Abstract
Social workers can promote resiliency among refugee families by referring them to evidence-based programs to reduce the stressors of resettlement. The purpose of this study was to complete a structured adaptation process with the SafeCare® program for implementation in a refugee resettlement community. Participants included 21 members of an adaptation team made up of administrators, supervisors, and family service providers from three community agencies and community health workers. Quantitative findings suggested that content, process, and literacy-related adaptations were necessary to ensure cultural relevance of program materials. Qualitative feedback suggested the adaptation approach was a meaningful process that engaged community members and resulted in an acceptable and feasible curriculum for delivery in the refugee resettlement community, which will be further tested in a forthcoming implementation trial. The multi-pronged, community-engaged approach to SafeCare adaptation is presented as a potential framework for other programs that could benefit refugee children and their families.
Keywords
There is growing consensus that refugee and immigrant health is the public health crisis of this century (Syed & Mobayed, 2017). While the physical health needs of displaced refugee populations are significant (Diaz et al., 2017), their mental health needs are equally important (Cheung Chung et al., 2018). Children under the age of 18 make up about 42% of the world’s refugees, but only about 30% of the world’s population (UNHCR, 2021). Children who are refugees are at especially high risk for mental health issues, with studies documenting high rates of PTSD and traumatic grief (Bronstein & Montgomery, 2011), depression (Bronstein & Montgomery, 2011), anxiety (Betancourt et al., 2012), and behavioral problems (Betancourt et al., 2012). Exposure to violence and other extreme adversity before, during, and after experiencing forced displacement and migration have been identified through systematic review as key risk factors for poor mental health outcomes among refugee children relocating to high-income countries (Fazel, Reed, Panter-Brick, & Stein, 2012). The primary support system for these children—their parents or caregivers—have also experienced extreme stressors that affect their ability to support their children (Fisak & Grills-Taquechel, 2007). Addressing and minimizing post-migratory risk factors and bolstering protective factors, such as positive parenting practices, are vital in addressing the mental health needs of refugee children.
Refugees have a range of needs upon resettlement and are often reliant on service agencies to fill those needs. Social workers can assist refugee and immigrant families in variety of ways: by connecting them to concrete services; by providing counseling; and by referring them to specific evidence-based interventions to address specific problems challenges and reduce detrimental outcomes (Rine, 2018). For children exposed to adversity, interventions targeting positive parenting practices can limit the negative impacts of that adversity and prevent future exposure (Gee, 2016; Weine et al., 2004). Positive parenting practices buffer children from the negative biological impacts of stressors (e.g., Brody, Lei, Chen, & Miller, 2014; Chen, Miller, Kobor, & Cole, 2011; Luby et al., 2012; Miller, Chen, & Parker, 2011), and from mental health diagnoses and externalizing behavioral problems (Eltanamly, Leijten, Jak, & Overbeek, 2021). Importantly, positive parenting practices, such as parental warmth, are negatively impacted by the types of trauma exposures that refugees commonly face, including war, political violence, armed conflict, and displacement (Eltanamly et al., 2021). Interventions to promote positive parenting can bolster parents’ ability to provide warm, supportive relationships, a safe environment (Taylor & Biglan, 1998), and can increase the positive social, emotional, and behavioral development of children (National Scientific Council on the Developing Child, 2004). Programs that target parenting skills can also improve parent mental health, with meta-analyses associating parenting programs with reduced depression, stress, and anxiety (Barlow et al., 2011).
Researchers have postulated that interventions to bolster parental support for children in the aftermath of trauma are a suitable approach for immigrants and refugees (Zagelbaum & Carlson, 2011). Parenting programs to attenuate the negative effects of the refugee resettlement process on children can also promote healthy development (Birchler, 2017). However, a recent systematic review (Slobodin & de Jong, 2015) reported that limited research exists documenting the effectiveness of parenting interventions or systematic therapy for traumatized immigrants and refugees, for whom the global number is growing exponentially (UNHCR, 2021). Therefore, increasing access to culturally relevant evidence-based parenting (EBP) interventions may promote health equity, social justice (National Association of Social Workers, n.d.), and enhance the lives of refugee children and families who have experienced extreme hardship before, during, and after their forced displacement and migration experience (Parra-Cardona, Bybee et al., 2017).
Though parenting programs have been used successfully around the world with a wide range of cultures and populations (Gardner, Montgomery, & Knerr, 2016), successful implementation with refugees, especially during the resettlement period, poses unique challenges. Certainly, logistical barriers exist, such as language and translation issues, but perhaps more important is to ensure the cultural relevance of the program curriculum and delivery approach. Amongst EBPs, there is the expectation that programs be delivered with fidelity and adherence to the standards developed for efficacy trials; however, strict adherence may be at odds with successful implementation in real-world practice settings (Aarons et al., 2012). Some studies suggest that EBPs without adaptations can produce positive effects (Chaffin, Bard et al., 2012), while others suggest a need to consider cultural adaptations for program effectiveness (Baumann et al., 2015; McCabe & Yeh, 2009; Parra-Cardona, Bybee et al., 2017) as well as acceptability and uptake by the targeted populations (Castro & Yasui, 2017; Kumpfer, Magalhães, & Xie, 2017; Lau, 2006; Mejia, Calam, & Sanders, 2017). Further, program engagement and uptake are generally poor, and prevention programing is underutilized by members of racial and ethnic minorities, including refugees (Byrow, Pajak, Specker, & Nickerson, 2020).
