A scoping review of recommendations in the English language on conducting research with trauma-exposed populations since publication of the Belmont report; thematic review of existing recommendations on research with trauma-exposed populations

Objective To identify recommendations for conducting public health research with trauma-exposed populations. Methods Researchers searched Embase, PubMed, Scopus, Web of Science, Open Grey, and Google Scholar for recommendations. Trauma that causes psychological impact was our exposure of interest and we excluded clinical articles on treating physical trauma. We reviewed titles and abstracts of 8,070 articles and full text of 300 articles. We analyzed recommendations with thematic analysis, generated questions from the existing pool of recommendations, and then summarized select gaps. Results We abstracted recommendations from 145 articles in five categories: community benefit, participant benefit, safety, researcher well-being, and recommendations for conduct of trauma research. Conclusions Gold standards to guide the conduct of trauma-informed public health research do not yet exist. The literature suggests participation in trauma research is not inherently harmful, and current recommendations concern using research to benefit communities and participants, protecting participants and researchers from harm, and improving professional practice. As public health researchers increasingly analyze trauma as a determinant of health, gold standards for the conduct of trauma-informed public health research would be appropriate and timely.

Introduction whether research participants are selected from the populations who are likely to benefit from results. This report was revolutionary when it was released, and the three main principles remain relevant today. However, when the Belmont Report was published, scant research had been conducted on participants' comfort discussing trauma in a research setting, or on what benefits participation in research on trauma may offer participants [23]. Research since the Belmont Report was published indicates that participants can safely discuss trauma in health care [24,25] and research settings [23, 26-31] without causing undue harm, and that participation in research that addresses trauma may be meaningful and beneficial for trauma-exposed participants [32][33][34]. In light of our expanded understanding of the pervasiveness and complexity of trauma, we wanted to examine what relevant guidelines and recommendations exist for public health researchers on conducting trauma research and research with traumaexposed individuals and communities.
This review identifies, describes, and summarizes recommendations for conducting public health research on trauma or with trauma-exposed populations. The key questions are: our 2018 search with a limited number of articles that may change our framework, a single reviewer screened each title/abstract. This resulted in 126 articles for (2020) full text review. Three reviewers ( �� , �� , and �� reviewed the full texts so that each was independently rated by two reviewers, or a third in cases where two reviewers disagreed. We included all articles of all study designs in both searches. Exclusion criteria were the same across both searches. Articles were excluded if they were not in English, did not focus on trauma or a trauma-exposed population, and did not include either guidelines or recommendations for researchers about how to conduct research on trauma or with trauma-exposed populations. From the final set of articles in our 2018 and 2020 searches (S1 File), we extracted information on population, study design or article/report type, type(s) of trauma experienced, setting, and guidelines/recommendations from the authors. Because we were not extracting quantitative data from the articles, we did not assess risk of bias in individual studies or across studies, and we did not use summary measures, as is common in reviews utilizing Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines [37]. We analyzed themes from the initial search (in 2018) using MaxQDA [38] to group the text of recommendations into categories based on what part of the research process they applied to, population, and setting. We first read extracted suggestions from ten articles and each co-author drafted potential codes in memos. We then met as a research team, developed a pilot codebook from the group's memos, and coded ten additional extracted suggestions. We collaboratively identified the following as themes from the recommendations and included them in our codebook (available upon request): Institutional Review Board guidance, training, community-informed research, study stages, data collection, researcher characteristics, referral services, protocols, participants, participant empowerment, participant distress, special populations, children, and secondary traumatization.
We resolved differences through discussion and reached consensus on a final codebook. Two separate members of the research team coded the remaining transcripts from the initial 2018 search to arrive at a final set of codes. Following coding, we wrote summaries of each code and used these to produce a narrative summary for the results. We then read across these themes and created tables summarizing recommendations for the following research stages: research design, recruitment, consent, data collection, and results dissemination (available upon request). Because several recommendations cut across research stages and we noticed emergent themes in the set of extracted guidelines, we did a second round of classification sorting the recommendations into the following categories: community benefit, participant benefit, safety, researcher well-being, and the nature and scope of trauma research. Several articles offered recommendations for research with specific populations, and we translated these into broader suggestions where possible.
