Applying best practices for health disparities work to create a treatment adaptation framework for culturally diverse communities: A mixed-methods approach

Abstract

Objective:

Integrating best practices for health disparities to adapt evidence-based treatments is imperative to adequately meet the needs of diverse cultures, particularly ones that therapists can apply flexibility across multiple diverse communities.

Method:

Using a mixed-methods, community-engaged approach, we examined how a range of community participants (N = 169) defined mental health, perceived barriers to treatment, and used culturally-based coping methods to manage their mental health. Phase 1 (n=49) included qualitative focus group data from five distinct racial/ethnic communities (African immigrants/refugees, Black/African Americans, Hispanics, Pacific Islanders, and American Indians). Phase 2 included quantitative surveys from members of four of these communities (n=59) and the front-line providers serving them (n=61).

Results:

The communities and providers highlighted chronic worry and distress related to daily activities as primary treatment concerns. Further, this mixed-methods data informed our proposed best practice treatment adaptation framework using chronic worry as an example.

Conclusion:

The main aims of this study were to exemplify best practices for addressing mental health inequities in communities of color in terms of (1) conducting health disparities research and (2) applying a treatment adaptation framework for culturally-responsive clinical care. Specific features of how this framework was conceived and applied provide a unique and critical view into integrating best practices to address health disparities in diverse communities.

Racial and ethnic cultural differences are consistently reported in the prevalence, diagnosis, treatment response, and outcomes among individuals with mental health concerns. For instance, despite a demonstrated need for treatment, people of color (POC) have consistently reported lower service utilization rates for mental healthcare than national averages; have less access to mental health services; receive poorer quality of care; are less satisfied with mental health services and have higher rates of dropout when compared to White individuals (e.g., Carpenter-Song et al., 2011; Maura & Weisman de Mamani, 2017). Factors exacerbating mental health disparities among POC include cultural barriers (e.g., stigma and norms about mental health) and structural barriers (e.g., cost of services, inequitable insurance, accessibility, and language biases inherent in existing healthcare systems; Derr, 2016). Gaps in knowledge about mental health symptoms and services further compound these disparities.

One of the efforts aimed at reducing mental health disparities and understanding best treatment practices has been the creation of numerous culturally-adapted (CA) frameworks to better meet the needs of culturally-diverse communities, such as the ecological validity framework (Bronfenbrenner, 1989), cultural sensitivity framework (Resnicow et al., 2002), cultural adaptation process model (Domenech-Rodríguez & Wieling, 2004), and the formative method for adapting psychotherapy (FMAP; Hwang, 2009). This list of CA frameworks is not exhaustive, and there have also been several proposed frameworks for best treatment practices for specific communities, such as Latino/a individuals with schizophrenia (López et al., 2002), Vietnamese refugees (Hinton et al., 2004), and Latinos with depression (Interián & Díaz-Martínez, 2007). Such best treatment practices have included delivering evidence-based practice (EBP) in clients’ native languages, using phrases or images from the client’s culture, and modifying interventions to align with clients’ conceptualization of mental health (e.g., the mind and body connection; Lee-Tauler et al., 2018). Further, previous research shows promising evidence that CA interventions are efficacious, though there have been several methodological concerns (e.g., power, sample bias, sample size; Hall et al., 2016; Van Loon et al., 2013).

Nonetheless, despite the advancement of various CA frameworks and guidelines, traditional non-participatory research methodologies have often struggled to translate research into actionable steps to reduce health disparities for POC. For instance, Chu and Leino (2017) conducted a systematic review to develop a new, data-driven CA framework that distinguishes between core (facets that cause symptom change) and peripheral (facets that increase feasibility and acceptability of the intervention) aspects that therapists can adapt in psychotherapy. Although this demarcation is helpful for understanding mechanisms of change in therapy, these characteristics may not be particularly useful to community therapists who carry large caseloads and provide treatment to clients with complex presentations, including poverty, clinical severity, and comorbidity (Rodriguez et al., 2021). Hence, community-based participatory research (CBPR) has emerged in the last decade as a transformative research paradigm to bridge the gap between science and practice through community engagement and social justice to address health disparities (Wallerstein & Duran, 2017). A CBPR approach may be instrumental when adapting treatment for ethnic minority or underserved populations. Primarily as CBPR seeks to bring together researchers and communities by equitably involving community partners throughout the trajectory of a research project. By including community partners in the adaptation process, CBPR has the potential to ensure that interventions are adapted to be culturally acceptable for the community in which clinicians will implement the intervention. Specifically, adapting interventions in partnership with community stakeholders enhances cultural sensitivity, helps to ensure communities’ authentic lived experiences are reflected, and facilitates sustainability (McElfish et al., 2019; Wallerstein et al., 2020). Thus, CBPR succeeds by expanding the reach of translational sciences into influencing practices and policies for eliminating social and health inequities (O’Mara-Eves et al., 2015; Wallerstein & Duran, 2017), underscoring the utility of a CBPR orientation in addressing health disparities for ethnic minorities (McElfish et al., 2019).

Moreover, several investigators have discussed best practices (i.e., the most accurate way to identify, collect, evaluate, disseminate, and implement scientific inquiry; Perleth et al., 2000) in conducting CBPR with communities of color (McElfish et al., 2019; Wallerstein et al., 2020). Specifically, they have emphasized (a) establishing strong relationships between community leaders and researchers; (b) engaging community partners through multiple avenues (e.g., as community co-investigators, research staff, or community advisory board members); (c) outlining roles of community leaders throughout the study; and (d) training community partners in research methods for active involvement in the project. Further, best practices for engaging in CBPR also include using participants’ native language and cultural perspectives, conducting meetings in an easily accessible location or format for community partners, honoring cultural practices, and spending time in the community to build trust and understand their values, customs, concerns, and preferred means of addressing issues.

