What can primary care services do to help First Nations people with unhealthy alcohol use? A systematic review: Australia, New Zealand, USA and Canada

Studies were included if: (1) they presented data on programs or interventions for reducing alcohol consumption or the harms caused by unhealthy alcohol use in First Nations individuals from Australia, New Zealand, United States of America, or Canada; (2) the program or intervention was aimed at individuals with known unhealthy drinking (i.e. above recommended limits ‘hazardous drinking’ or an alcohol use disorder), OR aimed at improving accessibility or acceptability of alcohol care in primary health care services; (3) the program or interventions were implemented in a primary care setting or other setting that was readily available to individuals in the community on an outpatient basis and; (4) data on First Nations participants was presented separately to other ethnic groups. Studies focusing on ‘Alcoholics Anonymous’ or similar mutual support groups were excluded as they are the subject of a separate review [48]. Studies which focused solely on screening were excluded. No language restrictions were implemented. Due to the dearth of research on this topic, no start date for the search was specified. Databases were searched from inception to 11 March 2020.

The search strategy was developed (GK, KC, KL) and then refined by two academic librarians. The search was conducted by one author (GK) and titles and abstracts of all retrieved articles were screened and reviewed independently by two authors (GK and KL). Discrepancies were discussed and resolved by consensus. Papers meeting inclusion criteria were subject to a full-text screen conducted independently by three authors (GK, KL, JC). Disagreements were debated and where unable to be resolved, an additional author adjudicated (KC). Reference lists of included articles from the full-text screen and articles recommended by colleagues were reviewed (GK) against the inclusion criteria. These additional papers were independently full-text screened (by GK and either KL or JC). No disagreements were found. A summary of the articles retrieved is presented in Fig. 1.

Data extraction was completed by GK. Two authors (KL and JC) independently extracted data from half of the articles each. Care was taken to not allocate these authors any article on which they were a co-author. Data extracted included country and Indigenous population, participant characteristics, data collection methods, nature of the conditions treated, system used to categorise participants’ alcohol use, treatment strategy type (western [i.e. mainstream or non-Indigenous] or cultural), study focus (e.g. treatment effectiveness, implementation), length of follow-up, intervention details and outcomes. Studies that focused on implementation research were categorised according to a framework developed for the WHO [49], as either implementation or effectiveness-implementation design. The latter examines both the treatment effectiveness and implementation processes. ‘Western’ studies were defined as those that did not explicitly mention Indigenous cultural values or practices. ‘Cultural’ studies were grounded in traditional practices and philosophies of the respective First Nations participants. The three authors resolved discrepancies to form a combined version of data. Data extraction was independently checked (KC).

Study quality and risk of bias assessment was conducted using two tools: a version of the AXIS critical appraisal tool [50], which was later adapted by its authors to include assessment of the quality of randomised controlled trial (RCT) and cohort designs; and an adapted version of a tool to assess community participation in research [51‐53]. The latter categorises First Nations community participation across seven levels, from no participation to self-mobilisation. Studies are classified on this scale based on participation at four time-points during research projects: diagnosis (identifying a community’s priorities), development (appropriate strategies to address the priorities), implementation (of the strategies) and evaluation (of the effectiveness of the project; Table 2). Scoring was conducted independently by three authors (GK, KL, JC). Discrepancies were discussed and resolved by consensus. Where participation data was not presented or unclear, an ‘unknown’ classification was allocated.

Studies examined treatments for a range of severity of unhealthy alcohol use, from hazardous use to alcohol dependence. Nine studies focussed on alcohol dependence or ‘alcoholism’ (sic) [24, 29, 56, 64, 69, 72‐75]. Two studies focused on alcohol use disorders more broadly (severity not specified) [30, 31]. No study was found to exclude clients with dependence. Of the remaining 17 studies, a variety of terms were used to describe the nature of the alcohol problem being treated (Table 3). No diagnostic criteria or screening thresholds were provided for the majority (n = 19, 68%) of studies (Table 4). Of the 28 studies, almost half (n = 13; 46.5%) were quantitative [24, 55, 57, 61, 62, 69, 70, 72‐77], nine were qualitative (32%) [31, 54, 56, 58, 60, 63, 64, 66, 71], and six were mixed methods (21.5%) [29, 30, 59, 65, 67, 68].

