Global Systematic Review of the Cost-Effectiveness of Indigenous Health Interventions

Abstract Background Indigenous populations around the world have consistently been shown to bear a greater burden of disease, death and disability than their non-Indigenous counterparts. Despite this, little is known about what constitutes cost-effective interventions in these groups. The objective of this paper was to assess the global cost-effectiveness literature in Indigenous health to identify characteristics of successful and unsuccessful interventions and highlight areas for further research. Methods and Findings A systematic review of the published literature was carried out. MEDLINE, PSYCINFO, ECONLIT, EMBASE and CINAHL were searched with terms to identify cost-effectiveness evaluations of interventions in Indigenous populations around the world. The WHO definition was followed in identifying Indigenous populations. 19 studies reporting on 27 interventions were included in the review. The majority of studies came from high-income nations with only two studies of interventions in low and middle-income nations. 22 of the 27 interventions included in the analysis were found to be cost-effective or cost-saving by the respective studies. There were only two studies that focused on Indigenous communities in urban areas, neither of which was found to be cost-effective. There was little attention paid to Indigenous conceptions of health in included studies. Of the 27 included studies, 23 were interventions that specifically targeted Indigenous populations. Outreach programs were shown to be consistently cost-effective. Conclusion The comprehensive review found only a small number of studies examining the cost-effectiveness of interventions into Indigenous communities around the world. Given the persistent disparities in health outcomes faced by these populations and commitments from governments around the world to improving these outcomes, it is an area where the health economics and public health fields can play an important role in improving the health of millions of people.


Introduction
There are almost 400 million Indigenous people living in countries around the world [1]. The World Health Organisation (WHO) defines Indigenous populations as those that live in distinct geographical territories, identify themselves as belonging to a cultural group separate from mainstream society and are descendent from groups present in the area before modern states and borders were defined [2]. Indigenous populations have been repeatedly shown to bear a greater burden of death, disease and disability than their non-Indigenous counterparts [1,[3][4][5][6]. Despite this, little is known about what constitutes cost-effective health interventions in these unique population groups and there is a lack of evidence as to the extent and nature of investment in programs to address the burden of ill-health in these populations [7,8].
Economic evaluation of health care programs has become an important area of applied economics over the last 30 years [8][9][10].
The field has had a significant impact on policy-making processes in countries around the world most prominently through high level policy initiatives such as the National Institute for Health and Care Excellence in the UK and the Pharmaceutical Benefits Advisory Committee in Australia. Economic evaluation analyses whether the additional benefits of an intervention is worth undertaking relative to another intervention or normal care [9]. Without a substantial evidence-base on the cost-effectiveness of policy options, policy-makers have little economic evidence to make resource allocation decisions in the field of Indigenous health.
The objective of this review is to systematically search the literature to pull together existing evaluations that estimate the cost-effectiveness of health interventions into Indigenous populations around the world. This review allows for an exploration of the type of interventions that have been shown to be effective in these unique population groups, the specific resource requirements needed to deliver programs to these populations and the aspects of these programs that are deemed to be of value by the populations to whom they are targeted.

Suitability of the traditional health economic paradigm for the Indigenous health field
There is an increasing recognition of the shortcomings of traditional methods of measuring health benefits in economic evaluations of healthcare programs [8,[10][11][12]. Indigenous populations have been shown to hold different conceptions of health to mainstream populations [1,8]. Specifically, family, community, connections to the land and cultural sensitivity have been shown to be qualitatively valued with regards to health by Indigenous populations [8]. While this has led some to argue that the traditional approaches of the health economic paradigm are inappropriate to measure the impact of interventions in these communities, it highlights the importance of further economic research into the field and suggests that valuation of outcomes within economic studies should ideally incorporate some form of patient or user-elicited valuation so that they reflect these broader conceptions of health.
To provide a basis for further evaluating these arguments, this review also highlights any explicit attempts by included studies to incorporate these concerns into cost-effectiveness evaluations.

Methods
A systematic review of the literature was conducted to find articles that provide an economic evaluation of interventions targeting or reporting on an Indigenous population. No protocol has been previously published for this review.

Inclusion Criteria
The inclusion criteria for this review specified three characteristics for studies. First, the studies had to examine interventions that were primarily aimed at improving the health of target populations. Second, included papers had to be economicevaluations of an intervention that met the definition of one of the types outlined in Table 1. This depicts a broad spectrum of economic evaluation methodologies from cost-benefit analysis, generally considered the form of economic evaluation that is most comprehensive in scope, to the narrower forms of evaluation including simple cost-analyses. The search strategy was adapted from previously published systematic reviews of economic evaluations [13,14]. Studies were required to report on either the cost impacts of the intervention of interest or some measure of cost-benefit to be included in the study. Finally the papers had to either focus on or separately report on a population that is either wholly or partially Indigenous. The WHO definition of Indigenous mentioned above was used as a basis for identifying these populations in the literature [2]. The specific search strategy was adapted from a previously published systematic review of this population group and is designed to encompass Indigenous populations around the world in line with this definition [15].