Despite this guidance, adequate research representation, engagement of multicultural populations, humility in research and practice, and cultural centeredness continue to lag in mainstream psychology (Rodríguez-Espinosa & Verney, 2021). Further, although the field of psychology has integrated a social justice lens when establishing research objectives and outcomes, the incorporation of CBPR principles of involving community partners at every stage of the research project has primarily remained at the establishment of study objectives and recruitment level (Rodríguez-Espinosa & Verney, 2021). For instance, De Las Nueces et al. (2012) noted that most studies using community-engaged research for mental health discussed community involvement in identifying study objectives, recruitment, developing and delivering interventions, and data collection. These efforts increased retention of community participants and effectiveness of interventions in improving behavioral- and health-related outcomes among POC. However, relatively few teams mentioned involving community partners in interpreting research findings, the manuscript preparation process, or other dissemination efforts. Thus, there is a gap in the current literature of CBPR principles of involving community partners at every stage of mental health research or in creating subsequent adaptation frameworks that provide actionable guidelines to community clinicians working with POC.

Therefore, the present study utilized a CBPR conceptual model (see Wallerstein et al., 2020) and the formative method of adapting psychotherapy (FMAP; Hwang, 2009) theoretical model to explore the mental health needs and treatment-related priorities of multiple racial/ethnic communities of color in a major city in the Mountain West region of the United States. Specifically, we used a community-based and bottom-up approach for applying best practices for health disparities work to create a treatment adaptation framework. Hence, we utilized the FMAP to guide our process due to its alignment with CBPR principles and previous work demonstrating the FMAP’s ability to ensure cultural acceptability (Hwang et al., 2015; Mkenda et al., 2017). The FMAP outlines five main phases: (1) generating knowledge and collaborating with stakeholders, (2) integrating generated information with theory and empirical and clinical knowledge, (3) reviewing the initial culturally adapted clinical intervention with stakeholders and revising the culturally adapted intervention, (4) testing the culturally adapted intervention, and (5) synthesizing stakeholder feedback and finalizing the culturally adapted intervention.

The research team conceived and designed the current study closely with Community Faces of Utah (CFU). CFU is a long-term partnership between the Utah Department of Health, the University of Utah, and five diverse community organizations: Best of Africa (BoA; African immigrants and refugees), Calvary Baptist Church (CBC; primarily Black/African Americans), Hispanic Health Care Task Force (HHCTF; Hispanic/Latino/a), National Tongan American Society (NTAS; Pacific Islanders), and Urban Indian Center of Salt Lake (UIC; American Indians/Alaska Natives). Thus, rather than focusing on only one community, our community-academic research team sought to find consistency and differences across and within all CFU communities using qualitative (constructivist grounded theory) and quantitative approaches. A mixed-methods approach is considered best-practice methodology when using CBPR principles (McElfish et al., 2019; Wallerstein et al., 2020). The partnership between CFU and the research team developed when CFU identified mental health as an important, unaddressed issue in their communities. The Community Collaboration Engagement Team at the University of Utah then contacted senior author A.A. as a potential collaborator with CFU and facilitated the initiation of this partnership. Extensive discussion on the foundation of the collaboration and essential elements for ongoing culturally engaged clinical work and research are detailed in Asnaani et al. (in press).

Due to the exploratory nature of our study, we did not approach the project with established hypotheses. Instead, we aimed to demonstrate application of best practices for addressing persisting mental health inequities in communities of color by (1) conducting health disparities research with a CBPR orientation and (2) applying a framework that can guide the adaptation of evidence-based treatments (EBTs) for culturally-responsive clinical care, stemming from this research in the previous aim. To achieve Aim 1, we intended to (a) conduct focus groups (i.e., Phase 1) to understand how each community broadly defines mental health, identify specific mental health priorities across communities, and understand how these mental health needs are currently being managed or treated in each community; and (b) jointly create and implement a short quantitative survey (i.e., Phase 2) on remaining questions from the focus group data. For Aim 2, we aimed to use a synthesis of data from Phases 1 and 2, utilize interpretations and reflections from our community partners, and establish theoretical guidelines for developing culturally-responsive treatments (namely, FMAP), with a firm footing in CBPR and social justice principles, to generate an overall treatment adaptation framework. More importantly, CBPR offers psychologists a unique opportunity to enhance patient-centered research and care, increase the effectiveness and sustainability of interventions, boost the real-world generalizability of research findings, and close the research-practice gap. Hence, we hoped that using such a bottom-up mixed-methods approach that clinicians and researchers could utilize across communities would ensure that communities of color are not siloed, which only impedes broad-scope efforts to appreciably address far-reaching health disparities. As a whole, this study aimed to demonstrate application of best practices for health disparities work in two main domains: health disparities research and culturally competent practice/clinical care via the creation of a treatment adaptation framework.

Phase 1 Qualitative Data

Method

For Phase 1, we conducted a series of focus groups to explore the mental health experiences of all the CFU communities by understanding how community members conceptualize mental health, their priority mental health concerns, their barriers to seeking help, and, most importantly, their mental health treatment preferences. In line with best practices and CBPR principles, the research team and CFU leadership (comprised of community health workers, advocates, church administrators, and academics) collaboratively decided that Phase 1 would involve community-specific focus groups with members from each CFU community.