Study interventions, primary outcomes, and analytic strategies were highly varied. We classified studies according to their research focus (Table 3). Two-thirds (n = 18) of the studies focused on treatment accessibility or acceptability from the perspective of clients (n = 6), staff (n = 8) or both (n = 4) [30, 31, 55, 57, 65, 76]; [54, 56, 58, 60, 68, 70, 71, 77]; [63, 64, 66, 74]. The remaining one-third focused on treatment effectiveness (n = 4) [24, 69, 73, 75]; effectiveness-implementation (n = 3) [29, 62, 72] or implementation research (n = 3) [59, 61, 67]. One effectiveness-implementation study also evaluated client access and staff access and acceptability [29], and another study evaluated implementation and client and staff acceptability [59].

The majority of studies (n = 17; 60%) reported on western treatment approaches [24, 29, 30, 54‐61, 68, 70‐73, 75], with only 3 reporting specifically on cultural approaches (n = 3) [74, 76, 77] for addressing unhealthy alcohol use. Eight studies reported on both [31, 62‐67, 69]. The balance of western compared to cultural approaches varied by country. Of the Australian publications, the majority (n = 13/17) examined use or implementation of a western approach, most often BI [30, 54‐61, 68, 70‐72]. The two studies from New Zealand solely focused on the importance and acceptability of culture in alcohol care [76, 77]. Half of US-based studies (n = 4/8) included cultural approaches [31, 66, 69, 74]. The two Canadian studies explored bicultural approaches [63, 64].

There were only two controlled trials (one blinded RCT [24] and one non-randomised controlled trial [73]). The remainder of the quantitative studies had no control group (n = 5) [55, 61, 62, 69, 72]. Most studies were of moderate quality when examined with the AXIS critical appraisal tool [50].

Objectives and aims were clearly described, and study design was appropriate for the stated aims in two-thirds of studies (n = 25, 89%; n = 20, 71% respectively) [24, 29‐31, 55‐61, 63‐74, 76, 77]; [24, 29‐31, 54, 56‐60, 63‐68, 70‐77]. Study populations were clearly defined (n = 25, 89%) and taken from samples which were likely representative in most studies (n = 23, 82%) [24, 29‐31, 54, 57‐70, 72‐77]; [24, 29‐31, 54, 57‐63, 65‐68, 70, 72‐77]. However, most studies (n = 25, 89%) did not describe non-responders or missing data [24, 29‐31, 54‐59, 61‐65, 68‐77]. Methods were insufficiently described in a third of studies (n = 10, 36%) to allow replication [30, 31, 54‐56, 62, 64, 66, 71, 75]. Basic data (e.g. demographics) were not described in nearly half of studies (n = 13, 46%) [54‐56, 59‐64, 70, 71, 74, 75], of which, nine were published in the last decade [54‐56, 59‐64]. Discussion and conclusions did not appear justified by the results in a quarter of studies (n = 6, 21.5%) [24, 54‐56, 61, 69] and ethical approval was not mentioned in just under a third of studies (n = 9, 32%) [54, 63, 64, 70‐75].

Community participation by First Nations peoples varied (Table 5). The highest levels of participation (Levels 5 to 7) were found in both the ‘diagnosis’ phase of the research, and the ‘development’ phase, with nine studies (32%) each [24, 29‐31, 62, 65, 69, 73, 75]; [29‐31, 62, 64, 65, 69, 73]. This was followed by the ‘implementation’ phase (n = 7; 25%) [29‐31, 62, 64, 65, 73]. Four studies scored highly in all of the first three phases (14%) [30, 31, 62, 73]. Only one study scored highly in all four phases [29]. Participation was described with insufficient detail to be assessed in twelve studies (43%) [56‐61, 63, 67, 68, 74, 76, 77].