When we created tables of the recommendations organized by research stage from our 2018 search, each author also produced a list of questions that the recommendations raised for them. One author ( �� ) then identified themes within the set of questions and consolidated these lists, dropping questions that are already widely discussed within the broader research literature (e.g., questions pertaining to community based participatory research (CBPR)). He grouped the retained questions into the following emergent themes on gaps specifically relevant to the conduct of trauma-informed research: research design and conduct, systemic issues, and research translation/applicability.
Because publication themes in our 2020 search were similar to article themes in our 2018 search, we did not code articles in our 2020 search in MaxQDA [38]. We directly sorted them into categories of community benefit, participant benefit, safety, researcher well-being, and the nature and scope of trauma research, as well as research design, recruitment, consent, data collection, and results dissemination. We did not generate questions in response to recommendations in the articles within our 2020 search as these articles did not offer substantially different recommendations than those identified in our earlier search.

Results
After removing duplicate articles using the automated processes in EndNote and Covidence and manually removing additional duplicates, we had 8,075 articles (6,234 in 2018 and 1,835 in 2020). After screening titles and abstracts, we had 300 articles for full text review (174 in 2018 and 126 in 2020). Co-authors conducting the full text review excluded 155 articles (96 in 2018 and 59 in 2020) for the following main reasons: Primary purpose of the article not to present guidelines or recommendations (129 articles), Only Clinical Care (19 articles), Book/ Book Chapter (9 pieces), Not Related to Trauma or Traumatized populations (4 articles), Language other than English (3 articles). Thus, we used abstracted data from 145 articles in this review (Fig 1).
The reviewed articles (S1 File) were heterogenous and consisted of a mix of non-research articles (46 articles), qualitative research (37 articles), review pieces (18 articles), mixed methods studies (15 articles), experimental studies (13 articles), cross-sectional studies (5 articles), and observational studies (1 article). The articles studied or commented on the impacts of trauma or trauma research ethics among diverse populations including, children (20 articles), service providers or researchers (19 articles), survivors of other forms of violence or abuse (11 articles), Indigenous people (9 articles), survivors of sexual assault (9 articles), women (9 articles), people with disabilities (7 articles), populations described by sample design (e.g., probability sample) (7 articles), people with unspecified trauma (6 articles), students (6 articles), gender and sexual minorities (4 articles), immigrants and refugees (4 articles), elders (3 articles), incarcerated individuals (2 articles), individuals affected by COVID-19 (2 articles), individuals in 'the Global South' (2 article), New York City residents after 9-11 (2 articles), survivors of intimate partner violence (2 articles), autistic people (1 article), emergency room patients (1 article), individuals who may be both survivors and perpetrators of violence (1 article), individuals affected by natural disasters (1 article), internet users (1 article), people using methamphetamines (1 article), residents in a rural community (1 article), survivors of trafficking (1 article), survivors of war (1 article), and veterans (1 article). The remainder of our Results sections address our findings by each of our two research questions.
Research Question 1: What are existing recommendations for conducting research with populations who have experienced trauma? (Table 1) Many of the recommendations we extracted pertain to more than one research stage, and each research stage has multiple recommendations pertaining to each category (S1 File). As a result, we chose to present narrative summaries of the recommendations pertaining to community benefit, participant benefit, safety, researcher well-being, and the nature and scope of trauma research in our Results section. Our full list of reviewed articles (S1 File) also organizes articles by community benefit, participant benefit, safety, researcher well-being, and the nature and scope of trauma research.
Community benefit. Trauma harms not only individuals who experience it, but also the communities that they are members of and/or constitute, and subsequent generations [140,141]. This review resulted in many recommendations that researchers support community well-being or empowerment throughout the research process, and many articles explicitly addressed systematic inequity [54, 57, 61, 63-65]. Most recommendations endorsed CBPR principles either explicitly or implicitly [142], including considering strengths and protective factors in community rather than just deficits [52, 56, 57, 61, 63, 64]. Recommendations also supported representation in research (as participants, during study design and ethical review, and during dissemination) of marginalized and trauma-exposed populations in order to ensure research equity

PLOS ONE
Thematic review of existing recommendations on research with trauma-exposed populations Table 1. Select recommendations for research with trauma-exposed populations by research stage, in 2018 and 2020 review conducted in the United States on research with trauma-exposed populations.