Participants

Participants in Phase 1, residing in a major urban area in the Mountain West region of the U.S., were selected via a purposive sampling approach (Boehnke et al., 2010), which the CFU leaders recommended due to success in previous CFU-research collaborations. Specifically, in line with best practices, community leaders could contact their community members via the most culturally-appropriate method (e.g., word of mouth, calling members, announcing at community meetings). Additionally, as requested, community leaders verified that they recruited members who were open to discussing issues regarding mental health, were from a range of ages, genders, and socioeconomic/occupational backgrounds, were not all from the same family unit, or individuals known to have mental health distress (although at least one-two participants in each focus group voluntarily shared a mental health diagnosis) in order to capture a representative sample of their communities. A total of 49 community members participated in the focus groups. Specifically, nine members were from the African immigrant/refugee (BoA) community, 10 from the primarily Black/African-American (CBC) community, 11 from the Hispanic/Latino/a (HHCTF) community, nine from the Pacific Islander (NTAS) community, and 10 from the American Indian/Alaska Native (UIC) community. Of these 49 participants, 31 identified as female and 18 identified as male. The age range of all participants was 19 to 72 years (M = 41.4 years, SD = 15.7 years).

Procedure

The University of Utah Institutional Review Board (IRB) approved the present study. Six focus groups (one for each CFU community, except for the Latino/a community, who requested two smaller groups) were conducted between July and August 2020 with 8–10 community members from each community. In line with CBPR principles and best practices, community leaders who received training in conducting focus groups and utilized a specifically-developed discussion guide (see Figure 1 in Supplementary Materials (Supp)) led the focus groups. Specifically, two trainings via Zoom were provided to the CFU-research team covering effective recruitment methods in diverse communities, foundations of clinical interviewing, qualitative data methods, and guidance from community leaders on how to obtain open-ended feedback from community participants in culturally-effective ways. Each community leader moderated the focus group for their respective community (authors J.V., E.E.N., F.T-P., V.M., & D.L.). One to two academic team members were present for all focus groups to observe, take notes, and ask follow-up questions for responses that required clarification. Thus, to avoid cultural immersion bias, the academic team, as outsiders to these communities, were present to ensure follow-up inquiry occurred to any assumed beliefs or statements (Napier et al., 2014). Due to COVID-19 restrictions, while we were planning these focus groups, the CFU-research team decided to conduct these focus groups remotely in an online, live format via password-enabled Zoom sessions. All participants provided verbal informed consent for the two-hour, audio-recorded focus groups, which an external service transcribed, with safeguards to protect participants’ identities. Participants and interpreters received USD $100 for their participation. Extensive details on the methodology, including details on the research trainings, are described in Asnaani et al., in press.

Focus group questions and handouts were translated into Spanish and Tongan using community-based translators (and re-checked for accuracy by CFU leaders). The BoA focus group was conducted in English with a live Kirundi interpreter. The HHCTF focus groups were conducted in Spanish by two of their community leaders (authors A.S.B. and J.V.), with live interpretation in English for the research team provided by a trained interpreter. After introductions and a brief description about navigating Zoom, the community leader introduced the topic of the focus group then opened the discussion with a question about what mental or emotional health means to the community members or their families. Facilitators used the topic guide to explore priority areas for mental health intervention, impact of sociopolitical determinants as barriers to mental health treatment, current culturally-congruent efforts to manage emotional health, and the positive or negative impact of COVID-19 on emotional health.

After the focus groups, we provided participants with handouts describing common mental health symptoms that CFU leaders previously identified as problematic in their communities (e.g., anxiety, stress, depression, post-trauma reactions, and self-harm/suicide) and common barriers to mental health treatment reported in the extant literature (e.g., poverty, discrimination, stigma, immigration/acculturative stress, isolation/loneliness). Due to focus group participants’ feedback about lack of knowledge about community resources for mental health, the research team also distributed lists of existing community emotional health resources.

Data Analysis

A qualitative, constructivist grounded theory (CGT) approach was utilized as the Phase 1 methodology. This approach allows for new theoretical insights to be developed directly from a systematic collection and analysis of data and is particularly useful in areas where existing theory is incomplete, unsuitable, or absent (Charmaz, 2014). In line with best practices and CBPR principles, data were generated through focus groups transcripts to enable participants’ conceptualizations and reported experiences of mental health. In CGT, concepts and theory are established through constant comparison, where we reviewed, compared, and contrasted codes within and across communities (Charmaz, 2014). This process generated further questions on mental health priority areas and intervention needs, which we then tested with further data collection (i.e., Phase 2). All analyses were performed in Atlas.ti version 9.

Data analysis was divided into stages – open, focused, and theoretical coding (Charmaz, 2014). The research team (RT; consisting of first author K.K. as the lead coder and three trained coders) first individually read each community’s focus group transcript and then conducted the initial analysis via line-by-line coding, which helped explain and understand conceptual reoccurrences and patterns in the data. In line with grounded theory studies, the RT assigned labels or codes to units of meaning within the data (e.g., words, phrases, sentences) in a process known as open coding. In line with Charmaz’s method, the RT utilized gerunds (i.e., verbs ending in “ing”) to capture processes and attempted to keep codes as similar to the data as possible (e.g., using communities’ terminologies). Initial coding was guided by the six questions posed in the focus groups and thus labeled according to their relevance to the research objective. The RT developed coding systems individually and through weekly team meetings and discussions over the 6-month data analysis period. Through initial coding, we began to look for similarities and differences between codes, and asked ourselves what these codes suggested or assumed, as well as what points of view were driving these codes, and what they represented overall. Additionally, we created memos for each community and discussed any issues or concerns experienced in the open coding phase. We used these memos to document our reflection about how and when processes occurred, how they evolved, and their consequences. After inductive thematic saturation was achieved, that is, until no new codes could be developed, there was a total of 1,576 initial codes generated across all communities (Urquhart, 2013).

Following initial coding, the RT refined their open codes by identifying recurring themes within communities and collapsing similar codes into one code, which was then used to assess which codes appeared to be the most significant (Charmaz, 2014). For instance, we categorized similar codes within and across communities and found commonalities and dissimilarities among categories. We did the constant comparison process individually and then during weekly meetings to assess for any biases or misinterpretations. This synthesis led to the development of new focused codes used to incorporate larger data segments. The process elevated the level of conceptual analysis so that broader categories could be developed. The focused coding reduced the initial codes down to a total of 845. Throughout the analytical process, the RT continued to compare codes, develop themes within and across communities, and wrote memos of their interpretations of the underlying codes and themes.

