Ethics and Community Involvement in Syntheses Concerning American Indian, Alaska Native, or Native Hawaiian Health: A Systematic Review

Abstract

Background

The objective of the research was to review reporting of ethical concerns and community involvement in peer-reviewed systematic reviews or meta-analyses concerning American Indian, Alaska Native, or Native Hawaiian (AI/AN/NH) health.

Methods

Text words and indexed vocabulary terms were used to query PubMed, Embase, Cochrane Library, and the Native Health Database for systematic reviews or meta-analyses concerning AI/AN/NH health published in peer-reviewed journals, followed by a search through reference lists. Each article was abstracted by two independent reviewers; results were discussed until consensus was reached.

Results

We identified 107 papers published from 1986–2012 that were primarily about AI/AN/NH health or presented findings separately for AI/AN/NH communities. Two reported seeking indigenous reviewer feedback; none reported seeking input from tribes and communities. Approximately 7% reported on institutional review board (IRB) approval of included studies, 5% reported on tribal approval, and 4% referenced the sovereignty of AI/AN tribes. Approximately 63% used evidence from more than one AI/AN/NH population study, and 28% discussed potential benefits to communities from the synthesis research.

Conclusions

Reporting of ethics and community involvement are not prominent. Systematic reviews and meta-analyses making community-level inferences may pose risks to communities. Future systematic reviews and meta-analyses should consider ethical and participatory dimensions of research.

INTRODUCTION

Current regulations focus only on individual participants. However, other individuals and communities not directly involved in research can also bear risks that are associated with a research study, but that are not acknowledged or mentioned in the regulations. If IRBs focus only on risks to individual research participants, they fail to apply fully the principle of beneficence, because the scope of this principle can extend beyond the individual participants in the research. –

National Bioethics Advisory Commission (2001, 73)

Academic research conducted on human subjects in the United States may aspire to the principles of beneficence, justice, and respect for persons articulated in the Belmont Report (US Department of Health and Human Services 1979). If the research is federally funded, managed, or regulated, it is subject to oversight by an Institutional Review Board (IRB) (U.S. Department of Health and Human Services 2009). The Food and Drug Administration (FDA) also regulates research according to human subjects research and IRB provisions given in Title 21 Parts 50 and 56 of the Code of Federal Regulations (). However, systematic reviews and meta-analyses of published results are often exempt from IRB review (U.S. Department of Health and Human Services 2009) because they are based on previously published findings. This does not imply that the activity of research synthesis (e.g., bringing together different studies for a larger inference, as in systematic review or meta-analysis) has no ethical implications. We argue that, as research syntheses draw stronger conclusions about population health than primary research papers, they allow greater potential for group-level harms.

Ethics reporting in systematic reviews has received little attention (Vergnes et al. 2010). Several systematic reviews have considered the ethical quality of published primary literature (Alexin and Salantera 2008; Block, Khitin, and Sade 2006; Cherny et al. 2009; Karlawish et al. 1999; Marziali, Serafini, and McCleary 2005; Munung et al. 2011; Pitak-Arnnop et al. 2009; Schroter et al. 2006) or in dissertations (Kjellstrom and Fridlund 2010; Kjellstrom, Ross, and Fridlund 2010). However, to our knowledge, no systematic analysis has been conducted to evaluate how ethics are reported in review articles or meta-analyses. Vergnes et al. (2010) identified four main ethics-related concerns in conducting a systematic review: (1) inclusion of unethical research; (2) validity of informed consent obtained at the individual study level for synthesis; (3) possible conflicts of interest; and (4) challenges in how to define and measure “ethical” research. Vergnes et al. (2010) also noted that the popular MOOSE (Stroup et al. 2000) and PRISMA (Liberati et al. 2009) guidelines lack ethical considerations as part of the checklists, and that IRB approval is not required for most synthesis research. We would like to elaborate on the ideas of Vergnes et al. by arguing that community-level inferences lent authority by the systematic review process (Greenhalgh 1997) could be more stigmatizing to communities targeted for inference than the findings of primary research studies. Stigma may be a driver of health inequalities (Hatzenbuehler, Phelan, and Link 2013), and therefore authoritative scientific research that entrenches perceptions about a community (Foucault 1980) might contribute to health disparities through generation or perpetuation of stigmatizing inferences.

Ethical concerns surrounding the dynamics between communities and research are also relevant for synthesized inference about community health. A recent review and framework (Oliver et al. 2008) noted that the public might be engaged as consultants or collaborators, or even control the research agenda, while researchers could be equal collaborators, take a hands-off approach, or dominate the project. Public stakeholders may be involved in the process of research (Boote, Telford, and Cooper 2002; Hanley et al. 2001) or in shaping a broader research agenda (Oliver et al. 2004; Rosenstock, Olenec, and Wagner 1998). While there are diverse perspectives on stakeholder opinions for research syntheses (Gøtzsche and Ioannidis 2012), in the absence of adequate communication, the preferences of stakeholders and researchers may be misaligned (Tallon, Chard, and Dieppe 2000), potentially resulting in mistrust of findings and strained research relationships in the future. In this paper, we focus on the involvement of American Indian, Alaska Native, or Native Hawaiian populations (hereafter referred to as AI/AN/NH) in synthesis research as a particular example of community relationships and the research process.

There are special ethical and community concerns when working with AI/AN/NH data. Complex histories between AI/AN/NH and United States institutions contribute to the present-day mistrust of academic research and the need for additional considerations by researchers when working with these geographically and culturally diverse populations (Boyer et al. 2011; Smith 1999). Historically, many research studies involving AI/AN/NH were developed, designed, implemented, and interpreted by outsiders, compromising the integrity of findings and alienating tribal communities (Boyer et al. 2011; Mihesuah 1993). Possible harms from research conducted in this manner include stigmatization of communities, undermining of traditional and ancestral knowledge, and possible disruption of economic viability (Boyer et al. 2011; Manson et al. 2004; Sahota 2007; Santos 2008). A recent lawsuit against Arizona State University, in which blood samples collected for diabetes research were used to draw conclusions about the Havasupai people without individual or tribal permission, illustrates how research with individuals might impact tribal communities (Sahota 2007). To prevent future harms, AI/AN/NH governments and affiliated organizations have organized to create policies and structures for regulating research, including guidance, suggestions for codes of research ethics, and resource kits for researchers and for AI/AN/NH communities (Committee on Native American Child Health and Committee on Community Health Services 2004; Lavelle, Larsen, and Gundersen 2009; Macaulay et al. 1998; Mihesuah 1993; Quigley 2006; Sahota 2007 ; Santos 2008; Sharp and Foster 2002). Synthesis developed collaboratively with communities is a form of community-based research (Israel et al. 1998) because it is a process that generates knowledge in partnership with communities.

A joint effort between the Cochrane Collaboration’s public health group and indigenous health groups identified challenges for systematic reviews concerning indigenous populations (McDonald et al. 2010). In particular, they noted a perspective among indigenous colleagues that reviews should be grounded in “[u]nderstanding of colonization, dispossession, racism and assimilation as key determinants of inequities and inequalities experienced by indigenous peoples,” and that “[i]ndigenous involvement in research processes is considered an issue of rigor” (McDonald et al. 2010, 643). Synthesis research neglecting the social, legal, and historical context of indigenous health issues, or imposed on communities by external investigators, may miss the mark (McDonald et al. 2010). Research should also be presented accessibly for the target populations (McDonald et al. 2010).