Western approaches were studied in a variety of ways. Some trialled the standard western treatment approach, while others engaged First Nations clients or staff in tailoring it for the Indigenous context. Western approaches were also used in conjunction with cultural therapies, sometimes as a bicultural strategy, and other times the cultural elements were outsourced to community organisations as ‘clip-on’ elements of care.

Brief intervention was the most studied approach; however, no study measured its effectiveness. Effective client–clinician communication is at the core of BI and is likely impacted by the communication preferences and cultural protocols of First Nations peoples [78, 79]. The one study which implemented a culturally-influenced version of BI [67], did not describe the development of the approach. Tailoring and implementation studies are needed to optimise BI delivery and should record client perceptions of acceptability, as well as effectiveness.

Ambulatory withdrawal management was examined in two Australian studies in relation to Aboriginal Community Controlled Health Services. For carefully selected individuals, this approach may help increase accessibility and acceptability of withdrawal management. The research also demonstrated how community partnerships can be a part of developing, running and evaluating such programs within a primary health care service.

A small number of studies trialled alcohol relapse-prevention medicines. The RCT of naltrexone reported significant reduction in alcohol-related consequences compared to placebo despite low participant numbers [24]. Inclusion criteria for the trial may have been a barrier for First Nation people’s participation as interested persons were required to provide confirmation of “tribal enrolment” or a “Certificate of Indian Blood”. Of the 28 included studies, this was the only study where verification of First Nations status was discussed. This may have been due to the study’s exploratory aim of looking at predicted responses to treatment based on genetic traits. Consultation with community partners may help to respectfully decide on terms of participation and reduce barriers to participation in clinical trials.

The two other studies involving relapse prevention medicines trialled disulfiram. Both were performed half a century ago (1968 and 1970) and involved administration of a test dose of alcohol to precipitate a reaction. This is not standard practice today and would raise ethical concerns. Due to the scarcity of relapse prevention medicine trials with First Nations peoples, it was decided to include these studies. Disulfiram is not as frequently prescribed as it once was, due to concerns over serious adverse reactions in the event of relapse to drinking and debate over its effectiveness, particularly with unsupervised dosing. Despite these concerns, disulfiram is still prescribed internationally for carefully selected and voluntary patients [80]. In Australia, lack of government subsidies on the cost has also been identified as a barrier to its use [81]. The models of support offered alongside disulfiram therapy in these two studies appeared to be culturally tailored and progressive for their era. However, given the studies were conducted over 50 years ago, client acceptability data may need to be re-examined. Acamprosate, was not trialled or mentioned in any study. In contrast, amongst the general population both acamprosate and naltrexone have been trialled extensively.

Underutilisation of these medicines is a population-wide trend, although a range of social determinants may see First Nations peoples being less likely to access alcohol pharmacotherapy than non-Indigenous populations [82, 83]. Based on the included studies, medicines to prevent relapse seem acceptable within a culture-informed framework [24, 56]. Methods to increase awareness of the medicines and their availability through primary care services, could be considered in future research [30].

Cultural healing practices and spiritual customs were a major theme in this review. Traditional healing approaches have also appeared in other areas of First Nation’s health research, responding to the communities’ call—“our culture is our treatment” [17, 42, 84, 85]. The integration of cultural protocols and traditional therapies into primary health care has been associated with the movement for cultural reclamation. These themes were highlighted in one of the earliest included studies [74]. Along with innovations in treatment content and tailoring of delivery to increase acceptability and accessibility for minorities, community leaders are pushing for control over what is studied and how research questions are framed [86]. First Nations Elders are resisting the framing of research in deficit-based rhetoric, in favour of strength-based approaches and stories of resilience [87‐89]. With this paradigm shift comes opportunities to create knowledge that incorporates the best of western, evidence-based practice with traditional knowledges, forming a third stream valuable in both spheres. This concept has been articulated in “two-eyed seeing” and “Ganma” philosophies [90, 91], from Canada and Yolngu Nation (Australia) respectively.