There is limited work on adapted interventions for resettled refugee and immigrant populations (Ballard, Wieling, & Forgatch, 2018; Fazel & Betancourt, 2018; Sim, Bowes, Maignant, Magber, & Gardner, 2021; Slobodin & de Jong, 2015), but several approaches have been proposed as part of the adaptation process for parenting programs delivered among other diverse populations. Approaches to adapting evidence-based programs for new cultural or ethnic groups vary in both the framework used to guide the adaptations and adaptation process and the extent of the adaptations that are made (Baumann et al., 2015; Kumpfer et al., 2017; Mejia, Calam, & Sanders, 2017); however, there are some similarities across approaches (Mejia et al., 2017). Several models that have been used to guide the cultural adaptation of evidence-based programs (e.g., Barrera & Castro, 2006; Bernal, Bonilla, & Bellido, 1995; Ferrer-Wreder, Sundell, & Mansoory, 2012; Resnicow, Soler, Braithwaite, Ahluwalia, & Butler, 2000). Previous research has focused on categorizing the extent of adaptations (e.g., deep structure vs. surface structure; Resnicow et al., 2000), differentiating between types of adaptations (e.g., process, content, ordering, presentation, dosage/intensity; Aarons et al., 2019), defining dimensions of what parts of the intervention can be adapted (e.g., language, persons, metaphors, content, concepts, goals, methods, and context; Bernal et al., 1995), and developing frameworks used to guide the adaptation process (e.g., information gathering phase and preliminary adaptation phase; Barrera & Castro, 2006). Regarding the extent of adaptations, one example of a minimal approach is to examine the cultural relevance of the content by having members of the targeted population review materials and language translation (Mejia et al., 2015, 2017). More in-depth adaptations can be made by embedding culturally and contextually relevant components into the intervention (Domenech Rodríguez et al., 2011), and tailoring Provider characteristics to the population of interest (such as in the Family Spirit program; Barlow et al., 2013, 2015; Mullany et al., 2012), to more positively impact parenting practices (Parra-Cardona, Bybee et al., 2017; Parra Cardona et al., 2012).
Purpose of Study
In 2019, Georgia State University was awarded a grant to establish a Prevention Research Center aimed at improving the health and well-being of the refugee, immigrant, and migrant populations of Clarkston, Georgia, one of thirteen refugee resettlement cities in the U.S. Within this community, more than 40,000 refugees have resettled in Clarkston from over 75 countries. Recent estimates suggest that refugees living in Clarkston experience high levels of financial insecurity (Feinberg, O’Connor, Owen-Smith, Ogrodnick, & Rothenberg, 2020; Feinberg, O’Connor, Owen-Smith, & Dube, 2021; Lyons et al., 2021) and have limited English proficiency (Feinberg et al., 2020, 2021). The purpose of this study is to engage in a structured adaptation process of a behaviorally based parenting program, SafeCare®, to be implemented in Clarkston with refugee ethnic groups from different countries of origin. The multi-pronged approach to SafeCare adaptation is presented as a potential framework for adapting other evidence-based programs to be culturally appropriate so that social workers may feel confident in referring families to the programs to reduce disparities and improve the health of refugee children and their families during the resettlement process.
SafeCare is a broadly implemented behavioral parenting program that addresses three parenting skills critical for the promotion of positive parenting and the prevention of child maltreatment among parents of children ages 0–5 with three modules: (1) parent–child interaction (targets positive parent–child interaction and the reduction of child behavioral problems), (2) home safety (targets parental knowledge about home hazards and child protection from unintentional injury in the home), and (3) child health (targets parental health literacy and prevention of risk for medical neglect). A substantial body of field-based research supports the SafeCare program (Gershater-Molko et al., 2002; Lutzker & Rice, 1987; Rogers-Brown et al., 2020), and rigorous randomized trials of SafeCare have shown positive results on both child maltreatment recidivism (Chaffin, Hecht et al., 2012) and parenting skills (Carta, Lefever, Bigelow, Borkowski, & Warren, 2013; Whitaker et al., 2020). The SafeCare parenting program also has been well received by parents and seen as culturally relevant by diverse groups, including Latinx and American Indian populations (Chaffin, Bard et al., 2012; Damashek, Doughty, Ware, & Silovsky, 2011; Damashek, Bard, & Hecht, 2012). In the majority of international implementations of the SafeCare program (e.g., United Kingdom, Australia, Spain, and Israel), adaptations have been made, but these are typically minor and include changing language, adding content specific to a particular culture, and altering activities (e.g., play activities) to fit within a cultural context. In the adaptations made to date, little of the primary content of the SafeCare program has been altered; most changes have been focused on the editing of written language, addition of home hazards discussed or targeted for removal in the home safety module, and/or changes to the health module to be commensurate with the local health system (Shanley et al., 2013).