There has been substantial health research on indigenous communities in the United States. A review of papers indexed in PubMed or PsychLit (Sanson-Fisher et al. 2006) found an increase over the past several decades in the number of publications concerning indigenous health in the United States. As the literature drawing synthesized conclusions about AI/AN/NH populations grows, it is important to consider the reporting and contextualizing of information concerning these populations, and identify positive examples for future syntheses. The objective of this study was to describe the reporting of research ethics and community involvement in systematic reviews and meta-analyses concerning AI/AN/NH health.

METHODS

Article Identification

On September 10, 2012 we began a systematic review following a pre-specified but not prospectively registered protocol (Appendix 1) by querying for controlled vocabulary terms and text words in four databases: PubMed, Embase, Cochrane Library, and Native Health Database (see Appendix 1).

After consolidating citations and removing duplicates there were 114 unique citations to evaluate. We gathered full text of each (except those that were abstracts only) and independently applied inclusion/exclusion criteria. Each included article must have been a peer-reviewed journal-published systematic review or meta-analysis; available in the English language; and either focused on AI/AN/NH health (including minority health or health disparities research that incorporated AI/AN/NH communities) or on a health topic presenting AI/AN/NH findings in a way that could pose a risk to communities. We defined “risk to communities” as either the separate presentation of any AI/AN/NH results, or conclusions about AI/AN/NH communities in the context of results for health topics that may be sensitive for communities (i.e., diabetes, obesity, depression, alcohol or other substance abuse, fertility, children’s health, maternal and reproductive health, social and psychological health, individualism and self esteem, genetic data, cardiovascular disease, or cancer). The idea motivating our inclusion criteria was to identify studies with potentially stigmatizing findings for AI/AN/NH communities.

Studies were excluded if they were not systematic reviews or meta-analyses; were only abstracts; were not published in a peer-reviewed journal; or if they reported only on First Nations (Canada) or Aboriginal (Australia or New Zealand) populations, or other populations not clearly identifiable as United States AI/AN/NH. We recognize that indigenous populations throughout the world have histories of colonization and may experience challenges similar to AI/AN/NH in the United States; however, we felt it would be inappropriate to extend our review to settings outside of the United States given our limited knowledge of the processes specific to areas outside the United States context. We defined “health” broadly but excluded physical anthropology papers (i.e., papers with conclusions about human evolution) both because we do not consider anthropology a health science and because several of the ethical problems in that literature have been recently discussed (Turner 2005). Disagreements about article eligibility were resolved by consensus.

We supplemented the initial set of records with a search (conducted by one reviewer) through the reference lists of each article that was identified in the database search, regardless of that article’s inclusion status, because the search might have found a text string in the full text of the article. The search through the reference lists of included articles looked for titles or subject headings suggestive of systematic reviews or meta-analyses. Both authors examined the full text of each identified article to evaluate it for inclusion. Several preliminarily included papers were revealed, on closer reading, to not satisfy inclusion criteria; we still allowed relevant citations in those papers’ reference lists to be included as part of the search process. We included only the most recent version of relevant Cochrane reviews for this paper. One article identified from reference lists could not be obtained and was excluded. The overall search process is summarized in a PRISMA (Liberati et al. 2009) flow diagram (Figure 1). We used Endnote X3 for citation management.

Figure 1.

PRISMA Flow Diagram.

Article Abstraction and Synthesis

Each author independently applied a standardized form to each paper to evaluate the reporting of specific items (Table 1).¹ Most of these questions were straightforward, referring to specific things that were either present or absent, but one was more holistic asking about the article’s discussion of benefits for AI/AN/NH communities. We defined “benefits for AI/AN/NH communities” broadly to encompass discussion of how the research could be used for positive change in AI/AN/NH communities. We computed the kappa coefficient to quantify the extent to which determining the presence of discussion of benefits is prone to inter-reviewer disagreement. Disagreements during article abstraction were resolved by consensus. We compared equivalence of proportions across two categories (e.g., proportion primarily concerned with AI/AN/NH health, comparing articles from database search vs. articles from reference lists) using the score test; and independence of multiple categories (e.g., specificity of identifying AI/AN/NH communities, by whether the article focused on AI/AN/NH health) by χ² test. These statistics are for data summary rather than hypothesis-testing purposes.

Table 1.

AI/AN Systematic Review Study Quality Form

Quality considerations for the systematic review of ethical reporting in systematic reviews of AI/AN health
Year of Publication:__________
First Author: __________
Topic: ____________________
Include which AI/AN groups are included in the studies in their review (population summary)
Journal: __________
Write “Cochrane” if this is a Cochrane review
Was this systematic review primarily interested in AI/AN health?
It counts as “yes” for any studies of ethnic/racial dimensions of health including AI/AN as a comparison group.
___Yes
___No
___Uncertain
Are the inferences that can be made about AI/AN health from this paper a possible risk to communities?
___Yes
___No
Does the article address any of the following potentially sensitive topics specifically for AI/AN communities?
___Drinking or Alcohol
___Non-Alcoholic Addictive Behavior
___Depression or Suicide
___Other Mental Illness
___Cardiovascular Disease
___Diabetes
___Cancer
___Diseases of Children
___Diet or Nutrition
___Drinking Water or Sanitation
___Sexual Behavior or Sexually-Transmitted Diseases
___Traditional Medicines
___Tribal Ceremonies or Other Traditional Practices
___Tobacco Use
___None (not offensive)
___Other: __________
Do the authors state included studies had IRB approval?
___Yes
___No
Do the authors state that included studies had tribal consent or approval?
___Yes
___No
Do the authors state that included studies had Memoranda of Agreements between research institutions and tribes?
___Yes
___No
Does the article include any discussion of genetic information from AI/AN communities?
___Yes
___No
Does the review ever mention issues of tribal sovereignty?
___Yes
___No
Does the paper mention urban/rural/reservation jurisdictional issues?
___Yes
___No
___Not relevant
Does the paper state that relevant tribes were contacted for permission or design of the systematic review, or appropriateness of the systematic review question?
___Yes
___No
Does the paper discuss potential benefits of the research for AI/AN communities?
___Yes
___No
How specific are the authors in naming the AI/AN/NH populations discussed?
___Use name of specific tribe(s) and/or reservation(s)
___Refer to region only (ex. 2 tribes in the Southwest)
___Use only general terminology (ex. American Indian, Alaska Native, Native Americans, etc.)
Was there more than one study of AI/AN/NH health summarized in their review?
Mark “Yes” if there were at least 2 relevant studies; Mark “No” if there was only one study which was used to generalize about AI/AN.
___Yes
___No
Did the authors describe in this paper how studies were ascertained for their systematic review?
___Yes
___No
Do the authors include unpublished studies in their systematic review?
“Published” means peer-reviewed literature; it does not include abstracts.
___Yes
___No
Did the authors mention any conflict of interests in the included studies?
___Yes
___No
Do the authors mention any conflict of interests in their own study?
___Yes
___No
Other Comments on Quality: ____________________
If no comments, write “none”, do not leave blank.