Community Benefit
Focus research on change and use community-based participatory research methods: promote participant agency; do not exclude people with trauma; support empowerment; address stigma, historical trauma, and systematic injustice; involve community partners and participants in oversight of your research from inception to completion-including data interpretation and dissemination, authoring publications, and ownership of the data-; build lasting relationships with communities; create tangible products that benefit the community over the course of your project; embed research findings in historical and cultural context for accurate interpretation; extend confidentiality to the whole community; and consider policy X X X X When advertising your study account for the possibility that some people may not remember or identify their trauma as abuse; develop protocols for trying to prevent such participants from ending up in your control group so as to not affect your results (e.g., ask about specific experiences and screen for trauma symptoms). Know that retrospective and prospective measures of trauma may measure different constructs and consider that people with trauma may differ from people without trauma in terms of pre-existing vulnerabilities. Also consider that some participants may be survivors and perpetrators of trauma. When screening participants do not ask them to self-identify or label their experiences, instead ask they about specific experiences they may have had and be aware of any mandating reporting standards Seek community and public involvement in research and investigate practical and ethical challenges in this research. Researchers who belong to populations that experience marginalization may seek to affect change through their research, or by providing bridging social capital between researchers and wider communities. Bonding social capital is required to do this. Researchers who are complete outsiders to a community should seek to partner with researchers who are also community insiders. Intersectionality [129], reflexivity, and community based participatory research are helpful tools [130][131][132][133][134][135][136][137][138][139]. Safety. Trauma fundamentally threatens the physical and/or psychological safety of those who experience it [143]. Many potential participants in trauma research may experience ongoing threats to their safety, and trauma researchers are often concerned about the possibility that discussing trauma may cause participants distress. Safety recommendations addressed ways to assess and minimize the possibility of harm due to research participation that could make participants more vulnerable to violence, or data collection that may cause distress. Articles also considered participant safety as a function of systemic exposures and social identity (e.g., incarceration or immigration status creating pressure to participate, or sexual orientation rendering parental consent for minors to participate in research dangerous) [82-84, 88, 89, 146] Goodwin & Tiderington (2020) present thoughtful trauma-informed recommendations on how to conduct research that take into account participants' social identities and systemic exposures [87]. Mwambari (2019) provides a particularly thick and rich description of the risks that participants and local researchers can incur from working with non-local researchers, as well as recommendations to lessen these risks [84]. For a more detailed discussion on reducing participants' vulnerability to violence or distress from research participation please see: Ahlin (2019)  Researcher well-being. When studying trauma, researchers may experience participants' trauma vicariously, which could cause distress or burn-out [31,104]. Several authors offered recommendations to minimize distress and promote researcher well-being through self-care and addressing secondary traumatization. These recommendations included pacing one's workload [103,108], cultivating one's support network [31,97,103,105,107], connecting with other trauma researchers [105,113,148,149], and preventing, recognizing, and responding to emotional distress experienced by researchers as a result of conducting research [31, 42, 72, 92, 102-107, 111] Some authors considered researcher well-being not just the responsibility of the researcher, but the responsibility of universities and Institutional Review Boards (IRBs) as well [105,106,111,112,114]. Recommendations to minimize distress and promote researcher well-being through environmental supports included trauma-informed supervision, de-briefing, and access to counseling [72, 102-108, 111, 112]. For more detailed discussions of secondary traumatization, self-care, and considerations for researchers with a trauma background themselves, please see Eadesa et al. (2020), [104] Fohring (2020 [105], Jeftic (2020) [113], Markowitz (2019) [106], and Wager (2011) [31]. Researchers interviewing perpetrators may want to read Jeftic (2020) [113] and Markowitz (2019) [106].