Our study will focus on one of the three modules of the SafeCare program, the Parent–Child Interaction module (PCI); its preventive focus directly addresses the type of positive, nurturing behaviors that have been demonstrated to promote positive parenting and improve children’s development outcomes. The PCI module includes two separate protocols, one for parents of infants 0 to18 months and one for parents of children 1.5 to 5 years old. Both protocols promote positive parent–child relationships by teaching parents to attend to their children and interact with them through language, stimulating play activities, positive verbalizations, and touch. Parents are also taught to structure activities using planned activities training (Sanders & Dadds, 1982) in which parents learn to talk with their children about activities, engage them in making choices, explain rules, provide positive reinforcers while ignoring minor misbehaviors, and give positive consequences for the child’s success (Lutzker & Bigelow, 2001). Data from studies focusing only on PCI of the SafeCare program, including a randomized trial, indicate that it is effective at changing parent behavior (Cordon, Lutzker, Bigelow, & Doctor, 1998; Dachman, Halasz, Bickett, & Lutzker, 1984; Lutzker, Megson, Webb, & Dachman, 1985) and child outcomes (Carta et al., 2013; Lefever et al., 2017). This module was selected as the focus of the project after initial discussions with community partners who indicated that parenting was a primary concern in the Clarkston community and the 6-session length of this module would likely be well received by community families.
In the current study, we document the adaptation process and examine the adaptation needs for the SafeCare PCI module in four refugee ethnic groups (Afghan, Burmese, Congolese, and Ethiopian-Eritrean). These groups were selected based on the partnering community agencies reported referral histories and the high number of parents of younger children served by the agencies. There is a strong need for researchers to document their adaptation process, describe adaptations made to evidence-based parenting programs, and evaluate the adapted version of the program (Baumann et al., 2015). While some researchers have done this well (Mamauag et al., 2021), to our knowledge, this is the first known study to document the adaptation process of an evidence-based parenting program to be delivered among a diverse community of refugee populations using community-based participatory methods. We followed Barrera and Castro’s Heuristic Framework (Barrera & Castro, 2006) to guide our adaptation process due to its iterative nature and emphasis on community engagement; and we categorized the types of adaptations based on Aarons and colleagues taxonomy of adaptations (Aarons et al., 2019), given this taxonomy was developed based on real-world adaptations to SafeCare during a large-scale implementation. Findings from this study will be used to inform the program delivery methods for a larger follow-up implementation trial (Whitaker et al., 2021).
Method
Study Overview: Adaptation Process of the SafeCare Parenting Program
Data from this article are drawn from an ongoing Center for Disease Control and Prevention-funded implementation trial launched in September 2019 that will test two implementation approaches to SafeCare PCI delivery for families residing in Clarkston, GA (Whitaker et al., 2021). In this paper, we focus on the first aim of this study: to engage in a structured approach to the adaptation of the SafeCare PCI curriculum for four refugee ethnic groups that are prominent in the Clarkston community (Afghan, Burmese, Congolese, and Ethiopian-Eritrean). The structured approach to adaptation discussed in this paper included the first two stages of Barrera and Castro’s (2006) Heuristic Framework for the cultural adaptation of interventions: (1) information gathering and (2) preliminary adaptation design (Barrera & Castro, 2006).
In Phase 1, the information gathering process included the formation of an Adaptation Team that included individuals with expertise in parenting program service delivery and individuals with cultural expertise for the communities of focus (N = 21) who served as study participants. The team also included a university faculty health literacy expert, SafeCare experts, and GSU research team members (Aarons et al., 2012). The team goals were to (1) complete an informational session on the English version of the SafeCare PCI module curriculum (two SafeCare experts provided an in-depth review of the basic concepts, intervention target skills, and parent materials for PCI), (2) examine the curriculum, and (3) create curriculum content modification recommendations for the relevant to the populations. Immediately following the curriculum information session, Adaptation Team members (excluding SafeCare experts) completed an online survey to assess their perceptions regarding the need for adaptations.
Members of the Adaptation Team were then divided into four “PCI-Adaptation groups” based on their expertise with each of the four identified ethnic groups to examine the curriculum and create modification recommendations. PCI-Adaptation groups met bi-monthly in virtual meetings to complete a critical review of the SafeCare PCI curriculum components, target skills, and parent materials. Each meeting focused on a different SafeCare curriculum category, such as the parenting skills taught during routine activities, child development information, or parent–child play activities. Five meetings were held for each group and were co-led by a SafeCare Trainer and a Research Coordinator. The curriculum being reviewed was shared on the screen on the virtual meeting platform, and as recommended changes were suggested and agreed upon, live edits were made so the group could come to a consensus on the modification. PCI-Adaptation group members were asked to complete a brief satisfaction survey after each PCI-Adaptation group meeting.
In Phase 2, the preliminary adaptation phase of Barrera and Castro’s (2006) framework, the live edits, and information collected during the four PCI-Adaptation groups were used to finalize a preliminary draft of written adaptations to the PCI curriculum by the project research team (principal and co-investigators and project coordinator), two SafeCare experts, and the health literacy expert. The goals of Phase 2 were to use the information gathered during Phase 1 to adapt the SafeCare PCI module for delivery for the refugee ethnic groups. After analyzing the suggested edits and feedback, it was determined that one adapted curriculum could serve the needs of the four target refugee populations. This decision was based on recommendations from the PCI-Adaptation groups that were focused on the revisions of content that were relevant to ethnic groups of focus. Further, there were no recommendations from the PCI-Adaptation groups that suggested adaptation to or exclusion of SafeCare key elements or target skills. Materials were modified using culturally and linguistically appropriate standards (CLAS) and health literacy guidelines, both critical to ensuring responsiveness to diverse cultural groups (Office of Minority Health, 2013, 2014; Plain Language and Action and Information Network, 2011).