RESULTS

Our search identified 107 studies satisfying our criteria (Figure 2, Appendix 2);² 44 of these were primarily about AI/AN/NH health, while 63 were on health topics and included AI/AN/NH data that might pose a risk to AI/AN/NH communities. The proportion of papers primarily about AI/AN/NH health was different between articles obtained by database searches and articles from reference lists (p = 0.001); most studies not primarily about AI/AN/NH were found in reference lists. We found no study reporting every dimension of research ethics and community involvement (Figure 2). None of these articles mentioned conflicts of interests in the summarized studies, but three mentioned possible financial conflicts of interest in their own study (Bangalore et al. 2011; Fowkes et al. 2008; Paradies 2006), and one review mentioned that its authors had been involved in all the relevant primary research (Naylor et al. 2003).

Figure 2. Patterns of Reporting in Systematic Reviews and Meta-Analyses Concerning American Indian, Alaska Native, or Native Hawaiian Health.

Each column represents an article, and each row represents a possible characteristic of that article. Filled dots indicate the article had that characteristic; empty dots indicate that the article did not have that characteristic. Row percentages describe the proportion of articles in our sample with each characteristic. Articles are sorted by whether they were primarily focused on American Indian, Alaska Native or Native Hawaiian populations.

Our independent judgments for discussion of research benefits had 87% agreement (kappa 0.68) (Figure 3). In our consensus judgment, only 28% of the included articles discussed the benefits of the synthesis for AI/AN/NH communities (Figure 2). There was no trend in the proportion of articles reporting benefits over time, either for all articles or stratifying articles by focus (or not) on AI/AN/NH health. There was a significant relationship between an article’s focus (or not) on AI/AN/NH health and the specificity of that article’s language identifying AI/AN/NH communities (p=0.03). However, there was not a pattern of articles focused on AI/AN/NH being consistently more or less specific in identifying communities.

Figure 3.

Reviewer Agreement on Standardized Abstraction Form Item “Does the paper discuss potential benefits of the research for AI/AN/NH communities?”.

No study reported asking AI/AN/NH communities about the appropriateness of the systematic review questions or the conduct of the review, though the two Cochrane reviews discussing global indigenous issues did work with aboriginal Australians. No articles we identified discussed Memoranda of Agreements between researchers and AI/AN/NH tribes and communities, at either the level of primary or synthesis research.

Although the overall reporting pattern we observed was one of neglect for reporting research ethics or community involvement, several studies set positive examples illustrating how future syntheses on AI/AN/NH health might address these issues. One study considered stakeholder inclusion in reviewed programs and evaluated the ethical quality of reviewed programs (Flynn et al. 2006). That review’s “propriety” evaluation for programs was based on the Center for Disease Control and Prevention’s (CDC) Framework for Program Evaluation (1999), which in turn was based on the recommendations of the Joint Committee on Standards for Education Evaluation (1994). Several studies acknowledged the limited quality of the evidence available for an inference to AI/AN/NH communities. One paper (Montag et al. 2012) juxtaposed the limitations of their work against the beneficial inferences that “[n]onetheless…may be gleaned” from their synthesis (441). Several studies (Montag et al. 2012; Pollack et al. 2012; Teufel-Shone et al. 2009; Vimaleswaran et al. 2008) noted community approval for primary research. The two Cochrane reviews specifically about indigenous health solicited indigenous (e.g., Aboriginal Australian) input throughout the conduct and reporting of the study (Carson et al. 2012a; Carson et al. 2012b). Several reviews acknowledged AI/AN/NH community heterogeneity and the differences between urban and rural experiences (Carson et al. 2012a; Carson et al. 2012b; Teufel-Shone et al. 2009). Several papers referenced sovereignty or the legal, historical, or psychosocial context (Carson et al. 2012b; Gone and Alcantara 2007; Pollack et al. 2012). Several studies went beyond journal requirements to provide a plain language summary to make findings accessible to a general reader. One of the Cochrane systematic reviews on indigenous health (Carson et al. 2012a, 4) included a “Why It Is Important to Do This Review” section heading and Teufel-Shone et al. (2009, S31) included a “SO WHAT? Implications for Health Promotion Practitioners” box. Perhaps most importantly, several reviews (Carson et al. 2012a; Montag et al. 2012; Pollack et al. 2012) had substantial discussions of AI/AN/NH benefits from doing the research. All these practices could be incorporated into future reviews on AI/AN/NH health. (See Table 2 for other recommendations.)

Table 2.

Suggestions for Various Stakeholders in Synthesis Research

Stakeholders	Suggestions
Researchers	Consider systematic reviews and meta-analyses concerning human populations as a form of human subjects research, with similar ethical principles at stake. Consider whom the research benefits and who bears what risks from the work. Form partnerships and solicit input from relevant stakeholders. Consult with tribal, regional, or national AI/AN/NH organizations, as appropriate, potentially with a third party to facilitate communication (i.e., AI/AN/NH healthcare providers, primary research authors) Feedback can be solicited at multiple stages of the systematic review process. Consider developing guidelines as a research community for reporting ethical issues.
Journal Editors	While ethical guidelines for publishing differ across journals (Pace et al. 2008), evidence suggests that editorial guidelines can have an impact on quality of ethics reporting (Karlawish et al. 1999). Editors should: Encourage authors to consider reporting on ethical aspects of any research. Discuss benefits of any evidence synthesis research for included communities, and if absent, ask authors to justify omission of these things from articles (e.g., the lack of discussion is per community request.
Tribes and Communities	Consider developing explicit terms for giving permission to use data for research (i.e., may it be pooled for original data meta-analysis) and whether investigators should state appropriate tribal permissions were obtained in all publications. Consider contacting researchers or academic institutions with whom you are already engaged or plan to engage, with suggestions for systematic reviews on community health priorities. It is a positive thing for ideas and priorities for community health to originate from communities.

DISCUSSION

Most systematic reviews and meta-analyses concerning AI/AN/NH health in this sample had little discussion of ethics or community involvement.

We estimated that only 28% of the identified papers discussed benefits of the synthesis research for AI/AN/NH communities. The kappa statistic for the pre-consensus responses to this question was 0.68, which the guidelines recommended by Landis and Koch (1977, 165) classify as “substantial.” However, there were systematic differences in how the two reviewers interpreted the “benefits” question. One initially felt that only studies mentioning specific benefits for AI/AN/NH health should get credit, while the other initially felt that any discussion of benefits for a population inclusive of AI/AN/NH communities was relevant. While we anticipated some discrepancy in defining “benefits” (Vanderpool and Weiss 1987), the high inter-reviewer agreement in our study for this item indicates that “discussion of benefits” might be sufficiently objective for other systematic reviews to incorporate this consideration into their evaluations. There was no apparent change in this proportion over time, but we hope that through drawing attention to this deficit that a greater proportion of future reviews may discuss the benefits to communities of research synthesis.