For theoretical coding, RT explored the relations between codes by connecting them and analyzing overarching categories. We fleshed out each primary focused code, analyzing the situations in which they appeared, when they changed, and the relations among them. For instance, we had several line-by-line focused codes of descriptions of trauma experienced by community participants. We then categorized these codes of trauma into types of trauma (e.g., childhood abuse, community violence, natural disasters, intimate partner violence, refugee trauma, discrimination and racism), which we then further categorized into primary or secondary trauma (i.e., hearing about another individual’s first-hand trauma), and then created the higher-order category of experiencing trauma. This iterative process of creating subcategories and overarching categories reduced the focused codes down to 152, which were further reduced during group meetings to 110 for all communities. After the categories were well refined, themes were developed using the categories as a framework. The RT did this for each community and identified similarities and differences across communities. Specifically, we looked for causal conditions (i.e., what impacts their mental health/how do they perceive mental health), the phenomena (i.e., their mental health symptoms), strategies (i.e., their culturally relevant coping strategies), context (i.e., the sociopolitical determinants), intervening conditions (i.e., what facilitates or constrains the strategies community members use to manage their mental health), and consequences (i.e., what tailoring/adaptations would community members like to see from clinicians and their communities). At the end of our qualitative analysis, we developed a tentative treatment adaptation model for each community, expressed as a set of cohesively related concepts. In line with best practices for CBPR and the FMAP theoretical model (specifically Phase 1), we presented the results of the qualitative analysis to the community leaders to assess accuracy and agreement.

Results

Altogether, every sample group from the community defined mental health as a balance between emotional, psychological, and social well-being, with three of the communities (HHCTF, NTAS, and UIC) also including spiritual well-being. Additionally, it is crucial to highlight that the communities chose to use the word emotional health instead of mental health, with the NTAS community sample group emphasizing that the Diagnostic and Statistical Manual of Mental Disorders (DSM-5) definitions of mental health should be avoided. Moreover, the top emotional health concerns across four communities (CBC, HHCTF, NTAS, and UIC) were anxiety, depression, and substance use problems. Two communities (HHCTF and UIC) also highlighted suicidal ideation as a top concern. The African immigrant/refugees and Black/African-American communities also emphasized parent-child cultural conflict as a significant concern. Additionally, all community samples highlighted themes related to a lack of knowledge about mental health symptoms and not knowing where to go to receive mental health services. A major barrier to receiving mental health treatment noted across all communities was minimizing or having no awareness of mental health concerns and stigma towards mental health. Common coping strategies across community samples were spending time with their community or family, praying, and meditating. Please see Supp Figure 2 for more detail of unique themes within individual CFU communities. The sample groups from the communities also identified job insecurity, COVID-19, changing immigration policies, and health insurance concerns as sociopolitical determinants of mental health. Notably, participants across the communities (except for HHCTF) reported positive impacts of the COVID-19 pandemic as well, such as getting time to reconnect to their native cultures and having more opportunity to spend time with family. All participants also emphasized becoming aware of mental health symptoms and learning about mental health as areas in need of improvement in their communities, and consistently expressed a desire for race-matched or culturally-embedded therapists and group therapy as ways of connecting with practitioners.

Based on these Phase 1 findings, the CFU-research team discussed the most appropriate next step for information gathering and the areas that were still unclear or needed to be confirmed on our tentative treatment adaptation framework. We decided that we needed to gather data from a larger sample of community members from each community and to obtain the concurrent perspectives on these same issues from the community providers serving them (many of whom were from the communities themselves). The entire CFU-research team deemed this additional information crucial in helping us understand the best way to present and intervene on emotional health issues for each community and showcase applying best practices to create a treatment adaptation framework that psychologists could apply and test across diverse communities.

Phase 2 Quantitative Data

Method

Participants

The CFU community leaders recruited participants via purposive sampling consistent with Phase 1. The only exclusion criteria were that participants, both community members (with a target of 15 participants per CFU community) and community providers (10 providers per community), could not have participated in Phase 1 of the study (qualitative) to be eligible for Phase 2 (quantitative) and participants had to be at least 18 years of age. However, due to an unexpected change in the American Indian/Alaska Native community leadership, we were unable to recruit community members and providers from this community. For the CBC community, our community partners were unable to identify providers who consistently provided mental healthcare to church members, even though they were easily able to provide community member respondents for the survey. Hence, additional providers serving multiple communities were recruited from the Utah Department of Health’s (UDOH) Community Health Worker (CHW) section (N=31), and we were able to obtain an adequate number of CHWs who verified that they serve the CBC community at least. CHWs were chosen due to the community leaders’ suggestion of positive relationships and trust embedded between CHWs and the communities of color under investigation. Moreover, significant literature discusses the benefits of using CHWs with communities of color or low-and middle-income communities (Balcazar et al., 2011). Thus, the total N for Phase 2 was 120 (59 community members from BoA, HHCTF, CBC and NTAS, 61 providers including the subsample of CHWs).

Measures

Community member survey.

The construction of this survey was guided by the general themes that were developed from the data analysis of Phase 1 data. Specifically, the team incorporated the communities’ definitions of mental health and the preference for the term “emotional health” as determined by the Phase 1 data into the Phase 2 survey’s introduction (e.g., As a reminder, mental or emotional health includes emotional, psychological, spiritual, and social well-being…). Further, in line with best practices, the surveys incorporated community leaders’ suggestions that wording should be at a fifth-grade reading level and to provide no more than three descriptive options per item (e.g., Not at all, A little, A lot). The survey was translated into Spanish for the HHCTF community, while a translator was hired to collect the survey responses from BoA community members via live translation in Kirundi. The other communities indicated that they were comfortable not providing any additional translation.