A revised written curriculum and supporting materials was shared with all PCI-Adaptation group members and invited them to participate in a qualitative interview. The purpose of the interview was to solicit feedback on the adaptation process and the finalized revised curriculum.
Participants
All study procedures involving human participants were approved and implemented in accordance with the ethical standards of the Georgia State University Institutional Review Board. Informed consent was obtained from all participants. There were no conflicts of interest among study participants. In Phase 1, the information gathering phase, an Adaptation Team was formed. The team consisted of administrators, supervisors, and family service providers from refugee resettlement agencies with expertise in parenting program service delivery in Clarkston; individuals with cultural expertise for the communities of focus (community health workers); SafeCare experts; and GSU research team members. Study participants included the 21 Adaptation Team members who were external to the GSU research team (administrators, supervisors, and family service providers from the three targeted implementing community agencies, and the community health workers recruited from the community as consultants to deliver the intervention). All study participants were either members of the relevant ethnic groups, fluent in the languages spoken by the groups, or had experience delivering services to refugees. Adaptation Team members (excluding the GSU research team and agency administrators) were then split into four PCI-Adaptation groups based on their own cultural identity, targeted language fluency, or experience working with each ethnic group (Afghan, Burmese, Congolese, and Ethiopian-Eritrean). In Phase 2, fifteen PCI-Adaptation Team members consented to participate and completed the qualitative interviews.
Data Collection Measures
SafeCare Modification Survey (Intent/Removing/Integration/Setting/Tailoring)
The modification scale developed by Meza and colleagues (2020) to assess mental health clinicians’ intent to modify an evidence-based mental health therapy post-training was adapted for this study by replacing the name of the EBP in the original items with “SafeCare” (Meza et al., 2020). The 9-item scale measured participants’ (1) general intent to modify the therapy model, (2) intent to remove aspects based on client needs, (3) intent to integrate aspects from other interventions, (4) intent to modify the setting of delivery, and (5) intent to tweak or tailor the intervention to the client. All items were rated on a 7-point Likert scale ranging from Strongly Agree to Strongly Disagree.
PCI-Adaptation Group Descriptive Data and Participant Satisfaction Survey
The frequency, length, and participation rates were collected for each PCI-Adaptation group meeting. After each meeting, participants were asked to complete a six-item survey to assess strengths and challenges of the meeting, feelings of comfort during the meeting, the overall effectiveness of the meeting, and any changes they would like to see in future meetings. The purpose of the survey was to gauge the extent to which different types of members of the Adaptation Team felt heard and part of the group throughout the process and to elicit feedback for making future meetings more inclusive and effective. Survey items included three Likert-scaled questions: How effective do you think the meeting was today? How comfortable did you feel sharing your opinion today? How comfortable do you think others felt about sharing their opinion and open-text response questions?: Please note up to 2 strengths and 2 challenges about the meeting today; What would you like to change about future meetings to make them more effective and/or ensure everyone’s voice is heard?
Adaptation Recommendations
Extensive written notes were taken by research team members during the adaptation team meetings and included the name of the team member, and specific feedback and recommendations on the SafeCare PCI components, target skills, and parent materials. Feedback from team members was cataloged in Excel by three SafeCare experts and research team members, as were the final decisions describing the adaptations made for each of the refugee ethnic groups. These included adaptations to the curriculum content itself and to the implementation process.
Qualitative Interviews
Interviews were conducted to elicit feedback from the PCI-Adaptation Team on the adaptation process and the drafted, revised curriculum developed during the preliminary adaptation phase. Interview questions focused on the adapted materials (e.g., in what ways are they more appropriate for the targeted populations) and the adaptation process (e.g., perceptions of inclusion and whether certain perspectives were not considered). Questions included: (1) How are the adapted SafeCare materials more appropriate for target families? How are they inappropriate? What additional adaptations are needed?; (2) To what extent was the SafeCare Adaptation Team involved in deciding on the adaptations made to the PCI and PII modules?; (3) What did you like most about the adaptation process? What did you like least?; (4) How satisfied are you with the opportunities for you and your fellow team members to express your concerns and opinions?; (5) Were any important voices not heard during the adaptation process? How could this process be more inclusive?; (6) What would you change about the adaptation process?; and (7) Can you comment on the quality of work related to reducing the reading level of the materials? Interviews were conducted via phone by two research team members. Audio recordings were electronically recorded and transcribed using a WebEx conference line. The average interview length was about 34 minutes and ranged from 21 minutes to 60 minutes.