We excluded several studies from this review that might be thematically relevant to AI/AN/NH health because we felt they neither were primarily about AI/AN/NH health nor, as presented, posed risks to AI/AN/NH tribes or communities. Specifically, we excluded studies not focused on AI/AN/NH health that found no AI/AN/NH literature and made no assertions about AI/AN/NH populations (Fuentes and Aranda 2012; Jaja et al. 2008; Park and Taylor 2007). We also excluded others not focused on AI/AN/NH health that found AI/AN/NH literature but did not discuss it in the AI/AN/NH context, did not present the AI/AN/NH results separately, and drew no AI/AN/NH-specific inferences (Herbst et al. 2005). We acknowledge this is a subjective line to draw and that others may have included these references in this review.

This may be the first systematic review of reporting of research ethics in systematic review and meta-analysis literature, or for literature on AI/AN/NH health. Methodological strengths of our study include a rigorous and systematic search strategy; the inclusion of an American Indian perspective in writing a review related to American Indian health; independent double article selection (except for initial identification of review articles from reference lists) and abstraction; and a standardized abstraction form that measured multiple domains of reporting research ethics and community involvement. Methodological weaknesses of our exploratory study include possible selection bias from a non-comprehensive search, and the possible lack of content validity (DeVellis 2003) (e.g., omission of relevant domains of ethics or community involvement reporting) of our improvised study evaluation form. Our ability to interpret observed gaps in ethics reporting is also limited, as this review did not consider the primary literature on which these 107 references were based, so we cannot distinguish whether the omission is due to review authors or primary literature authors; if the omission was in the primary literature it may have been at the request of participating tribes. In general, we are summarizing what has been published in the peer-reviewed literature.

Another limitation of this review, realized during our review process, is inattention to non-systematic narrative reviews, or books, on AI/AN/NH health. While searching for systematic reviews and meta-analyses, we encountered many non-systematic review articles drawing inferences specifically on AI/AN/NH communities (Alcantara and Gone 2007; Brown 1995; Butler et al. 1999; Gahagan and Silverstein 2003; Quigley 2006). We also excluded several academic books (Joe 2008; Rhoades 2000; Roubideaux and Dixon 2001) that may also be viewed as resources for integrated AI/AN/NH health inference. How these neglected writings discussed issues of ethics and community involvement may be more germane to characterizing the “synthesis literature” published about AI/AN/NH health than the set of articles we included that were not primarily about AI/AN/NH health but made a comment about AI/AN/NH health based on a single citation.

It was beyond the scope of our review to systematically evaluate the primary literature of the published reviews, but we noticed that the Strong Heart Study (Lee et al. 1990) and the Pathways to Health Study (Davis et al. 1995) have been heavily cited, and therefore especially influential for AI/AN/NH inference, in systematic reviews and meta-analyses on AI/AN/NH health. Both these studies had input from tribes and communities as a core feature of the design and implementation of the research. However, even if the evidence base is ethical and participatory, when there is only a single study or unreliable evidence on which to base inference, the main effect of synthesis may be to reduce discussion of the context under which the original evidence was obtained. This may include missing caveats about to whom appropriate inferences may be extended: research syntheses may attempt a broader generalization than primary studies, and not all primary research is based on representative samples of communities. Even the Strong Heart Study, which enrolled participants from tribal rolls with a high participation rate (Lee et al. 1990), is not representative of AI/AN/NH, as it is a cohort of rural communities that may differ in many respects from urban AI/AN/NH populations, and its study population reflects the study design (~1,500 participants each from three study centers: Arizona, Oklahoma, and North or South Dakota) rather than the population distribution of the AI/AN/NH communities across those geographic regions. The scientific goals of primary research studies may be different from research syntheses incorporating those studies (Vergnes et al. 2010).

Even though we are not drawing any conclusions about American Indian health, our review nevertheless poses risks and benefits to AI/AN/NH tribes and communities. Consolidating a large number of reviews facilitates inferences about AI/AN/NH that are only as appropriate as the cited reviews and primary literature were rigorous, and our review did not evaluate the general scientific rigor of identified studies. Someone reading only our bibliography could get a stilted impression of the health issues that are important for AI/AN/NH communities, since the relative frequency of systematic reviews on a given topic reflects the priorities of researchers and the limitations of our search, not necessarily the priorities of AI/AN/NH communities.

Systematic reviews have an important role in shaping research agendas (Armstrong et al. 2010). Including community partners in the planning of a systematic review is one way to ensure that communities will have a voice in the research agenda going forward. It may also be fruitful to share the findings from a completed systematic review with a community targeted for inference to see that community’s response and thus identify additional research needs; this has an analogy to “closing the loop” in other forms of community-based research (Macdonald et al. 2012; Payne-Sturges, Schwab, and Buckley 2004).

In general, we think that communities who are at risk of community-level research harms should be included in the process of defining, and potentially restricting, the research that could generate those harms. In that vein, we have solicited feedback on our manuscript from AI/AN/NH researchers who were not authors of any included studies in our methodologic review. These correspondents have perspective on both communities facing likely harms from our review: research communities that may be negatively affected in the short term by an article pointing out deficiencies in current practice, and AI/AN/NH communities that may be affected by the inferences facilitated by the discussed systematic reviews on AI/AN/NH health. This is not a replacement for community discussion in research with greater potential for stigmatizing inferences.

We feel it is important to include stakeholders at the level of the potential harms (i.e., community-level voices for studies allowing potentially stigmatizing inferences about community health) and would encourage future reviews making health inferences about AI/AN/NH community health to contact communities rather than AI/AN/NH individuals. The most appropriate level of community consultation may be tribal, regional, or national (i.e., National Congress of American Indians) depending on the scope of the synthesized inference. It may help researchers performing synthesis-level research to contact the authors of primary research studies to seek help in obtaining feedback from their partner communities. Similarly, AI/AN/NH healthcare providers (i.e., the Indian Health Service) might be another point of contact for AI/AN/NH communities about the appropriateness of synthesis research. If relevant community partners cannot be reached or cannot provide feedback, authors should recognize that community voices are being excluded, and take deliberate steps to minimize the potential harms (e.g., not present findings separately for AI/AN/NH communities). However, this investigator-solipsistic approach has serious drawbacks: investigators may not be able to identify or mitigate all risks to communities from synthesis research; investigators may deprive communities of benefits of targeted research; and investigators could compromise the validity of results if pooled study populations come from very different contexts. When possible, we think it is far better to consult with community stakeholders to conduct appropriate community-level research.

Conclusions

Current systematic review and meta-analysis literature on AI/AN/NH health does not often discuss issues of research ethics or community involvement; however, our review identified several positive examples for best practices that we hope will improve the reporting of ethics and community involvement in future syntheses. We hope that in the future researchers will consider the potential for community-level benefits and harms from research synthesis, and incorporate community stakeholder perspectives into the research synthesis process.

Appendix 1. Additional Details of the Systematic Review Protocol

The objective of this systematic review is to investigate whether systematic reviews and meta-analyses on AI/AN/NH health include discussion of ethical issues and/or tribal consent.