Nature and scope of trauma research. Articles on the scope and nature of trauma research concerned particular trauma topics, risk/benefits to participation, methodology, and application of theory to trauma research. Particular trauma topics included interventions [125], disclosure [124], cognitive distortions [150], and interviewing children [151] [153][154][155][156], a need to use consistent measures and not conflate different constructs [123,127,128], a need to see participants as potentially being both survivors and perpetrators of violence [122], and the nocebo effect in trauma research [119]. Researchers offered recommendations on improving internal research validity when potential control group members with trauma do not identify themselves as having trauma [31,121], and many recommended using participatory research methods to conduct better research [132,135,137,138,155]. Gultekin et al. (2019) [133] introduced the eco-social trauma intervention model, which may be used across disciplines to develop and pilot trauma-informed interventions, and Tol (2020) [139] described using a social justice framework for research. A few articles concerned research and insider-outsider dynamics [130,131,136]. For example, Carr (2019) [130] describes situated solidarity as a possible approach for academics who also belong to communities that the focus of their research proposes to benefit, and Gaillard & Peek (2019) [131] recommend a structure for outsider and local researchers to work together in the aftermath of disaster.

Research Question 2: What gaps exist for public health practitioners to be able to plan and implement trauma-informed research studies?
Following our review of existing recommendations, we identified a select number of gaps specific to the literature on trauma research that, should they be addressed, may advance public health research on trauma or with trauma-exposed populations. The gaps we identify are far from exhaustive, but we hope they may catalyze further conversation and future scholarship. We acknowledge that work is being done to address the gaps we identify, and we hope to see this work continue. We comment briefly on three themes pertaining to gaps below (research design and conduct, systemic issues, and research translation/applicability) and provide more detailed explanation and a select number of questions to consider in (S1 File).
Research design and conduct. We identified a need for additional scholarship on: 1) how to apply theory in the design of trauma-relevant research; and 2) how research populations should be defined and investigated in trauma research. The first is a need because effective use of theory in research facilitates knowledge building and communication between researchers [157], and can also inform intervention design and testing [158]. Without additional guidance on how to apply theory in trauma-relevant research, scholars may utilize theories in a fragmented manner that does not support knowledge unification. The second is a need because insufficient consensus on who is traumatized and how to define their exposure limits both epidemiological and intervention research. For example, people of color and other marginalized groups are under-represented in psychological trauma intervention research [159][160][161] which limits the generalizability and applicability of psychological trauma treatment research [159,160,162].
Systemic issues. We identified a need for research on psychological trauma that: 1) utilizes multi-level frameworks [163,164]; 2) conceives and attends to multiple dimensions of trauma simultaneously (e.g., historical, cultural, racial, and systemic trauma, plus individuallyoriented conceptions of trauma); and 3) focuses on the role that systemic injustices (e.g., institutionalized racism and colonialism) and abuse of institutional power (e.g., state-sponsored torture or terror) have in shaping trauma exposures and responses. These are gaps because trauma exposure, immediate harm done by trauma, and societal responses to trauma are largely determined within complex social systems [17]; as such, research that does not account for multiple social ecological exposures, multiple dimensions of trauma, or societal power dynamics, may be ineffective at addressing trauma epidemiology and interventions. They may even promote distorted perceptions of and inappropriate responses to trauma.
Research translation/applicability. We identified two pressing needs for scholarship pertaining to research translation or applicability: 1) literature should address how to translate findings to wider -and multi-ethnic- [160] audiences and how to scale up evidence-based or promising interventions; and 2) researchers need guidance on how to design or use research to effectively impact policy. The first is a need because the lag between trauma-relevant research and public knowledge regarding trauma creates treatment barriers for trauma survivors and promotes public disinvestment in trauma prevention and treatment research [165,166]. The second is a need because policy is an important determinant of health [167] (that many sectors are presently seeking to leverage to prevent and treat trauma), but health researchers are not typically trained to design studies with immediate relevance to informing policy or stakeholders working within cross-sectorial partnerships to influence policy.