Emotional health symptoms were listed out in the survey using similar language as was provided to the team by community members in Phase 1; this was done to consistently represent the communities’ stated terminology for mental health symptoms in the Phase 1 focus groups. The following symptoms were described: feeling sad (depression), feeling a racing heartbeat, sweating, or tense muscles (anxiety), worrying about different areas in life (chronic worry), thoughts about harming or killing self (suicidal ideation), feelings of stress about upsetting events (trauma), and stress related to COVID-19, racism, and changing immigration policies (these stressors were prominent when data were collected in Fall of 2020). Stigma toward emotional health was another common theme that was raised by participants across the Phase 1 focus groups. Thus, the survey in Phase 2 further assessed self, social, and structural stigma (e.g., I would feel ashamed of myself if I was having emotional health problems), using items as derived from the Stigma-9 Questionnaire (STIG-9) and the Self-Stigma of Seeking Psychology Help (SSOSH) scale (Gierk et al., 2018; Vogel et al., 2006).

To further corroborate focus group results, the survey assessed for perceived barriers to receiving treatment for emotional health, matching themes that were apparent in Phase 1. Specifically, these questions were shaped by the language and examples used in the focus groups (e.g., not having health insurance, fear of racism, lack of trust in providers, language barriers, not knowing where to go). Additional questions regarding common physical symptoms were also included based on the definition of emotional health as including one’s balance with physical health, an area that was not explicitly explored in Phase 1. Finally, given that the broader aim of the overall funded study was to develop an emotional health intervention for these communities, questions regarding participants’ openness to treatment and preferences for treatment modality were also included in the survey (see Supp Figure 3 for community members’ survey). The final Qualtrics survey took approximately 10–15 minutes to complete.

Provider survey.

The provider survey was constructed similarly to the community member survey, with results from the focus groups and CFU leadership guiding its creation. Providers answered personal demographic questions and about the community members they serve (e.g., sexual orientation, ethnicity, ages), the most common emotional health symptoms they observe, and how confident they are in understanding those symptoms. The survey included questions about the kind of support they provided and if and how they modified the presentation of their care for their ethnically diverse clients. Providers’ interest in training, feelings of burnout, and their clients’ physical symptom complaints were also assessed. This survey was also kept at a fifth-grade reading level but included a greater number of questions and answer options per item (per community leader input that this was appropriate), taking approximately 15–20 minutes to complete via Qualtrics (see Supp Figure 4).

Procedure

Participants (community members and providers) were informed that this study was designed to understand how their community was managing or treating current emotional health needs. Additionally, they were informed that the survey would ask them about perceived barriers to obtaining mental health treatment. A University of Utah IRB-approved consent cover letter was provided to all participants and the CFU leaders collected verbal consent from both community members and providers. Research staff delivered the consent cover letter to the CHW providers recruited from UDOH and obtained their consent via email. Upon receiving participants’ contact information, Qualtrics survey links were sent out for completion by study staff via email. Study staff was available to aid in the completion of surveys, including aid with technology, aid with Spanish-translated material, and support during live Kirundi interpretation sessions with BoA community members. CFU leaders requested that participants have several compensation options, including a physical gift card for a local grocery store and digital gift cards for Walmart and Amazon. Each participant received a $10 gift card of their choice for their participation. Payment contact information was not linked to survey answers to assure confidentiality. For more details on methods/procedure for Phase 2, see Asnaani et al. (in press).

Data Analysis

Descriptive statistics were used to analyze survey variables of interest. Phase 2 data were analyzed with the intent of exploring how survey results from a broader subsection of each community and their providers mapped on to Phase 1’s qualitative findings and either corroborated or modified the initial treatment adaptation framework. For the community data, our intent was to pinpoint the specific mental health concerns of these ethnically diverse communities which were most in need of intervention, the current barriers to care for such issues, and the intervention methods that were most likely to be acceptable and useful for each community. The provider data were analyzed to determine if there was consensus between providers and community members on the communities’ top mental health concerns and barriers per each of the 4 CFU communities included in this phase of data collection. CHW data were analyzed separately since they service multiple CFU communities.

Results

The Phase 2 findings for the four CFU communities that we were able to collect data from (i.e., African American, Latino/a, African immigrant and Pacific Islander samples) generally further confirmed the trends observed in Phase 1 data, with the inability to pay for treatment as a barrier identified across all communities. Notably, due to discussions in the focus groups about the current political climate (in the summer before the 2020 elections), racism and discrimination were identified as circumstances affecting mental health. Thus, in the Phase 2 surveys, we specifically asked about stress from these concerns. Not surprisingly, the top emotional health concern identified across community samples was feeling stress and anxiety about COVID-19, racism or racial unfairness, and changing immigration policies. This most frequent concern was followed by worrying about different areas in life, such as finances, family relationships, and health. Providers serving these groups confirmed that the top emotional health concern they see in the community was chronic worrying about different aspects of life. Additionally, depression was reported as the third most frequent emotional health concern by both community members and providers. See Supp Figure 5 for community-specific and provider results for the four CFU communities for which we have this data.

Best Practices for a Community-Informed Treatment Adaptation Framework

Using constructivist grounded theory to interpret Phase 1 qualitative data and as confirmed by Phase 2 quantitative data, we created a treatment adaptation framework that integrates best practices for health disparities research for racially/ethnically diverse communities overall (Figure 1) and community-specific frameworks for four of the CFU communities (see Supp Figures 6–9). We did not obtain adequate data to justify applying this framework to our American Indian/Alaska Native community sample. For the other four communities, this framework depicts the circumstances impacting the community members’ emotional health, the culturally relevant coping strategies they prefer to use, the sociopolitical determinants that hinder them from seeking treatment, the barriers that interfere with seeking support, and, importantly, the preferences community members and the clinicians who work with them express as suggestions to improve mental health services. The elements of this framework were informed by the consistent findings from both Phase 1 qualitative data and Phase 2 survey data, in conjunction with the theoretical model (i.e., FMAP) and community-based methodological principles guiding the study (i.e., CBPR and social justice, both of which place the voices of stakeholders in a central place). Notably, the elements of the framework also reflect a need for adapted interventions to address social determinants of mental health (see “Causal conditions: Macro-level” and “Context”). Such elements ensure that we refrain from using the solely individual deficit-model that has dominated our field, which tends to imply that mental health distress only occurs due to individual factors or deficiencies (McCormack et al., 2016). Acknowledging the macro-level factors and social determinants aligns with social justice efforts to address broader issues perpetuating health disparities in communities of color (Collins et al., 2018).