Steps of Protocol

I. Search Strategy

1. Generate a list of keywords for AI/AN/NH literature from brainstorming session and review of other literature.

2. Compare this set of terms against the MESH and Emtree controlled vocabulary terms in PubMed and Embase.

3. Consider sensitivity analyses of various search strings for identifying systematic reviews and meta-analyses of American Indian text, including: alternate spellings of meta-analysis (“meta analysis”), omitting “systematic” from “systematic review”, omitting “systematic review OR meta-analysis”, using “systematic review” AND “meta-analysis” instead of OR.

4. Consider sensitivity analyses of additional terms for AI/AN/NH including “AI/AN/NH” and “Native”.

After optimization per the above steps, the following search terms were used to identify systematic reviews or meta-analyses specific to AI/AN/NH health topics:

PubMed (67 results)

(“indians, north american”[MeSH Terms] OR “American Indian” OR “American Indians” OR “Native American” OR “Native Americans” OR “Alaska Native” OR “Alaska Natives” OR “AI/AN” OR “Native Hawaiian” OR “Native Hawaiians”) AND (“systematic review” OR “meta-analysis”)

Embase (97 results)

(‘american indian’/exp OR ‘american indian’ OR ‘native american’/exp OR ‘native american’ OR ‘american indians’ OR ‘native americans’ OR ‘alaska native’ OR ‘alaskan native’ OR ‘alaskan natives’ OR ‘ai/an’ OR ‘native hawaiian’) AND (‘systematic review’/exp OR ‘systematic review’ OR ‘meta-analysis’)

Cochrane Library (2 results)

(“indians, north american”[MeSHTerms] OR “American Indian” OR “American Indians” OR “Native American” OR “Native Americans” OR “Alaska Native” OR “Alaska Natives” OR “AI/AN” OR “Native Hawaiian” OR“Native Hawaiians”)

Native Health Database (14 results)

“systematic review” OR “meta-analysis”

II. Citation Management

Lists of citations from each database search were saved separately and then loaded into Endnote X3. Citations recovered by searching through relevant articles’ reference lists were tracked using Google Doc excel tables. Articles identified as relevant through the reference list search were added one at a time to the Endnote X3 library. We used the “groups” function in Endnote X3 to keep separate the articles from original database searches, the articles from the reference list search, and articles that we used to contextualize our findings. Article eligibility was evaluated by two independent reviewer assessments and differences were discussed and resolved by consensus. Decisions concerning article relevance were tracked in Excel or Google Doc tables. We used Powerpoint to produce the PRISMA flow chart.

III. Data Abstraction, Management and Analysis

We used a Google Doc Form questionnaire to standardize abstraction across articles, available for viewing here: https://docs.google.com/spreadsheet/viewform?formkey=dHJXcHI2Qi04SGFIMjZXak9KY2E1NGc6MA#gid=0

Responses to our Google Doc Form populated a table, including a time stamp for each abstraction, which we downloaded into Excel. Reviewer disagreements were highlighted in Excel and discussed. A consensus abstraction table was saved as an Excel file and loaded into R for data analysis. The code used to generate Figure 2 and calculate statistics reported in this paper is available on request.

IV. Pre-Submission Peer Review

It is essential for this manuscript that we incorporate diverse perspectives from members of AI/AN/NH communities. Therefore, after preparing a draft of our manuscript, we emailed individual contacts that were members of AI/AN/NH communities, interested in research ethics, and currently active in research. We asked them if they would provide feedback on our draft concerning appropriateness for AI/AN/NH and in return be acknowledged in our paper. If the contact replied that they would provide this service, we explicitly requested that they keep the draft strictly confidential.