Analyses
Data analyses included descriptive summaries of survey data collected after the curriculum information session and in follow-up to the PCI-Adaptation Team meetings. Feedback was coded by the research team according to Aarons and colleagues (2019) Taxonomy of Process and Content Adaptations (Aarons et al., 2019). According to this framework,
Four research team members coded the Excel sheet notes into the potential taxonomy categories. All coding of meeting notes was completed individually, and the team held a final meeting to discuss ambiguous items and discrepancies identified by the research team. Coding reliability for the PCI-Adaptation group meetings was high (approximately 95%; 232/241 in agreement), with 5% of items needing further discussion to attain agreement. The adaptation categories recommended by the PCI-Adaptation groups were: Process adaptations related to the presentation and ordering of materials, content adaptations (i.e., exclusion and supplementation), and implementation adaptations. Dosage/intensity, forestalling, and selective integration did not emerge as types of adaptations recommended by the PCI-Adaptation groups.
Qualitative analyses were led by a research team member (Willging) who trained interviewers and coders. The two team members that conducted the interviews checked the transcriptions for accuracy, and qualitative interviews were transcribed and analyzed in a two-step approach. In the first step, the two-team members open-coded eight randomly selected transcripts to develop an initial set of codes with guidance from an expert in qualitative research. Then, additional interviews were randomly selected and double-coded using NVivo software until consensus between the two coders was met, resulting in four double-coded transcripts. During the second phase, the two coders and the Research Coordinator met after double coding each transcript to review reliability, adjust the codebook structure and coded definitions when necessary, and finalize the codebook used for analysis. The remaining 11 transcripts (including the 8 transcripts that were open-coded) were randomly divided between the two coders to code independently using NVivo. The coders used NVivo to query data by participant attributes, including adaptation group and project role, and then summarized results in a descriptive matrix.
Results
Quantitative Results
SafeCare Modification Survey Summary
Surveys were sent to the 21 Adaptation Team members to determine the participants baseline thoughts regarding the need for SafeCare program adaptations; 14 participants (67%) completed the survey. Six completers were agency supervisors, 5 were family service providers, and 3 were community health workers. All six supervisors identified English as their predominant language and were born in countries not representative of the target ethnic groups (U.S., U.K., and Russia), with 67% having 5+ years of experience working with a target ethnic group. Of the family service providers and community health workers, all were immigrants, 86% were fluent in relevant languages, 57% had 3 or more years of experience working with a target ethnic group, and 43% had 0–3 years of experience. The majority of participants (79%) agreed that general modifications to SafeCare were needed. Specifically, over one-third of participants indicated they would need to modify SafeCare to fit the needs of their clients by removing aspects, and the majority noted that they would consider integrating aspects from other interventions, modify the delivery setting (typically delivered in the home), and tailor or tweak SafeCare. Approximately 40% of participants reported that there were some aspects of SafeCare that may not apply to the community they serve. In sum, data suggest that an adaptation process for SafeCare was highly necessary based on the perceptions of the community-based research partners with experience and knowledge of the refugee communities who will be targeted for recruitment in the future implementation trial.
PCI-Adaptation Group Meeting Data and Participant Satisfaction Survey Summary
Results regarding meeting participation and team members satisfaction with the group process are presented in this section. Meeting attendance varied slightly among the groups; the average attendance rate across the five meetings for each PCI-Adaptation group was 75% among the Ethiopian/Eritrean group and 80% for the other three groups (i.e., Afghan, Burmese, and Congolese). Satisfaction surveys were administered after every PCI-Adaptation group meeting. Participant survey completion was low, with an average of 45% of group members completing the surveys following each meeting session. The Ethiopian/Eritrean had the highest post-workgroup meeting survey completion rates (average 78%), followed by Afghan (67%), Burmese (35%), and Congolese (25%).
Meeting topics were divided into SafeCare curriculum categories, with a large focus on the parent materials that are disseminated during session delivery. On average, participants rated the meetings to be moderately to very effective (average 3.86/5). Notably, participants indicated that they were very comfortable with sharing their opinions about the materials (4.43/5) and that they viewed other group members as comfortable sharing their opinions (4.19/5).
Adaptation Recommendations Summary
Adaptation Groups Recommendations for SafeCare PCI
In total, across the 4 PCI-Adaptation groups, 113 adaptations were recommended. Adaptations were distributed between process (38%; 43/113) and content adaptations (62%; 70/113). For process adaptations, approximately 41 out of 43 recommendations were related to the Presentation of materials, while a few (2 of the 43) were related to Ordering. Examples of the process adaptations for Presentation included simplifying language in the parent materials to meet literacy needs, removing language that does not make sense in translation (i.e., “Curious George”), and translating parent materials to the native languages of the parent. Ordering adaptations were mainly focused on the reorganization of the presentation of the parent materials or combining some of the parent–child play activities.
For Content adaptations, 49 out of 70 (70%) recommendations centered on Supplementation. Examples of content Supplementation included adding information not typically covered in PCI (e.g., common child safety practices in the United States such as safe infant sleep practices and car seats as well as United States (Georgia) laws about child maltreatment and mandated reporting), or including additional examples or explanation to help clarify SafeCare content, target skills, and activities (e.g., explaining that the reason you do not leave children to play independently for long periods is because young children get bored quickly and developmentally cannot be expected to play on their own for much time). The remaining Content adaptations focused on Exclusion (21 of 70; 30%). Content suggested for removal included play activities deemed not appropriate for refugee populations (either they would not commonly have items to organize the activity in their home [i.e., finger painting] or there was a cultural appropriateness issue [e.g., engaging child in a word hunt while shopping in stores if parent’s literacy was low]), references to U.S. foods, games, terms of play or endearment that would not translate (e.g., we changed the name of the game, “I Spy” to “Find a Toy” to ensure the spy reference would not make families uncomfortable), supplemental materials on topics not relevant or consistent across cultures (i.e., daily routines or bathing), and pictures that were not culturally relevant or helpful in conveying parent skills (e.g., the U.S.-based parent materials have pictures of parent and child interacting in various ways to make the materials more friendly in appearance; however, many do not actually depict the target skill being taught).