Search strategy
A search was conducted of CINAHL, MEDLINE, EMBASE, PSYCINFO and ECONLIT (from inception to May 2014) using variations of the search string contained in Table 2. Reference lists of included studies were also searched for further applicable studies.

Data Extraction
Study review, selection and data extraction were independently undertaken by two authors (BA and JM). Abstracts, titles and keywords of the studies returned from the search were screened for compatibility with the inclusion criteria. Once studies were identified for potential inclusion, full texts were reviewed. Data were extracted from the studies using a form developed for the review based on standard techniques used in the literature and included the following items: country of origin, methodology including type of evaluation, comparators used, outcome measures, settings and participants, results and evidence of inclusion of Indigenous conceptions of health [16]. The primary outcome measures were reported measures of cost-effectiveness.

Results
The search yielded 559 abstracts (see Figure 1). One further study was identified through a hand search of relevant journals. Once duplicates were removed, 458 unique abstracts were reviewed. Three published abstracts were found that appeared to meet the inclusion criteria. Relevant authors were contacted to identify if further publications resulted from these abstracts. The authors of one study did not respond [17], one was published as a short-report that included no additional information to the abstract [18] while one was being readied for submission and not yet available to other researchers [19]. Two of the abstracts were included as studies as they contained enough information to meet the data extraction requirements [17,18] while the other was excluded as there was insufficient information in the abstract to be included [19]. In total, nineteen studies reporting on 27 interventions met the criteria to be included in this review [7,11,17,18,[20][21][22][23][24][25][26][27][28][29][30][31][32][33][34]. Tables 3 and 4 summarise the characteristics of included studies.

Outcome Measures
Costs were the primary outcome measure reported in eight of the studies [11,17,20,22,24,25,30,33]. One study reported costbenefit ratios for five injury-prevention interventions [34]. The study of a remote dental service in Western Australia attempted to use published valuations of equivalent government services as an estimate for the value of services provided and reported on the cost-benefit ratio in this regard [21]. Five studies reported costs per quality adjusted life years gained [23,26,31,32] and two reported on cost per disability life years gained [7,18]. Costs per life years saved were reported by Sheerin et al. in their study on Hepatitis C treatment for injecting drug users in New Zealand [29], while cost per avoided hospitalisation was the focus of the study of Reeve et al. on palivizumab injections for high risk infants [27].
Each included paper made some judgment as to the costeffectiveness or cost-impact of the interventions being studied. In total, only five of the 27 interventions were deemed to be strictly not cost-effective or cost-saving [7,11,27,32]. Of the eight costanalysis studies, six found interventions would be cost-saving over time relative to the comparison [17,20,22,24,25,30]. The study of the group midwifery program in the Top End of Australia found no significant cost differences between the provided intervention and usual care [33], while the urban midwifery program was deemed not cost-saving although qualitative results demonstrated that patients valued the service [11]. The cost-benefit analyses carried out all found the injury prevention interventions to be cost-beneficial in that they had benefit-cost ratios above one [34]. Two of the three cost-effectiveness studies found their interventions to be cost-effective [21,29]. Three interventions analysed using costutility analysis were found to cost-ineffective [7,32]. Table 5 depicts included interventions by cost-effectiveness.

Comparators Used
The choice of comparator against which the cost-effectiveness of the intervention is assessed plays a large part in determining whether a particular intervention is cost-effective or not. Included studies could be grouped into two main categories in this regard. Fourteen of the studies assessed the cost-effectiveness of their intervention against so-called 'business as usual' cases where they were compared to a situation with no intervention, either through the use of a control group [11,28,30,32], projecting historical trends [7,17,20,25,26,29,31,33,34] or in one case actual hospital records with the impact of the intervention being retrospectively assessed [27]. The remaining five studies compared the costeffectiveness of the intervention to a hypothetical or alternatively modelled scenario where participants would be forced to obtain the intervention through alternative service providers [18,[21][22][23][24].