Appendix 2. Articles Meeting Inclusion Criteria

Authors, Year	Title	Field	Journal Info
Rhoades DA. 2006.	National health data and older American Indians and Alaska Natives.	Aging	Journal of Applied Gerontology 25(1): 9S–26S.
Li D, Zhao H, Gelernter J. 2012.	Further clarification of the contribution of the ADH1C gene to vulnerability of alcoholism and selected liver diseases.	Alcohol	Hum Genet 131(8): 1361–1374.
Montag A, Clapp JD, Calac D, Gorman J, Chambers C. 2012.	A review of evidence-based approaches for reduction of alcohol consumption in native women who are pregnant or of reproductive age.	Alcohol	Am J Drug Alcohol Abuse 38(5): 436–443.
Williams RC, Jacobsson LTH, Knowler WC, Del Puente A, Kostyu D, McAuley JE, et al. 1995.	Meta-analysis reveals association between most common class II haplotype in full-heritage native Americans and rheumatoid arthritis.	Autoimmunity	Human Immunology 42(1): 90–94.
Mahoney MC, Michalek AM. 1991.	A meta-analysis of cancer incidence in United States and Canadian native populations.	Cancer	Int J Epidemiol 20(2): 323–327.
Marcus AC, Crane LA. 1998.	A review of cervical cancer screening intervention research: implications for public health programs and future research.	Cancer	Prev Med 27(1): 13–31.
Maskarinec G, Takata Y, Chen Z, Gram IT, Nagata C, Pagano I, et al. 2007.	IGF-I and mammographic density in four geographic locations: a pooled analysis.	Cancer	Int J Cancer 121(8): 1786–1792.
Michalek AM, Mahoney MC. 1990.	Cancer in native populations-lessons to be learned.	Cancer	J Cancer Educ 5(4): 243–249.
Sheth S, Bedford A, Chopra S. 2000.	Primary gallbladder cancer: recognition of risk factors and the role of prophylactic cholecystectomy.	Cancer	Am J Gastroenterol 95(6): 1402–1410.
Sun L, Yu S. 2012.	Diabetes mellitus is an independent risk factor for colorectal cancer.	Cancer	Dig Dis Sci 57(6): 1586–1597.
Vasilevska M, Ross SA, Gesink D, Fisman DN. 2012.	Relative risk of cervical cancer in indigenous women in Australia, Canada, New Zealand, and the United States: a systematic review and meta-analysis.	Cancer	J Public Health Policy 33(2): 148–164.
Whop LJ, Valery PC, Beesley VL, Moore SP, Lokuge K, Jacka C, et al. 2012.	Navigating the cancer journey: A review of patient navigator programs for Indigenous cancer patients.	Cancer	Asia Pac J Clin Oncol.
Hardeman W, Griffin S, Johnston M, Kinmonth AL, Wareham NJ. 2000.	Interventions to prevent weight gain: a systematic review of psychological models and behaviour change methods.	Cardiovacsular	Int J Obes Relat Metab Disord 24(2): 131–143.
Bangalore S, Kumar S, Lobach I, Messerli FH. 2011.	Blood pressure targets in subjects with type 2 diabetes mellitus/impaired fasting glucose: observations from traditional and bayesian random-effects meta-analyses of randomized trials.	Cardiovascular	Circulation 123(24): 2799–2810, 2799 p following 2810.
Carroll S, Dudfield M. 2004.	What is the relationship between exercise and metabolic abnormalities? A review of the metabolic syndrome.	Cardiovascular	Sports Med 34(6): 371–418.
Carty CL, Buzkova P, Fornage M, Franceschini N, Cole S, Heiss G, et al. 2012.	Associations between incident ischemic stroke events and stroke and cardiovascular disease-related genome-wide association studies single nucleotide polymorphisms in the Population Architecture Using Genomics and Epidemiology study.	Cardiovascular	Circ Cardiovasc Genet 5(2): 210–216.
Coble JD, Rhodes RE. 2006.	Physical activity and Native Americans: a review.	Cardiovascular	Am J Prev Med 31(1): 36–46.
Conn VS, Phillips LJ, Ruppar TM, Chase JA. 2012.	Physical activity interventions with healthy minority adults: meta-analysis of behavior and health outcomes.	Cardiovascular	J Health Care Poor Underserved 23(1): 59–80.
Danesh J, Lewington S, Thompson SG, Lowe GD, Collins R, Kostis JB, et al. 2005.	Plasma fibrinogen level and the risk of major cardiovascular diseases and nonvascular mortality: an individual participant meta-analysis.	Cardiovascular	JAMA 294(14): 1799–1809.
Davis AM, Vinci LM, Okwuosa TM, Chase AR, Huang ES. 2007.	Cardiovascular health disparities: a systematic review of health care interventions.	Cardiovascular	Med Care Res Rev 64(5 Suppl): 29S–100S.
Dumitrescu L, Carty CL, Taylor K, Schumacher FR, Hindorff LA, Ambite JL, et al. 2011.	Genetic determinants of lipid traits in diverse populations from the population architecture using genomics and epidemiology (PAGE) study.	Cardiovascular	PLoS Genet 7(6): e1002138.
Fernandez A, Sorokin A, Thompson PD. 2007.	Corneal arcus as coronary artery disease risk factor.	Cardiovascular	Atherosclerosis 193(2): 235–240.
Fesinmeyer MD, North KE, Ritchie MD, Lim U, Franceschini N, Wilkens LR, et al. 2012.	Genetic Risk Factors for BMI and Obesity in an Ethnically Diverse Population: Results From the Population Architecture Using Genomics and Epidemiology (PAGE) Study.	Cardiovascular	Obesity.
Fowkes FG, Murray GD, Butcher I, Heald CL, Lee RJ, Chambless LE, et al. 2008.	Ankle brachial index combined with Framingham Risk Score to predict cardiovascular events and mortality: a meta-analysis.	Cardiovascular	JAMA 300(2): 197–208.
Gami AS, Witt BJ, Howard DE, Erwin PJ, Gami LA, Somers VK, et al. 2007.	Metabolic syndrome and risk of incident cardiovascular events and death: a systematic review and meta-analysis of longitudinal studies.	Cardiovascular	J Am Coll Cardiol 49(4): 403–414.
Heald CL, Fowkes FG, Murray GD, Price JF. 2006.	Risk of mortality and cardiovascular disease associated with the ankle-brachial index: Systematic review.	Cardiovascular	Atherosclerosis 189(1): 61–69.
Kirk JK, Bell RA, Bertoni AG, Arcury TA, Quandt SA, Goff Jr DC, et al. 2005.	Ethnic disparities: Control of glycemia, blood pressure, and LDL cholesterol among US adults with type 2 diabetes.	Cardiovascular	Annals of Pharmacotherapy 39(9): 1489–1501.
Kurian AK, Cardarelli KM. 2007.	Racial and ethnic differences in cardiovascular disease risk factors: a systematic review.	Cardiovascular	Ethn Dis 17(1): 143–152.
Maskarinec G, Takata Y, Pagano I, Carlin L, Goodman MT, Le Marchand L, et al. 2006.	Trends and dietary determinants of overweight and obesity in a multiethnic population.	Cardiovascular	Obesity (Silver Spring) 14(4): 717–726.
Mau MK, Sinclair K, Saito EP, Baumhofer KN, Kaholokula JK. 2009.	Cardiometabolic health disparities in native hawaiians and other pacific islanders.	Cardiovascular	Epidemiologic Reviews 31(1): 113–129.
McDowell SE, Coleman JJ, Ferner RE. 2006.	Systematic review and meta-analysis of ethnic differences in risks of adverse reactions to drugs used in cardiovascular medicine.	Cardiovascular	BMJ 332(7551): 1177–1181.
Mottillo S, Filion KB, Genest J, Joseph L, Pilote L, Poirier P, et al. 2010.	The metabolic syndrome and cardiovascular risk a systematic review and meta-analysis.	Cardiovascular	J Am Coll Cardiol 56(14): 1113–1132.
Ruige JB, Assendelft WJ, Dekker JM, Kostense PJ, Heine RJ, Bouter LM. 1998.	Insulin and risk of cardiovascular disease: a meta-analysis.	Cardiovascular	Circulation 97(10): 996–1001.
Siri-Tarino PW, Sun Q, Hu FB, Krauss RM. 2010.	Meta-analysis of prospective cohort studies evaluating the association of saturated fat with cardiovascular disease.	Cardiovascular	Am J Clin Nutr 91(3): 535–546.