Health Literacy Expert Recommendations
In addition to the Process and Content adaptations, the team health literacy expert had feedback that influenced the draft of the final curriculum based on CLAS and health literacy guidelines. Specific recommendations were to change the language to be more action-oriented (for example, instead of saying “physically redirect the child,” say “redirect the child to a different area or activity”) and categorize materials according to child age group by year (between ages 0 to 5 years), with a table of contents to depict this.
Qualitative Interview Results
Across the fifteen interviews, all study participants agreed that the adapted materials were appropriate and ready for use in the implementation trial. Participants noted that the simplified language (26.7%; 4/15) and revised pictures (6.7%; 1/15) would make it easier for parents to understand the program lessons. “I think that through [the adaptation] process we really brainstormed and troubleshooted and thought about potential challenges and potential best vocabulary and pictures that should be used throughout all of the materials. I think there was a lot of focus and care taken to make sure that it was not going to be something too unfamiliar or something that was unable to be understood.”
The Adaptation Team participants also noted that the revisions made to the materials improved the cultural appropriateness (46.7%; 7/15) and suitability for refugee families’ living situations (33.3%; 5/15) in the Clarkston community. “There are also a lot of examples that were changed to be more reflective of the refugee experience, rather than examples that would be appropriate for an American audience. An example of that we talked about, like, your normal craft supplies at home. Well, that’s not really a thing in the refugee community… So I think just in general, [after addressing those types of aspects] it’s culturally more appropriate.”
Participants reported feeling very positive that the revised materials will allow for increased parent comprehension of the program and ultimately increase the benefit of the program to parents and their children (100.0%; 15/15). “The adaptation process, I think it helps babies or children grow and [gives] mothers an opportunity to learn… I think this [SafeCare] makes a big difference, and it makes a positive impact on mothers with children raising them, and especially in a new country, the United States, which can be different from their home countries. And I think it, it helps [and] matters a lot...”
In regard to the adaptation process, participants consistently agreed that the meetings allowed for open interactions (86.7%; 13/15), were collaborative (73.3%; 11/15), and encouraged the sharing of diverse perspectives (40.0%; 6/15). Participants described the research team as receptive to the feedback offered (93.3%; 14/15) and commended the open process of making live changes to the curriculum as part of the group virtual meetings (20.0%; 3/15). “They shared the screen on the virtual platform we were meeting on so as we went through every piece, we were all looking at the same thing together as we gave notes and feedback… We definitely had a voice in things… and it did feel like suggestions that we had and concerns that we had were taken very seriously… seeing this feedback being typed out on the screen in real-time [helped me] see my concern, I [could] see my opinion right there… that was definitely beneficial from our perspective.”
All members noted that they would have rather met in person versus virtually to allow for maximum engagement in the adaptation process (information was collected during the height of the COVID-19 pandemic). Participants also suggested that the Adaptation Teams could have benefited from further inclusion of community members (26.7%; 4/15), especially parent or migrant families who have previously participated in home visiting services (6.7%; 1/15). Further, participants recognized that not all sects of each ethnic group were represented, especially for the Afghan community (the agency family service provider is from Iran and speaks Farsi, which is a dialect of the targeted language, Dari; 6.7%; 1/15), and that future processes could benefit from broader inclusion.
Discussion
Exposure to violence and extreme adversity prior to, during, and after experiencing displacement and forced migration has a profound effect on the mental health and behavioral outcomes of refugee children (Fazel et al., 2012). Positive parenting programs among refugee populations may be a potential way to buffer the negative effects of experiencing extreme adversity and bolster post-migratory protective factors for both children and adults (Eltanamly et al., 2021). Increasing access to culturally and linguistically relevant parenting interventions among underserved and diverse families like refugees is one way to promote health equity and social justice, and enhance the lives of those who have experienced extreme hardship before, during, and after their forced displacement and migration (Parra-Cardona, López-Zerón et al., 2017). To our knowledge, this is the first study to document the structured adaptation process of an evidence-based parenting program for a diverse community of refugees resettling in the United States, using community-based participatory methods focused on the first two stages of Barrera and Castro’s (2006) Heuristic Framework for the cultural adaptation of interventions: (1) information gathering and (2) preliminary adaptation design (Barrera & Castro, 2006).
