24,33,40,51,52]. Some very context-specific barriers were identified, such as adherence challenges for remote-dwelling Indigenous Australians during travel away from home communities [32][33][34][35][36], however many barriers to adherence faced by Indigenous Australians are universal. Forgetting doses, complex dosing schedules, the cost of medicines, inadequate social support and alcohol use were all identified as key factors in a 2008 international review of adherence [58]. This study also found that minority groups had poorer compliance, but concluded that Table 2 Opportunities to improve adherence socio-economic status was more likely to explain lower compliance than race [58].

Strategies to enhance adherence
The studies provided a range of suggested and proven adherence support strategies relevant for Indigenous Australians however there was only one example of a strategy being tested and evaluated in a research study (see Additional file 3 for details) [16]. Two other studies suggest there could be further scope for pill burden minimisation for Indigenous Australians. The Kanyini GAP study (not included in this review because results were not disaggregated by ethnicity) showed that dose regimen simplification (through the use of a polypill) can enhance adherence in a population with a large proportion of Indigenous Australians [59,60]. Dose simplification also improved adherence in a New Zealand study which included a large proportion (50%) of Maori participants [61].
Other intervention studies testing adherence support strategies for Indigenous Australians have been published [62][63][64][65] or are being undertaken [66]; they provide relevant insights such as the value of directlyobserved treatment, but did not meet the selection criteria for this review.

Specificity of findings
Some of the findings reported in this review may be specific to the study contextfindings from remote settings for instance should not be extrapolated to Indigenous Australians living in urban areas, however it is worthwhile highlighting the consistency in information reported which was perhaps unexpected given the diversity of participant groups, which varied by condition, location, sex and age. Our findings are supported by previous non-systematic reviews of adherence among Indigenous Australians [5,6].

Remaining evidence gaps
Unfortunately few papers accurately quantified adherence and just five articles linked adherence to clinical outcomes (none of which reported any statistical results). Consequently the size of the problem, the validity of health professional perspectives and the potential gains which could be achieved by focusing on adherence support strategies remain unknown. In addition, very limited information has been collected from Indigenous Australians who live in remote communities and those who do not speak English fluently; so further research is required to capture their views [67].
While many solutions can be identified from the international literature, specific locally-relevant strategies are also needed. It is heartening to see some translation of research findings into practice; the 2005 National Health and Medical Research Council guidelines for strengthening cardiac rehabilitation of Indigenous Australians includes evidence-based strategies for adherence support [68]. Continuing concerns about adherence and the ongoing significant burden caused by cardiovascular disease [69] and the absence of evidence-based recommendations for the management of other chronic conditions indicate more needs to be done.
We need more evidence on which activities effectively support Indigenous Australians requiring long-term medicines. This review indicates we have the necessary information to develop tailored, locally-relevant strategies.

Strengths and limitations
The key strengths of this review are the inclusion of qualitative and quantitative studies, the inclusion of grey literature and the use of a quality assessment tool. A limitation of this study is that initial study selection, data extraction and quality assessment were conducted by one individual (J.L.dD), however ongoing input from researchers with experience in systematic literature reviews and chronic conditions ensured the rigour of the implementation of the methodology and the results.

Conclusions
'Closing the gap' in health outcomes for Indigenous Australians with chronic conditions requires lifestyle modification, changes in health-seeking behaviour and adherence support. Data on adherence rates are limited, but it is likely that, as is the case in many patient populations, suboptimal adherence means that Indigenous Australians are not receiving the full benefit of medications. The maximum benefit of medications for chronic conditions is obtained with 100% adherence and in some chronic conditions such as HIV and rheumatic heart disease, missed or late doses can be particularly harmful. Therefore the target adherence for all people living with chronic conditions should be 100%.
This review informs clinical practice in several key ways. Clinicians who presume low adherence among Indigenous Australian patients, and may choose not to prescribe certain medications based on this presumption, need to acknowledge that this belief is not evidence-based. Clinicians prescribing for Indigenous Australians need to utilise methods which improve medication adherence, including dose administration aids. Additionally, methods which Indigenous Australians have requested or practitioners have suggested (Table 2) should be incorporated, including involving Indigenous health practitioners, using family-and community-centred approaches, and culturally-appropriate educational resources to achieve good rapport.
This review also informs future research priorities. Rigorous measurement and reporting of adherence using standard measures is needed and this could be achieved by using routinely collected dispensing data to calculate the medication possession ratio (an indicator widely used in adherence research). Randomised controlled trials of interventions, incorporating strategies such as those described herein would provide the robust data needed in this field to be able to improve adherence and therefore improve outcomes among people with chronic conditions.
In addition to gathering adherence data, existing information on facilitators should be used to develop, implement and evaluate interventions. Policy makers and health service managers should allocate appropriate resources for program delivery and evaluation and the findings should be published with detailed methodology information to maximise the evidence base.
Finally, the term 'adherence' itself may not be ideal, since it fails to acknowledge the partnership required to optimise chronic condition management. The focus should be on the development and maintenance of a respectful, trusting relationship between patient and health professional, one that is not overshadowed by assumptions of poor adherence.  Table: Characteristics and findings of evaluation reports and correspondence. A summary of the relevant information extracted from evaluation reports and correspondence. (DOC 73 kb)

Acknowledgements
The lead author is supported by an Australian Postgraduate Award scholarship. The provider of this scholarship had no involvement in the design, conduct or analysis of this review. APR is supported by an Australian National Health and Medical Research Council fellowship (1084656).

Funding
Not applicable.
Availability of data and materials Data sharing not applicable to this article as no datasets were generated or analysed during the current study.

Authors' contributions
JdD developed and conducted the search strategy, identified the eligible articles, extracted the data and applied the quality assessment criteria. JdD was a major contributor to the writing of the manuscript. AR made substantial contributions to the design of the review and the analysis and interpretation of the data. AR was a major contributor to the writing of the manuscript. AC made substantial contributions to the design of the review and the interpretation of data. All authors read and approved the final manuscript.
Ethics approval and consent to participate Not applicable.

Consent for publication
Not applicable.