Teufel-Shone NI, Fitzgerald C, Teufel-Shone L, Gamber M. 2009.	Systematic review of physical activity interventions implemented with American Indian and Alaska Native populations in the United States and Canada.	Cardiovascular	Am J Health Promot 23(6): S8–32.
Vimaleswaran KS, Luan J, Andersen G, Muller YL, Wheeler E, Brito EC, et al. 2008.	The Gly482Ser genotype at the PPARGC1A gene and elevated blood pressure: A meta-analysis involving 13,949 individuals.	Cardiovascular	Journal of Applied Physiology 105(4): 1352–1358.
Yancey AK, Kumanyika SK, Ponce NA, McCarthy WJ, Fielding JE, Leslie JP, et al. 2004.	Population-based interventions engaging communities of color in healthy eating and active living: a review.	Cardiovascular	Prev Chronic Dis 1(1): A09.
Fagot-Campagna A, Pettitt DJ, Engelgau MM, Burrows NR, Geiss LS, Valdez R, et al. 2000.	Type 2 diabetes among North American children and adolescents: an epidemiologic review and a public health perspective.	Childhood Diabetes	J Pediatr 136(5): 664–672.
Bautista-Castano I, Doreste J, Serra-Majem L. 2004.	Effectiveness of interventions in the prevention of childhood obesity.	Childhood Obesity	Eur J Epidemiol 19(7): 617–622.
Bluford DA, Sherry B, Scanlon KS. 2007.	Interventions to prevent or treat obesity in preschool children: a review of evaluated programs.	Childhood Obesity	Obesity (Silver Spring) 15(6): 1356–1372.
Bond M, Wyatt K, Lloyd J, Welch K, Taylor R. 2009.	Systematic review of the effectiveness and cost-effectiveness of weight management schemes for the under fives: a short report	Childhood Obesity	Health Technol Assess 13(61): 1–75, iii.
Campbell KJ, Hesketh KD. 2007.	Strategies which aim to positively impact on weight, physical activity, diet and sedentary behaviours in children from zero to five years. A systematic review of the literature.	Childhood Obesity	Obes Rev 8(4): 327–338.
Connelly JB, Duaso MJ, Butler G. 2007.	A systematic review of controlled trials of interventions to prevent childhood obesity and overweight: a realistic synthesis of the evidence.	Childhood Obesity	Public Health 121(7): 510–517.
Doak CM, Visscher TL, Renders CM, Seidell JC. 2006.	The prevention of overweight and obesity in children and adolescents: a review of interventions and programmes.	Childhood Obesity	Obes Rev 7(1): 111–136.
Dobbins M, De Corby K, Robeson P, Husson H, Tirilis D. 2009.	School-based physical activity programs for promoting physical activity and fitness in children and adolescents aged 6–18.	Childhood Obesity	Cochrane Database Syst Rev(1): CD007651.
Flodmark CE, Marcus C, Britton M. 2006.	Interventions to prevent obesity in children and adolescents: a systematic literature review.	Childhood Obesity	Int J Obes (Lond) 30(4): 579–589.
Flynn MA, McNeil DA, Maloff B, Mutasingwa D, Wu M, Ford C, et al. 2006.	Reducing obesity and related chronic disease risk in children and youth: a synthesis of evidence with ‘best practice’ recommendations.	Childhood Obesity	Obes Rev 7 Suppl 1: 7–66.
Kamath CC, Vickers KS, Ehrlich A, McGovern L, Johnson J, Singhal V, et al. 2008.	Clinical review: behavioral interventions to prevent childhood obesity: a systematic review and metaanalyses of randomized trials.	Childhood Obesity	J Clin Endocrinol Metab 93(12): 4606–4615.
Salmon J, Booth ML, Phongsavan P, Murphy N, Timperio A. 2007.	Promoting physical activity participation among children and adolescents.	Childhood Obesity	Epidemiol Rev 29: 144–159.
Sharma M. 2006.	School-based interventions for childhood and adolescent obesity.	Childhood Obesity	Obes Rev 7(3): 261–269.
Snethen JA, Broome ME, Cashin SE. 2006.	Effective weight loss for overweight children: a meta-analysis of intervention studies.	Childhood Obesity	J Pediatr Nurs 21(1): 45–56.
Stice E, Shaw H, Marti CN. 2006.	A meta-analytic review of obesity prevention programs for children and adolescents: the skinny on interventions that work.	Childhood Obesity	Psychol Bull 132(5): 667–691.
Stone EJ, McKenzie TL, Welk GJ, Booth ML. 1998.	Effects of physical activity interventions in youth. Review and synthesis.	Childhood Obesity	Am J Prev Med 15(4): 298–315.
Waters E, de Silva-Sanigorski A, Hall Belinda J, Brown T, Campbell Karen J, Gao Y, et al. 2011.	Interventions for preventing obesity in children.	Childhood Obesity	Cochrane Database of Systematic Reviews; doi: 10.1002/14651858.C D001871.pub3.
Carter JS, Pugh JA, Monterrosa A. 1996.	Non-insulin-dependent diabetes mellitus in minorities in the United States.	Diabetes	Ann Intern Med 125(3): 221–232.
Golden SH, Robinson KA, Saldanha I, Anton B, Ladenson PW. 2009.	Prevalence and incidence of endocrine and metabolic disorders in the united states: A comprehensive review.	Diabetes	Journal of Clinical Endocrinology and Metabolism 94(6): 1853–1878.
Hunt KJ, Schuller KL. 2007.	The Increasing Prevalence of Diabetes in Pregnancy.	Diabetes	Obstetrics and Gynecology Clinics of North America 34(2): 173–199.
Kim C, Newton KM, Knopp RH. 2002.	Gestational diabetes and the incidence of type 2 diabetes: a systematic review.	Diabetes	Diabetes Care 25(10): 1862–1868.
Li S, Shin HJ, Ding EL, van Dam RM. 2009.	Adiponectin levels and risk of type 2 diabetes: a systematic review and meta-analysis.	Diabetes	JAMA 302(2): 179–188.
Naqshbandi M, Harris SB, Esler JG, Antwi-Nsiah F. 2008.	Global complication rates of type 2 diabetes in Indigenous peoples: A comprehensive review.	Diabetes	Diabetes Res Clin Pract 82(1): 1–17.
Naylor JL, Schraer CD, Mayer AM, Lanier AP, Treat CA, Murphy NJ. 2003.	Diabetes among Alaska Natives: a review.	Diabetes	Int J Circumpolar Health 62(4): 363–387.
Porter C, Skinner T, Ellis I. 2012.	The current state of Indigenous and Aboriginal women with diabetes in pregnancy: A systematic review.	Diabetes	Diabetes Res Clin Pract.
Yu CH, Zinman B. 2007.	Type 2 diabetes and impaired glucose tolerance in aboriginal populations: a global perspective.	Diabetes	Diabetes Res Clin Pract 78(2): 159–170.
Dinos S, Khoshaba B, Ashby D, White PD, Nazroo J, Wessely S, et al. 2009.	A systematic review of chronic fatigue, its syndromes and ethnicity: prevalence, severity, co-morbidity and coping.	Fatigue	Int J Epidemiol 38(6): 1554–1570.
Colosia AD, Masaquel A, Hall CB, Barrett AM, Mahadevia PJ, Yogev R. 2012.	Residential crowding and severe respiratory syncytial virus disease among infants and young children: A systematic literature review.	Infection	BMC Infectious Diseases 12.
Goodman KJ, Jacobson K, van Zanten SV. 2008.	Helicobacter pylori infection in Canadian and related Arctic Aboriginal populations.	Infection	Canadian Journal of Gastroenterology 22(3): 289–295.
Hirsch-Moverman Y, Daftary A, Franks J, Colson PW. 2008.	Adherence to treatment for latent tuberculosis infection: systematic review of studies in the US and Canada.	Infection	Int J Tuberc Lung Dis 12(11): 1235–1254.
Nair H, Nokes DJ, Gessner BD, Dherani M, Madhi SA, Singleton RJ, et al. 2010.	Global burden of acute lower respiratory infections due to respiratory syncytial virus in young children: a systematic review and meta-analysis.	Infection	The Lancet 375(9725): 1545–1555.