Specifically, Figure 1 directs us to explore and identify the macro-level causal conditions (e.g., systemic racism, forced migration, health inequities, racial discrimination) and individual-level factors (e.g., not having tools to deal with emotional health challenges) that can lead to a community’s reported mental health phenomena/sequelae of most interest (e.g., in our current sample, chronic worry and daily stress from sociopolitical events such as COVID-19, racial strife, and political unrest). Further, the framework guides an investigation of the mitigating/coping strategies a community uses to handle these mental health symptoms (e.g., praying, physical activity, being around community members) to ensure that these factors are integrated into a chosen evidence-based treatment. Simultaneously, we must assess for intervening conditions/barriers (e.g., stigma, language barriers, unawareness of mental health issues) and externally imposed sociopolitical factors (e.g., immigration status, racial discrimination, health insurance issues) that may impede successful treatment within a particular community. Taking all these parts of the framework into account demonstrates how we can apply best practices to address health disparities by thoroughly evaluating the consequences of a chosen intervention and incorporating community preferences for treatment in order to maximize the success of a chosen treatment approach for that community.

Application of Framework with Current Sample

Using the current community samples as an applied example of this framework, our CFU-research team generated a culturally-adapted, testable intervention for the main areas of concern that were consistently and most frequently identified areas of mental health distress across the four partner communities for which we had adequate data to inform the framework (i.e., CBC, HHCTF, BoA, and NTAS) or target “phenomena” (see Figure 1). The phenomena include chronic worry (formally defined as pervasive and excessive patterns of anxious rumination about a variety of life areas; Roemer & Orsillo, 2009) and daily distress (defined by our community samples as feeling emotionally upset or stressed) about COVID-19, racism or racial unfairness, and changing immigration policies. Then, we consulted the literature and selected the mindfulness and values-based action protocol by Roemer & Orsillo (2009), given the considerable empirical support for its effectiveness broadly (Hofmann, Asnaani, Vonk, et al., 2012) and with specific communities of color or issues related to racism specifically (Fuchs et al., 2013). Briefly, this treatment entails providing psychoeducation about the function of worry, guided self-reflection on how to differentiate between how one feels and how one behaves, mindfulness (present-centered awareness) as a skill, and guided exercises around values clarification and engaging in values-congruent actions. With these specific treatment elements and structure in mind, we applied best practices to create a framework that delineates what cultural factors would need to be incorporated (and then subsequently tested) to culturally adapt this gold-standard treatment for use in each of these communities.

Going from left to right in this treatment adaptation framework, we first proposed additions to the treatment that directly address some of the macro-level and individual level causal conditions. For instance, for those with more recent immigration histories (such as those identifying as African refugee/immigrant or first-generation Latino/a from our current communities), we will include integrated services from other community centers with which we have already established relationships. This integration ensures we provide services around adjusting to various facets of acculturative stress (including instruction often provided by these centers on practical skills such as finding job placements, navigating the education system, or learning English, if desired). For the individual level, and align with best practices to culturally-responsive clinical care, we proposed a module to be added before the already-existing psychoeducation pieces of this mindfulness and chronic worry treatment protocol. The new module would introduce culturally-congruent definitions of mental and emotional health, describe commonly occurring symptoms, and discuss existing resources for various mental health concerns in each community (all directly derived from the current mixed-methods study). The lack of awareness of mental health symptoms present across all our partner communities drove the decision to include this module.

Next, we incorporated ways to address some of the intervening conditions directly. For instance, in addition to our direct efforts to address the gaps in awareness and literacy about mental health, we also included a module on stigma reduction (again, quite common across the communities, but particularly the African immigrant, Latino/a, and Pacific Islander communities). The stigma reduction module allows the therapist to directly address stigmatizing thoughts or beliefs by helping the client work through the origin of such thoughts or their utility in getting the help they need. In addition, we deliver our treatments (as we did for our focus groups and some of the surveys) in the preferred language of the client, which is in line with best practices and addresses language barriers that can impede receipt of evidence-based treatments, as widely documented in the literature (McElfish et al., 2019).

Following this, we have proposed the inclusion of community-specific coping skills into the delivery of core treatment strategies. For instance, for those communities indicating a strong faith base (e.g., the African American and Latino/a communities in our region), we ensured that one of the major domains of values discussed within the values-based protocol for chronic worry is spirituality, and that praying features as a major values-based action, along with other faith-based activities that are personally meaningful to each client. Similarly, communities expressed specific forms of physical activity that we could integrate as core actions into the treatment that were most culturally acceptable or relevant to each. For instance, some communities specifically liked the concept of yoga. However, others found it to be a divisive/religiously-laden term. Hence, in line with best practices, we instead discussed stretching movements while engaging in prayer or while using culturally-acceptable spiritual terms to still anchor to core mindfulness skills.