Wider conceptions of health
Only one study that met the inclusion criteria explicitly set out to capture wider benefits of culturally appropriate service provision [11]. This was done through broader qualitative evaluation of the value of the provided midwifery service and the use of a cost-consequence approach designed to incorporate broader values than narrowly defined health outcomes. Another included study also included qualitative results from patient interviews to document the appropriateness of the service to the Indigenous community [22]. The work of the ACE prevention project in Australia modelled the different impact of delivering interventions via Aboriginal Medical Services and mainstream general-practitioners, suggesting that health benefits to Indigenous communities would be higher from those services delivered through Aboriginal Medical Services as a result of increased

Cost Analysis
A partial form of economic analysis where only the costs of an intervention are compared to another potential intervention.

Cost-Effectiveness Analysis
Provide a measure of the effectiveness of an intervention using natural units such as life-years gained relative to the cost incurred to obtain that outcome.
Cost-Utility Analysis A particular form of Cost-Effectiveness Analysis that measures effectiveness using a measure of utility such as Quality Adjusted Life Years or Disability Adjusted Life Years.
Cost-Benefit Analysis An economic evaluation that sees both the costs and benefits of a particular intervention valued in monetary terms.

Database/s Search terms
Cinahl, Medline, EMBASE and PsycINFO (1) The following terms as words within the title, abstracts or texts of papers: aborigin* or american indian* or eskimo* or Ethnic Group* or first nation* or greenlandic or indigenous or inuit* or inupiat* or inuvialuit* or kalaallit* or maori or maoris or mapuche* or native american* or native people* or native population* or native siberian* or navaho* or nunangat* or sami* or skolt* or taiga* or tribe or tribal or yuit or yupik or zuni (2) ''cost-effectiveness'' or ''economic evaluation'' or ''cost impact'' as words within the title, abstracts or texts of papers or containing a subject heading under ''cost analysis/'' and $1556 (USD) were avoided as a result of screening engagement of the target population [7]. The remaining studies did not explicitly attempt to measure any wider or Indigenousspecific conceptions of health. It is important to note, though, that of the 27 interventions included in the analyses, only four were not targeted specifically at Indigenous populations [26][27][28][29]. Of the modelled studies, six of the eleven interventions examined were based on estimates of intervention effectiveness that had been obtained from studies carried out in Indigenous populations [7,[26][27][28][29]31]. The other five were based on effectiveness estimates emanating from the general non-Indigenous literature and applied to Indigenous populations.

Discussion
This systematic review has found that very few cost-effectiveness studies are available in the published global Indigenous health literature. This has implications for generating investment into Indigenous health programs since the lack of such evidence limits our ability to assess the investment-case of interventions based on the criteria of cost-effectiveness, rather than solely for equity reasons or broader policy objectives. This is concerning given the significant disparities in health and access to health care that exist between Indigenous and non-Indigenous populations worldwide. Nonetheless, the evidence-base that does exist in the literature provides some isolated insights into the potential cost-effectiveness of specific types of interventions. There is potential for further work to both increase the use of economic evaluation in this area and methodological work to ensure that health economic methodologies are relevant to Indigenous populations.

Limited Economic Research in the Field of Indigenous Health
A total of nineteen studies met the inclusion criteria for this review. Given the broad scope of the research question and search strategy, this depicts a very limited evidence-base from which to draw insights on the potential cost-effectiveness of interventions into Indigenous populations. This finding implies that investment into the area is largely being undertaken blind, based on assumptions rather than evidence of the cost and effectiveness of particular policies and interventions.
Three main reasons are offered here as potential factors explaining the lack of research in the field. First, there is a general lack of effectiveness studies in the field of Indigenous health, with the majority of research carried out in the field being observational rather than interventional in nature [35]. The findings of this review build upon previous literature that has emphasised the need to focus further research on finding effective interventions to improving the health of these unique groups. Second, the lack of studies could be a reflection on the attitudes of policy-makers and service providers in the area of Indigenous health that due to the severe inequalities faced by Indigenous population groups, interventions are justified purely on equity grounds rather than cost-effectiveness considerations. Such an argument overlooks the role that economic research could play in not only highlighting the relative importance of investing in Indigenous health, but providing guidance as to the best use of resources within the sector to maximise their impact. Third, as a result of the unique conceptions of health held by Indigenous populations, traditional economic evaluation methodologies may be inappropriate for the field of Indigenous health [8]. Specifically benefits have generally been measured using individualistic and Westernised constructs of health, which have been shown, at least in the Australian context, to be distinctly different from the communitarian values of Indigenous culture [8]. This has potentially led to missed opportunities to improve indigenous health and wellbeing as the full range of benefits, costs and costsavings that potentially arise from indigenous health interventions may not be captured. There is room for further refinement of economic methodologies to incorporate these concerns and particularly in the assessment of the applicability of existing health utility instruments, and potentially the development of new ones that may be more sensitive to Indigenous conceptions of health. While the evidence-base identified in this review demonstrates that traditional health economic approaches can ostensibly be used to show the cost-effectiveness (or otherwise) of interventions aimed at improving Indigenous health, there has rarely been much attempt to incorporate Indigenous valuations of the potential gains from such programs.
Examples do exist in the literature of attempts to incorporate these values into economic evaluations of health interventions.
The study of the Daruk-controlled midwifery service included above attempted to do so by taking a broader cost-consequence approach rather than restricting the analysis to narrowly defined health outcomes [11]. This involved a strong qualitative component to the study, which is an approach also taken by Jin et al. in their included study above [22]. Cost-consequences analyses or 'the basket of goods approach' has been viewed with some disdain in the health economic literature because of the potential for data mining and its inability to generate a clear decision rule [37]. It has been suggested however that this can be to some extent addressed by pre-specification of a conceptual framework in which the link between the multiple outcomes are linked to the intervention [38]. Further research into the suitability to of existing health economic techniques used to elicit preferences from target populations to Indigenous groups could provide a means to incorporate these values in a meaningful and rigorous way.