Nair H, Brooks WA, Katz M, Roca A, Berkley JA, Madhi SA, et al. 2011.	Global burden of respiratory infections due to seasonal influenza in young children: A systematic review and meta-analysis.	Infection	The Lancet 378(9807): 1917–1930.
Orr P. 2011.	Adherence to tuberculosis care in Canadian Aboriginal populations, Part 1: definition, measurement, responsibility, barriers.	Infection	Int J Circumpolar Health 70(2): 113–127.
Siber GR, Chang I, Baker S, Fernsten P, O’Brien KL, Santosham M, et al. 2007.	Estimating the protective concentration of anti-pneumococcal capsular polysaccharide antibodies.	Infection (Vaccine)	Vaccine 25(19): 3816–3826.
Destefano F, Pfeifer D, Nohynek H. 2008.	Safety profile of pneumococcal conjugate vaccines: systematic review of pre- and post-licensure data.	Infection (Vaccines)	Bull World Health Organ 86(5): 373–380.
Pollack KM, Frattaroli S, Young JL, Dana-Sacco G, Gielen AC. 2012.	Motor vehicle deaths among American Indian and Alaska Native populations.	Injury	Epidemiol Rev 34(1): 73–88.
Jowett CR, Morcuende JA, Ramachandran M. 2011.	Management of congenital talipes equinovarus using the Ponseti method: A systematic review.	Orthopedics	Journal of Bone and Joint Surgery - Series B 93 B(9): 1160–1164.
Kelley GA, Kelley KS, Tran ZV. 2000.	Exercise and bone mineral density in men: a meta-analysis.	Orthopedics	J Appl Physiol 88(5): 1730–1736.
Bluestone CD. 2004.	Studies in otitis media: Children’s Hospital of Pittsburgh-University of Pittsburgh Progress Report - 2004.	Otolaryngology	Laryngoscope 114(11 III): 1–26.
Smith DF, Boss EF. 2010.	Racial/ethnic and socioeconomic disparities in the prevalence and treatment of otitis media in children in the United States.	Otolaryngology	Laryngoscope 120(11): 2306–2312.
Flores G, Vega LR. 1998.	Barriers to health care access for Latino children: a review.	Pediatrics	Fam Med 30(3): 196–205.
Flores G. 2010.	Racial and ethnic disparities in the health and health care of children.	Pediatrics	Pediatrics 125(4): e979–e1020.
Guo Y. 2005.	Filial therapy for children’s behavioral and emotional problems in mainland China.	Pediatrics	J Child Adolesc Psychiatr Nurs 18(4): 171–180.
Brach C, Fraser I. 2000.	Can cultural competency reduce racial and ethnic health disparities? A review and conceptual model.	Primary Care	Med Care Res Rev 57 Suppl 1: 181–217.
Fisher TL, Burnet DL, Huang ES, Chin MH, Cagney KA. 2007.	Cultural leverage: interventions using culture to narrow racial disparities in health care.	Primary Care	Med Care Res Rev 64(5 Suppl): 243S–282S.
Atkinson DR. 1983.	Ethnic Similarity in Counseling Psychology: A Review of Research.	Psychology	The Counseling Psychologist 11(3): 79–92.
Coleman HLK, Wampold BE, Casli SL. 1995.	Ethnic Minorities’ Ratings of Ethnically Similar and European American Counselors: A Meta-Analysis.	Psychology	Journal of Counseling Psychology 42(1): 55–64.
Dhalla S, Zumbo BD, Poole G. 2011.	A review of the psychometric properties of the CRAFFT instrument: 1999–2010.	Psychology	Curr Drug Abuse Rev 4(1): 57–64.
Fan JB, Sklar P. 2005.	Meta-analysis reveals association between serotonin transporter gene STin2 VNTR polymorphism and schizophrenia.	Psychology	Mol Psychiatry 10(10): 928–938, 891.
Gone JP, Alcantara C. 2007.	Identifying effective mental health interventions for American Indians and Alaska Natives: a review of the literature.	Psychology	Cultur Divers Ethnic Minor Psychol 13(4): 356–363.
Griner D, Smith TB. 2006.	Culturally adapted mental health intervention: A meta-analytic review.	Psychology	Psychotherapy (Chic) 43(4): 531–548.
Kim G, Decoster J, Huang CH, Chiriboga DA. 2011.	Race/ethnicity and the factor structure of the Center for Epidemiologic Studies Depression Scale: a meta-analysis.	Psychology	Cultur Divers Ethnic Minor Psychol 17(4): 381–396.
Oyserman D, Coon HM, Kemmelmeier M. 2002.	Rethinking individualism and collectivism: evaluation of theoretical assumptions and meta-analyses.	Psychology	Psychol Bull 128(1): 3–72.
Paradies Y. 2006.	A systematic review of empirical research on self-reported racism and health.	Psychology	Int J Epidemiol 35(4): 888–901.
Smith TB, Silva L. 2011.	Ethnic identity and personal well-being of people of color: a meta-analysis.	Psychology	J Couns Psychol 58(1): 42–60.
Stone EJ, Baranowski T, Sallis JF, Cutler JA. 1995 Supplement.	Review of Behavioral Research for Cardiopulmonary Health: Emphasis on Youth, Gender, and Ethnicity.	Psychology	Journal of Health Education 26(2): S9–S17.
Twenge JM, Crocker J. 2002.	Race and self-esteem: meta-analyses comparing whites, blacks, Hispanics, Asians, and American Indians and comment on Gray-Little and Hafdahl (2000).	Psychology	Psychol Bull 128(3): 371–408; discussion 409–320.
Devries KM, Free C, Jategaonkar N. 2007.	Factors associated with condom use among Aboriginal people: a systematic review.	Reproductive	Can J Public Health 98(1): 48–54.
Shah PS, Zao J, Al-Wassia H, Shah V, Knowledge Synthesis Group on Determinants of Preterm LBWB. 2011.	Pregnancy and neonatal outcomes of aboriginal women: a systematic review and meta-analysis.	Reproductive	Womens Health Issues 21(1): 28–39.
Wu X, Zhao L, Zhu H, He D, Tang W, Luo Y. 2012.	Association between the MTHFR C677T polymorphism and recurrent pregnancy loss: A meta-analysis.	Reproductive	Genetic Testing and Molecular Biomarkers 16(7): 806–811.
Bryant J, Bonevski B, Paul C, McElduff P, Attia J. 2011.	A systematic review and meta-analysis of the effectiveness of behavioural smoking cessation interventions in selected disadvantaged groups.	Tobacco	Addiction 106(9): 1568–1585.
Carson KV, Brinn MP, Labiszewski NA, Esterman AJ, Chang AB, Smith BJ. 2011.	Community interventions for preventing smoking in young people.	Tobacco	Cochrane Database Syst Rev(7): CD001291.
Carson KV, Brinn MP, Labiszewski NA, Peters M, Chang AB, Veale A, et al. 2012a.	Interventions for tobacco use prevention in Indigenous youth.	Tobacco	Cochrane Database Syst Rev 8: CD009325.
Carson KV, Brinn MP, Peters M, Veale A, Esterman AJ, Smith BJ. 2012b.	Interventions for smoking cessation in Indigenous populations.	Tobacco	Cochrane Database Syst Rev 1: CD009046.
Doolan DM, Froelicher ES. 2006.	Efficacy of smoking cessation intervention among special populations: review of the literature from 2000 to 2005.	Tobacco	Nurs Res 55(4 Suppl): S29–37.
Gould GS, McEwen A, Watters T, Clough AR, van der Zwan R. 2012.	Should anti-tobacco media messages be culturally targeted for Indigenous populations? A systematic review and narrative synthesis.	Tobacco	Tob Control.
Sherman EJ, Primack BA. 2009.	What works to prevent adolescent smoking? A systematic review of the National Cancer Institute’s Research-Tested Intervention Programs.	Tobacco	J Sch Health 79(9): 391–399.
Thomas Roger E, Perera R. 2006.	School-based programmes for preventing smoking.	Tobacco	Cochrane Database of Systematic Reviews; doi: 10.1002/14651858.C D001293.pub2.
Tyas SL, Pederson LL. 1998.	Psychosocial factors related to adolescent smoking: a critical review of the literature.	Tobacco	Tob Control 7(4): 409–420.
Tobler NS, Stratton HH. 1997.	Effectiveness of School-Based Drug Prevention Programs: A Meta-Analysis of the Research.	Substance Use	The Journal of Primary Prevention 18(1): 71–128.