Finally, we ensured that our treatment incorporates direct efforts to address the broader societal level determinants driving health disparities. For instance, we are explicitly targeting job insecurity (achieved via a partnership with the community-specific social agencies we mentioned before, facilitated by the mental health provider) and actively providing support of community-organized COVID-19 vaccine and information clinics. For instance, we have partnered with health ministries within the African American community and advocacy organizations within the Latino/a community to advertise and obtain experts to provide virtual townhalls specifically focusing on COVID-19 education and misinformation. We also partnered with the state health department to obtain low-cost or no-cost options for providers to address the disparate health insurance context. These features all dovetail nicely with the final element of our treatment adaptation framework related to centering the treatment preferences for the communities (“Consequences”) whereby these more affordable providers are the CHWs who are already embedded in and trusted by these communities whom we are now training in this evidence-based treatment. CHWs also have training in connecting clients to vocational, financial, and other health services. Using a survey in this particular workforce for the provider data in Phase 2 verified the CHWs’ desire and openness to obtain such training. In this way, we can meet the mental health needs of these communities by using culture-matched therapists while also building desired capacity and knowledge within the communities themselves.

Discussion

The main aims of this study were to exemplify best practices for addressing mental health inequities in communities of color in terms of (1) conducting health disparities research and (2) applying a treatment adaptation framework for culturally-responsive clinical care that stems from such research. For the first aim, we generated an overall treatment adaptation framework using a bottom-up mixed-methods approach to more fully demonstrate how we can apply best practices in health disparities research to create such a framework. Specifically, the treatment adaptation framework we present here draws on the qualitative and quantitative data findings from 169 individuals from four ethnically, racially, and culturally diverse communities in the Mountain West region of the U.S. and synthesizes these findings with strong theoretical models for culturally-responsive intervention work (i.e., the FMAP; Hwang, 2009). Particularly, our two rounds of data collection mapped onto the FMAP phase 1, where we generated community-specific and overall knowledge in collaboration with community leaders. Following data collection, the CFU-research team discussed the results and began formulating a treatment adaptation framework that integrated theory, empirical, and clinical knowledge consistent with FMAP phase 2. The research team then refined the framework and presented a proposed treatment intervention for chronic worry and stress using mindfulness and values-based action to community leaders for feedback and suggestions in line with FMAP phase 3. Further, this study incorporated and tested several community-based participatory research (CBPR) methods (Wallerstein & Duran, 2017). The resulting framework also takes social justice principles into account in how it prescribes a broader view to addressing sociopolitical context and directly targeting barriers to equitable mental healthcare for communities of color.

In terms of the actual results from the study, the current study replicated findings around the mental health symptoms of most concern to racially and ethnically diverse communities and the barriers to mental healthcare for such symptoms, which are largely in line with previous studies (Derr, 2016). Hence, the individual findings from the community samples themselves are not new. However, it is notable that this study was conducted entirely within the context of the COVID-19 pandemic. Thus, this is one of the few mental health studies to be conducted in such a range of communities of color during the height of this historic public health crisis, the mental health impact of which is expected to be long-lasting and particularly exacerbated for these specific communities. Notably, despite considerable disruptions in work and childcare and frequent illness due to the pandemic, each of these communities showed admirable strength and resolve to share their perspectives about their emotional health needs to ensure their voices in this domain were heard within the context of this study.

In addition, this study also reveals a unique and less commonly used approach to guide best practices for mental health disparities research in underserved communities by its use of a two-phased, mixed-methods approach that integrates qualitative and quantitative data to generate the treatment adaptation framework for these communities. Such an approach has undoubtedly been heralded as an important step to appreciably addressing health disparities by informing more scalable and testable treatment frameworks (Lau et al., 2010; Bernal & Scharrón-del-Río, 2001), but the implementation of CBPR guidelines (in larger samples and with sufficient methodological rigor) still remains scarce within mainstream psychology (Rodríguez-Espinosa & Verney, 2021). Further, the strong community partnership in the current study that allowed us to examine such factors across multiple communities of color is even less commonplace and highly valuable to our understanding of tangibly addressing ongoing mental health disparities in a range of culturally diverse communities.

To achieve the study’s second aim, we exemplified an application of the treatment adaptation framework we developed in Aim 1 to guide best practices in culturally competent clinical care. For this, we chose the major intervention target areas (chronic worry and stress due to ongoing political and social stressors) that were highlighted as consistent areas of need across the four very different communities in our mixed-methods study and discussed how we could apply an evidence-based treatment for this problem area (mindfulness and values-based action; Roemer & Orsillo, 2009) in a culturally adapted way for further testing in each of the communities. Indeed, this balance afforded by the proposed treatment adaptation framework between finding common target areas among multiple communities of color while maintaining flexibility in incorporating unique community-specific factors that fall into the major categories shown in the framework is paramount for our work to address health disparities going forward. The framework we have proposed here that allows us to create tailored adaptations for a range of specific communities certainly extends important existing work in the field examining integration of cultural factors into EBTs more broadly, such as Culturally Informed Therapy for Schizophrenia (Weisman de Mamani et al., 2021) and cultural adaptions for depression (Ramos & Alegría, 2014), among many others.

Taken together, the current study is instrumental in modeling several key features of applying best practice to address health disparities for researchers and practitioners alike. First, our study’s strong CBPR orientation provides a valuable roadmap for forming community-engaged partnerships, whether for a specific research purpose or to determine how to deliver a clinical intervention (Asnaani et al., in press). Specifically, the close collaboration with community leaders and extensive opportunities for community input via qualitative and quantitative data are part of best practice approaches to addressing health disparities in culturally diverse communities because they allow for a more focused and knowledgeable approach with multiple viewpoints taken into consideration. Second, as we mentioned previously, this study highlights a great application of a mixed-method approach, another hallmark of best practices for health disparities work (Lau et al., 2010). Finally, a lesson we have learned keenly in this work is the propensity for researchers and policymakers to enter community-based health work with a perspective based predominantly on our own models of training and our conceptualizations of health. This study’s findings serve as an important reminder that a best practice approach to health disparities work with communities of color includes centrally placing the perspectives of those very individuals we seek to serve. Thus, we must take a ground-up, community-partnered approach from initial conceptualization through data collection and subsequent intervention, rather than a post-hoc or superficial approach, which applies to both research and practice. By doing so, we can push ourselves as clinical scientists to generate the most culturally-responsive and adaptive solutions while also empowering our communities to be equal partners in this work (Collins et al., 2018).