In the information gathering phase, the majority of the Adaptation Team members, all of whom had cultural expertise or extensive experience working with refugee populations in Clarkston, indicated that they would modify SafeCare to fit the needs of their clients by removing aspects, integrating elements from other interventions, modifying the delivery setting, and tailoring or tweaking SafeCare. By creating cultural and linguistic adaptations to EBPs aimed at program efficacy and cultural relevance at the client- and provider-levels, EBPs can respond to the preferences and needs of diverse populations (Cabassa & Baumann, 2013) while improving intervention recruitment, retention (Kumpfer, Alvarado, Smith, & Bellamy, 2002), and effectiveness (Griner & Smith, 2006; Kumpfer et al., 2002). Notably, several Team members initially indicated that some aspects of SafeCare might not apply to the communities they serve; however, over the course of the adaptation meetings, curriculum edits, and qualitative interviews, the key adaptations that emerged were relatively minor in scope, focused on the presentation and ordering of certain materials, or supplements or exclusion of language and materials to ensure cultural relevance. No changes were recommended for SafeCare drivers of behavioral change, such as target parent skills or the approach to parent training (i.e., explain, model, and practice feedback approach).
The Adaptation Teams recommended 113 Process and Content adaptations to ensure literacy of the materials and to better align SafeCare PCI with the cultural values and practices of the refugee ethnic groups. Process adaptations suggested were mostly focused on the presentation of materials and included rephrasing or simplifying explanations and terms to accommodate literacy levels. Many of the content adaptations offered a more specific explanation of the concepts or activities and replaced U.S.-based cultural references and play activities. These adaptations are consistent with recommendations in a meta-analysis conducted by van Mourik and colleagues (2017), which found that EBPs adapted for delivery with a different cultural group that included adaptations to better address such social, cultural, and environmental factors were more effective at improving parenting behavior among the population of focus compared to those that focused on more simple adaptations, such as translation alone (van Mourik, Crone, de Wolff, & Reis, 2017).
Other content adaptations intended to orient caretakers to safety practices that are common or are laws in the United States. Examples of these adaptations included adding information about car seat use, safe sleep practices, and mandated reporting laws. Such laws and/or societal norms are critical for refugees and immigrants to be aware of but may not be in place in their home countries (Greenberg et al., 2019). Often parenting programs like SafeCare assume some base-level knowledge of U.S. laws and social norms, and thus content around those laws and norms is not included. This may represent an important gap in the parenting programs for immigrants and refugees as it is important for parents to be aware of these laws or practices and to know what actions constitute child maltreatment.
In compiling the feedback from the PCI-Adaptation groups as part of the preliminary adaptation design (Phase 2), we found that the suggested adaptations were similar across and relevant to all refugee ethnic groups and were largely focused on the presentation of content and the inclusion or exclusion of similar types of information. Further, the suggested adaptations did not alter the core SafeCare elements, target skills, or the theoretical drivers of behavior change. Thus, the SafeCare experts determined that one major curriculum overhaul would be sufficient with language translations for each ethnic group to follow. This finding is consistent with prior research documenting similarities in parenting norms and skills (Leijten, Melendez-Torres, Knerr, & Gardner, 2016; Parra Cardona et al., 2012; Kaminski et al., 2008) and intervention principles (i.e., increasing positive parent–child relationships and modification of parent and child behavior) across cultures and contexts (Devlin, Wight, & Fenton, 2018; Gardner et al., 2016; Leijten et al., 2016, 2019; Mamauag et al., 2021).
During qualitative interviews, Adaptation Team members reported that they felt the adapted materials were culturally appropriate for the ethnic groups of focus and ready for Stages 3 and 4 in Barrera and Castro’s Heuristic Framework in the implementation trial (Stage 3: preliminary adaptation tests and Stage 4: adaptation refinement in the randomized trial; Barrera & Castro, 2006). Most members mentioned that the simplified language and culturally appropriate pictures would help with parent comprehension and acquisition of target skills. Language modifications and visual aids have been used in other evidence-based parenting programs that have been adapted for delivery in different cultural contexts (McCoy et al., 2021). Respondents also indicated that they thought the adaptations were responsive and sensitive to the unique experiences of refugees and would help create some level of familiarity for parents and children who, as refugees, are now required to adapt to and navigate an unfamiliar context. Lastly, the Adaptation Team members noted during the qualitative interview that they felt their expertise and suggestions were valued throughout the adaptation process. Collective empowerment and internal power-sharing are key components of community-based participatory research methods and may be one strategy for advancing social change and promoting health equity (Sánchez et al., 2021; Wallerstein et al., 2020).
Limitations
This study had several limitations. First, the SafeCare Modification Survey did not include open-ended responses. Allowing participants to explain their responses to the Likert scale items would have given us valuable information about participants’ initial impressions for modification needs. Second, the SafeCare Modification Survey and Participant Satisfaction Survey have not been previously validated. However, the Modification Survey has been used in prior research (Meza et al., 2020) and is based on a published coding system (Wiltsey-Stirman, Miller, Toder, & Calloway, 2013). Further, because of the small number of participants who completed the SafeCare modification survey, we were unable to assess the scale for internal reliability in our sample. However, the overall purpose of using this scale was to understand the general impressions from our community partners about whether adaptations to the SafeCare curriculum and implementation process were needed. We learned through this initial survey and throughout the adaptation process reported in this paper that adaptations were needed.
Third, there was a relatively low response rate for the post-PCI-Adaptation group satisfaction surveys; across all group meetings, only an average of 45% of the surveys were completed. Thus, we cannot ascertain any information about the experiences of those who did not complete the surveys. The low response rate may have been because the satisfaction surveys were sent out after each meeting (5 meetings per ethnic group) and may have been burdensome to the Adaptation Team members, who took part in this adaptation process in addition to their normal job roles and responsibilities. Further, the adaptation meetings met during the height of the COVID-19 pandemic in 2020; thus, the increased social and occupational pressures, uncertainty about the future, and physical distancing may have also played a role in the low response rate.