The notion of Indigenous culture in health care is tied to holistic definitions of health. This invariably involves connectedness to family, traditional homelands, spirituality, community identity and reciprocal relationships [26, 92, 93]. In research of the general population, factors such as connectedness have also been identified as important on the recovery journey for addictions [94, 95]. Likewise, connectedness has been identified as a protective factor against AUDs by First Nations peoples [66, 86, 96].

Overarching health policy has recognised the need for culturally-informed and tailored health care [38‐40, 97]. However, cultural and bicultural therapies are still largely unconsidered in the national alcohol treatment guidelines of all four included countries [16, 25, 98‐100].

The quality of the included studies was assessed from two lenses—from a (western) scientific lens and from a community values lens. We found reporting was more complete for scientific measures. Privacy concerns for smaller study sites may account for the lack of basic demographic data. Over half of the studies collected data through qualitative interviews and observations. While qualitative methods are important in understanding participants’ thoughts and feelings, it does not allow quantification of effectiveness. Limitations also apply to pre-post treatment comparison studies, as individuals are most likely to enrol in a trial when they are highly motivated or at a point of crisis. This can falsely inflate intervention success rates, as the participants may have already been on a path to abstinence or moderated drinking (regression to the mean). While there was only one RCT, strict conditions maintained in the study environment may not hold up in many routine clinical settings. In addition to further efficacy trials, studies establishing the effectiveness of interventions for improving unhealthy alcohol use outcomes in real world settings are also needed.

Almost half of the studies did not report on community participation. This is concerning as the value of community engagement (even outside community-based, participatory study designs) is well recognised and promoted through research ethical guidelines [46, 101, 102]. According to these guidelines, best practice with First Nations communities involves meaningful engagements, where all parties are consulted before the ‘problem’ is defined and well before study aims and designs are planned. Reporting details of the tailoring process may be challenging due to journal word constraints. Until this is a standard requirement of reporting, onus is placed on researchers to succinctly describe the processes undertaken.

It is clear from the included studies, there is often a divergence of beliefs (ontology), value systems (axiology) and ways of knowing what is truth (epistemology), between the scientific and First Nation communities. This is particularly apparent with the qualitative data identified in this review. An individual’s or community’s experiences, perceptions and opinions (anecdotal evidence) rate low in the scientific pyramid of evidence. Meanwhile, this type of evidence has an important place in Indigenous ways of knowing [42]. Future research collaborations with First Nations communities will have to navigate these differences to generate knowledge that is valuable in both knowledge bases (pluralism).

The small number of controlled trials and effectiveness studies identified in this review, is one example of the pressing need to build on community partnerships. For example, First Nations communities have expressed concern over ethical issues with RCT designs (withholding beneficial treatment from one part of the community) [103]. Alternative study designs such as the ‘stepped-wedge trial’ and ‘wait-list control’ have been approved as culturally safe by several communities [86, 104]. Given the cultural diversity within First Nations tribal groups, tailoring of treatment approaches to suit local cultural beliefs, values, protocols, and communication preferences is encouraged to promote client engagement with primary care services. Future studies could also consider attitudes to alcohol as a factor influencing treatment acceptability or delivery. Such attitudes may vary greatly between and within First Nations communities. Some communities may have a strong preference for abstinence as goal, based on past community experience of alcohol’s harms, desire to avoid risk behaviours arising from non-Indigenous culture, or missionary influence [6, 105]. For such communities, the idea of a controlled drinking goal (for non-dependent drinkers or for community) may be controversial [70]. Anecdotal evidence suggests in many communities (Indigenous and non-Indigenous) that there is a high awareness of the social harms from alcohol, but less awareness of its health effects. Any treatment efforts may need to be accompanied by sensitively conducted, community-based health promotion around alcohol.

To ensure comparability between studies, improved reporting of inclusion and exclusion criteria is needed, for example the thresholds used to define unhealthy drinking or alcohol use disorders (i.e. DSM or ICD criteria, or screening questionnaire score). Lastly, many promising treatment approaches have been abandoned after a single descriptive study, so there are many potential treatment approaches to build on and study.