So, where do we go from here, with this treatment adaptation framework that, while it was certainly labor-intensive to create, is a translatable deliverable? We now should move to another key feature of best practices in addressing health disparities (and phase 4 of the FMAP): the testing and implementing of such frameworks, both in our CFU communities and with application to other underserved communities. To this end, such efforts are already underway. For instance, we are currently pursuing testing of our modifiable, culturally-responsive, mindfulness-based and values-action approach for chronic worry (Fuchs et al., 2013; Roemer & Orsillo, 2009) as we described in detail as an applied example of the framework in our results. We are planning on (1) specifically targeting the macro-level causes highlighted for our current sample for our target area of chronic worry (i.e., partnering with community-specific social agencies and advocacy organizations to address issues related to immigration and acculturative stress), (2) addressing cultural barriers (e.g., adding a stigma intervention module to our treatment, conducting treatments in preferred languages), (3) directly targeting sociopolitical determinants to effective treatment options for these communities (e.g., high cost of treatment and lack of trust in providers from external cultural backgrounds by training CHWs who are themselves a part of these communities), (4) incorporating culturally-specific coping mechanisms into the core strategies in this protocol (e.g., prayer and different types of culturally-acceptable mindful practices, values exploration in spiritual or family/community relationships domains), and (5) ensuring community preferences from community members and community providers alike (i.e., providing training to CHWs, partnering with our local health department to make treatments more accessible and to centralize training) significantly frame the treatment adaptation we test.

Limitations

Although this study employed a mixed-methods approach to garner rich data across multiple racially/ethnically/culturally diverse communities, it has several limitations. First, we could not adequately assess intersectional areas of identity (e.g., age, sexual/gender identity, religion) and incorporate this more fully into our culturally-responsive framework, which is central to understanding communities of color (Lau et al., 2010; Plaut, 2010). Therefore, future studies should consider how intersectionality influences best practices for culturally-diverse communities within the various facets of the proposed framework. Second, due to an unexpected administrative leadership turnover at our site serving the American Indian/Alaska Native community, we were unable to collect quantitative data for this community. This recruitment challenge directly impeded our ability to generate a community-specific framework for this community that adequately represented the voices of this community as we were able to achieve for the other CFU communities. In addition, it highlighted a broader discrepancy for our research team in terms of the mental health conceptualization and its relation to the unique historical, spiritual, and holistic background for this community compared to the others (for an excellent review on the unique contributors to mental health for the American Indian community, see Gone, 2022), which we believe deserves its own focus and exploration in future work and that we are currently pursuing with community leaders. In line with good CBPR practices, we jointly agreed as a community-research team to retain presentation of this community’s data for the Phase 1 focus groups to ensure they were represented in our work here, and to ensure complete transparency around both the successes and difficulties encountered in our research process. Third, the survey for Phase 2 of the study was not a validated measure. However, aspects of the survey came from validated measures (e.g., questions regarding stigma). Moreover, the survey was created in collaboration with community leaders to ensure that length and content were most appropriate to the target communities, in line with best practices for community research.

Lastly, this study included assessments of a relatively small number of community members from the CFU communities compared to extensive epidemiological studies that have examined some of the same issues in much larger samples (e.g., Asnaani et al., 2010). However, the qualitative data we gathered were much richer than data collected from a quantitative survey or prevalence rates alone. Ideally, future studies should consider larger quantitative examinations of these phenomena in addition to high-quality qualitative data as done here to yield a more comprehensive picture of mental health needs and disparities impeding best practices in culturally diverse communities.

Conclusion

Despite significant guidance from health disparities experts in the field, previous studies evaluating EBTs have seldomly employed a combination of quantitative and qualitative methodologies in conjunction with strong theoretical models, CBPR principles, and a social justice lens to guide best practices in health disparities work. In contrast, the present study applied best practices for health disparities work to create a treatment adaptation framework using a bottom-up mixed-methods approach that clinicians, researchers, and community partners can employ across culturally diverse communities to ensure that we do not silo these communities. Specifically, this study exemplifies best practices for health disparities research given (a) the rigorous mixed-methodology used, (b) the implementation approach (CBPR, through all phases of the research), and (c) the process in which the final deliverable (a testable framework for treatment development) was created. Additionally, this framework itself presents a best practices exemplar for clinicians looking for specific guidance on adapting their EBP to meet the needs of diverse clientele/specific communities in a culturally-responsive fashion.

As an applied example of this framework at work, we have been able to engage community leaders, our state Department of Health, the CHW workforce, and community members to use the framework to adapt an intervention focused on addressing the top mental health priority (chronic worry) of each community sample, with the flexibility to tailor to the specific needs, barriers, and contextual factors (such as the stressful context of COVID-19) impacting each of the different CFU communities. Importantly, this work highlights a responsibility for all of us as psychologists to engage in advocacy efforts to better support the communities we work with, including more centrally incorporating social justice principles that address broader social determinants (Asnaani et al., 2020). Indeed, our field would greatly benefit from more consistent adoption of such a multi-faceted approach, as this can empower communities of color and other underserved groups and has the real potential, as is our sincere hope, to significantly reduce ongoing mental health disparities.

This study highlights that close collaboration with community leaders and opportunities for community input using mixed data methods are the best practice approaches to addressing health disparities. Also, a ground-up approach to community partnership, from initial conceptualization to data collection and subsequent intervention design, empowers racially diverse communities.

Appendix

Figure 1.

Communities of Color Best Practice Treatment Adaptation Framework