Fourth, the suggested adaptations were made for four refugee ethnic groups who were living in Clarkston, Georgia. These adaptations may not reflect the needs of refugees from other ethnic groups or refugees living in different contexts. It should be noted, however, that the purpose of this adaptation process was to adapt SafeCare for implementation within the context of a research study, which will be conducted in Clarkston, Georgia (Whitaker et al., 2021). Thus, our aim was to develop adaptations to SafeCare that would be relevant in the populations that would be participating in the research trial and not necessarily generalize these preliminary findings to other refugee ethnic groups or refugees living in different cultural contexts (i.e., living in another city in the United States or another country). Fifth not including parents in the adaptation groups is a primary limitation of the current study. However, given this work was conducted in the midst of the COVID-19 pandemic, with stay-at-home orders and social distancing restrictions in place, it was challenging to consider the best ways to engage parent participants, especially with limitations on accessible technology and language barriers. Sixth, analyses did not allow for specific exploration of outcome by gender, age, or racial/ethnic group variations in responses.
Lastly, further testing of the effectiveness of this intervention in a randomized controlled trial is necessary to complete the adaptation process according to Barrera and Castro’s Heuristic Framework to confirm whether this adapted program is effective at improving positive parent–child relationships and other SafeCare target outcomes among the refugee ethnic groups of focus. The current adaptation contributes to an implementation study that will examine how SafeCare Provider delivery methods (delivery by typical agency-based human service provider [standard SafeCare] vs. delivery by a community member [task-shifted SafeCare]) influence implementation and outcomes.
Conclusions and Future Directions
Throughout this community-based adaptation process of an EBP, we noted several recommendations for future adaptations that add to the scant information on this topic currently available in the literature (as noted by van Mourik et al., 2017). First, we felt that Barrera and Castro’s (2006) Heuristic Framework provided valuable structure to our adaptation process (Barrera & Castro, 2006), and their later work provides further support for this staged approach to adaptations (Barrera et al., 2013). In addition to providing structure, Barrera and Castro’s framework emphasized the inclusion of potential participants and experts who are experienced in working with the targeted groups during the adaptation process. This bottom-up approach to the cultural adaptation process, or the active engagement and input of community members, has been emphasized in other studies documenting cultural adaptations of parenting programs (Mamauag et al., 2021).
Second, the Adaptation Team members did indicate that, in the future, it would be important to include refugee and migrant parents in the adaptation process. The Adaptation Team included service providers from refugee resettlement agencies and community health workers, most of whom were members of the targeted ethnic groups; they were expected to represent the target population for the intervention. Prior research suggests that community participation, understanding the community, and community support are vital factors that impact program sustainability (Ceptureanu, Ceptureanu, Luchian, & Luchian, 2018). By including the target audience in the adaptation process, we can better capture the diversity in the voices of the community and include the people for whom this program is designed in the adaptation process. Further, when exploring adaptations to evidence-based programs, it is important to recognize there is heterogeneity within cultural and ethnic groups (Ford & Harawa, 2010). This was a concern voiced by the Adaptation Team during the interviews, specifically regarding the diversity between sects in the Afghan community. Future researchers should account for the diversity within each cultural or ethnic group.
Third, the Adaptation Team members indicated that they would have preferred to have met in person versus via virtual methods (video conference). While the COVID-19 pandemic restricted our meeting format, we suggest that future researchers consider holding adaptation meetings in person if it is safe to do so. Despite these uncontrollable changes to our approach, participation rates across the PCI-Adaptation group meetings were high, suggesting the adaptation process was acceptable. The format of both the curriculum informational meeting and the PCI-Adaptation group meetings allowed for an open, collaborative dialog in that there was adaptability and openness among collaborators and stakeholders. Prior research documents that continuous dialog, openness, and adaptability among collaborators and stakeholders are key ingredients in the success of cultural adaptation and transportation of EBPs to new contexts (Mamauag et al., 2021). Thus, we recommend that future researchers focus on cultivating an open, collaborative environment to value and elevate the voices of stakeholders and community members. Another way to value stakeholder and community member contributions to research is to compensate them for their time commitment and contribution. In this project, agencies were compensated for the time of their staff, and community health workers were compensated directly for participating.
Lastly, we echo the recommendations by Mejila and colleagues (2017) that a detailed reporting process for adaptations made to EBPs, similar to CONSORT (Schulz et al., 2010) guidelines for reporting quantitative findings, is necessary (Mejia et al., 2017). This uniform reporting could guide the field of dissemination and implementation science and public health as the fields move to increase equity and to advocate for social justice by implementing EBPs that are responsive and sensitive to the cultural and ethnic diversity of the people for which these programs are intended. Ultimately, these detailed processes reported in the literature could inform the scale-up dissemination of EBPs to increase the reach of evidence-based prevention strategies to a wider audience.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This publication was supported by Grant or Cooperative Agreement number U48DP006393, funded by the Centers for Disease Control and Prevention. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the Centers for Disease Control and Prevention or the Department of Health and Human Services.