Discussion
Although a robust body of public health research on trauma and with trauma-exposed people and communities exists, gold standards for the conduct of trauma-informed public health research have not yet been authored. Many researchers draw upon principles described in the Belmont Report to inform the ethical conduct of their research. While the Belmont Report offers guiding principles for the ethical conduct of research that remain relevant today, its principles are necessarily part of a larger living standard of research ethics that should be evolving over time. Since the Belmont Report was published, the Council for International Organizations of Medical Sciences published "International Guidelines For Ethical Review of Epidemiological Studies" [168] in 1991, and updated "International Ethical Guidelines for Epidemiological Studies" [169] in 2009. The World Health Organization (WHO) has also published "Putting Women First: Ethical and Safety Recommendations for Research on Domestic Violence Against Women" [170] in 2001, "WHO Ethical and Safety Recommendations for Researching, Documenting and Monitoring Sexual Violence in Emergencies" [171] in 2007, and "Ethical and Safety Recommendations for Intervention Research on Violence against Women" [172] in 2016. While these guidelines do elaborate on the ethical conduct of research with vulnerable populations, and do consider the importance of access to research participation for diverse groups, gaps remain in the ethics of inclusion in research [160,173] and they do not provide guidance on the conduct of trauma-informed research that may be applied across populations [169]. CBPR has also emerged as an ethical approach to research since the publication of the Belmont report, and offers an extensive set of guidelines for working with communities that may inform research with trauma-exposed populations or on systemic, cultural, racial, or historical trauma. However, studies may be able to be traumainformed without implementing full CBPR standards, which can be resource-intensive and time-consuming.
Despite the current lack of gold standards for trauma-informed research, this review identified recommendations in the literature for trauma research or research with trauma-exposed populations. These recommendations include using research for change (e.g., advocacy and de-stigmatization) [ 115-119, 121-123, 125, 127, 128, 131, 133, 139, 146, 151, 154, 177]. We also identified needs for guidance and continued discussion pertaining to the design and conduct of research (e.g., how to use theory and how to define and investigate research populations), systemic issues (e.g., how to use multi-level frameworks and multi-dimensional understandings of trauma, and how to address systemic injustice), and research translation and applicability (e.g., how to translate findings widely and scale up evidence-based interventions, and how to impact policy.).
This review, while contributing to literature, still has several limitations. First, though we used current literature and worked closely with a health sciences research librarian to define our search string, there could be important terms we missed. For example, we did not include "complex trauma" or "developmental trauma" in our search string, and we did not search the PsychINFO database. In addition, due to the large number of references reviewed, we did not add articles to our review from the reference lists of the articles we found through our search string. Additionally, due to our restriction to English-language articles, many articles we reviewed presented research based in the United States, and further research on trauma globally is necessary. Finally, we completed a thorough qualitative analysis on the final set of articles, but additional coders or a more extensive reconciliation process mays have further refined our findings.
This review also incorporates several strengths. First, by using a systematic, replicable search strategy, future researchers can recreate or easily expand on our methodology. Second, we attempted to incorporate non-peer reviewed and grey literature sources by including two grey literature databases in our search. Third, our research team draws on a diverse set of content and professional expertise (immigrant health, substance use, historic and systemic trauma, intimate partner violence, medicine, nursing), allowing for incorporation of distinct perspectives into the search string, strategy, and qualitative analysis.
It is not the goal of this review to create guidelines; however, developing specific and universal trauma-informed research guidelines that could be implemented and tested would improve research quality and support the mentoring of early career researchers. Researchers and community partners could use gold standards that provide a customizable "menu" of guidelines to inform protocol development, data analysis, dissemination, and application of results. This article highlights the need for such guidelines to be authored. Given public health's history of unethical research (e.g. Tuskegee Syphilis Trials) [178] that prompted publication of the Belmont Report, combined with the potential public health research has to be a vehicle of community and individual empowerment, we contend that individuals and communities affected by trauma must be involved in authoring a set of gold standard guidelines. We also contend that such guidelines must provide actionable and evaluated steps to increase the participation of people of color and other marginalized populations in research, in order to be relevant for and accountable to a majority of trauma-exposed populations [160]. We comment that gaps pertaining to research design and conduct, systemic issues, and research translation/ applicability require additional inquiry. These gaps are not exhaustive and scholarship is addressing them. Nonetheless, additional conversations about gaps and how to address them undoubtedly are needed, and those conversations must involve both researchers across disciplines and those affected by trauma. Institutes that are critical to the conduct of public health research, such as the National Institutes for Health, could gather individuals and communities affected by trauma to address gaps in a more comprehensive manner, and to author guidelines for the conduct of trauma-informed public health research. Given that public health researchers are increasingly embracing and researching trauma as a determinant of health, such guidelines and focus on gaps would be appropriate and timely.