Characteristics of Included Studies
While there was a limited amount of cost-effectiveness research of interventions to improve the health of Indigenous population Cost-Effectiveness in Indigenous Health groups, a number of conclusions can be drawn from the studies that were identified.
First, this review highlights that interventions into Indigenous populations, in particular rural and remote Indigenous communities, can be cost-effective, a broad but important finding given the often isolated and small populations of these groups when compared to non-Indigenous populations. Of the 27 interventions examined by included studies, 21 were deemed to be cost-effective or cost-saving while of the twelve interventions targeted solely to rural and remote populations, ten were found to be cost-effective or cost-saving by the respective studies. Being able to point to a body of evidence highlighting the cost-effectiveness of such interventions is important to justify widespread implementation of such programs on more than solely equity grounds and ensure that domestic debates on service provision are informed and based on the best available evidence.
The evidence-base drawn together by this review provides insights into particular interventions. Outreach programs were shown to be consistently cost-effective or cost-saving in all six interventions studied. These interventions were assessed relative to populations obtaining the services from alternative service providers. Thus they were found to be cost-saving or cost-effective despite often representing relatively high-costs for the health gains that occurred. They may be prohibitively costly in low-income environments outside of donor provision and it is also unclear how sustainable or community appropriate such models of service provision are. Explicit decisions need to be made by policy-makers in assessing the appropriateness of these services to local conditions.
The four injury prevention interventions were all found to be cost-effective as was the study of telehealth. Conversely, neither of two midwifery programs studied was found to be strictly costeffective (one found non-significant cost-savings). Nor were the two studies focused solely on interventions in urban areas, highlighting the need for further work in this area. The majority of the interventions were targeted specifically at Indigenous groups (twenty-three interventions) and most were delivered through culturally specific medical providers, such as Aboriginal Medical Services in Australia and Canada. This is in line with findings from the literature that culturally specific services are more effective in reaching these populations [39]. The studies identified were largely drawn from high-income nations and focused at a primary-care level in rural or remote populations. There is a large gap in the literature for Indigenous populations of low and middle-income nations. Similarly, urban Indigenous communities have been largely overlooked in the literature despite these communities often representing the bulk of Indigenous populations within countries. In Australia, for example, it is estimated that the 60% of the 'gap' in health outcomes between the Indigenous and non-Indigenous populations is a result of the health of urban Indigenous communities [40]. The results of this review are consistent with previous studies that have pointed to a lack of studies in the area of urban Indigenous health relative to remote and rural communities [41]. Indigenous conceptions of health were rarely explicitly acknowledged in the cost-effectiveness literature. While the finding that the majority of included interventions were Indigenous-specific suggests that these factors may be implicitly be worked into most of the studies (at least to the extent that interventions are appropriately designed for these cultural groups), there is room to better include these ideas into general economic methodologies.

Conclusion
Despite global commitments to reducing Indigenous health disadvantage, relatively little is known about what constitutes costeffective investments into Indigenous populations around the world. Furthermore, the evidence that exists has often relied on pivotal evidence extrapolated from non-Indigenous settings and been based on methods that have not allowed for the values that such communities place on health to be included. Nevertheless, in light of the limited available evidence, this review suggests that interventions into these often hard to reach populations can be cost-effective. Further economic research has the potential to provide much needed guidance to policy-makers on resource allocation decisions and help improve the health of Indigenous people around the world but it needs to be based on the development of methods that incorporate values specific to the communities in question. .

Supporting Information
Checklist S1 PRISMA checklist.