Engaging stakeholders to identify implementation strategies for a faith-based physical activity program

Background: Stakeholder engagement is critical to the acceleration of evidence-based interventions into community settings. Harnessing the knowledge and opinions of community stakeholders increases the likelihood of successful implementation, scale up, and sustainment of evidence-based interventions. Faith in Action ( Fe en Acción ) is an evidence-based promotora -led physical activity program designed to increase moderate-to-vigorous physical activity among churchgoing Latina women. Methods: We conducted in-depth interviews using a semi-structured interview guide based on the Consolidated Framework for Implementation Research (CFIR) at various Catholic and Protestant churches with large Latino membership in San Diego County, California to explore barriers and facilitators to implementation of a Faith in Action and to identify promising implementation strategies for program scale-up and dissemination. We interviewed 22 pastors and church staff and analyzed transcripts using an iterative-deductive team approach. Results: Stakeholders described barriers and facilitators to implementation within three domains of CFIR: characteristics of individuals (lack of self-efficacy for and knowledge of PA; influence on churchgoers’ behaviors), inner setting (church culture and norms, alignment with mission and values, competing priorities, lack of resources), and outer setting (need for buy-in from senior leadership). From the interviews, we identified four promising implementation strategies for the scale-up of faith-based health promotion programs: 1) health behavior change training for pastors and staff; 2) tailored messaging; 3) developing community collaborations; and 4) gaining denominational support. Conclusions: While churches can serve as valuable partners in health promotion, specific barriers and facilitators to implementation must be recognized and understood. Addressing these barriers through targeted implementation strategies at the adopter and organizational level can facilitate improved program implementation and lead the way for scale-up and dissemination. This research examines the barriers and facilitators to implementation of a faith-based promotora -led physical activity promotion program. The Consolidated Framework for Implementation Research (CFIR) was used to identify factors affecting program implementation of health promotion programs in faith-based settings. Findings provide specific suggestions on implementation strategies for scale-up of health promotion programs in faith-based settings.


Introduction
Despite the well-established evidence for physical activity (PA) in the prevention and control of cancer and other chronic diseases, approximately 80% of US adults and adolescents are insufficiently active. [1][2][3][4] In particular, Latinas only 13.6% of Latinas meet the PA guidelines, compared to 22% of non-Latina white women. 5 Furthermore, Latinas engage in fewer total minutes of PA than Latino men (19 vs. 30 min/day). 6 Dissemination and implementation of evidence-based PA programs are needed to help mitigate health disparities among Latinas and other at risk groups. 7 Faith-based organizations (FBOs) provide a unique setting to implement evidence-based interventions (EBIs) and are increasingly recognized as important partners in Latino-focused health promotion efforts, including physical activity. [8][9][10][11][12] The strong connection between health messages and the spiritual mission of the FBO, in addition to its reach into underserved communities, further justify the church as an effective health promotion setting. 13,14 Studies have documented the effectiveness of health promotion programs in FBOs; however, significant gaps exist in translating EBIs into faith-based settings to have broader reach. 15,16 Understanding the contextual factors is a critical step in translating research to practice, particularly in community settings. 17 Engaging stakeholders in the process of translation can lead to more effective assessments of contextual factors influencing implementation, and therefore more effective translation of behavioral interventions. 18 Faith in Action ( Fe en Acción) is a faith-based multi-level promotora-led intervention promoting moderate-to-vigorous PA (MVPA) among Latina women through group PA classes and Motivational Interviewing calls. 19,20 Its effectiveness was tested using a cluster randomized controlled trial in 16 churches. 21 At 12 months, there were significant increases in accelerometer-based MVPA and self-report leisure-time MVPA among Latinas in the intervention versus comparison condition, which suggests a greater increase in PA than many other PA interventions. 22 Intervention participants, compared to those in the attention-control condition, had a 66% higher odds of meeting the PA guidelines, reduced BMI, and used more behavioral strategies for engaging in PA. Attendance to PA classes was associated with increased self-report leisure-time MVPA; number of Motivational Interviewing calls was associated with meeting the PA guidelines. There were approximately 25,000 attendees in 8 intervention churches over the course of the 5-year study. Efficacy-effectiveness trials have promoted PA in FBOs, but Faith in Action is the only one that has focused on Latinas. [23][24][25][26] Because the outcomes paper focused on participant outcomes, little is known of how the organizational context contributed to the success (the "how" and the "why").
Understanding the organizational context is critical for scale-up and dissemination of evidence-based programs. 27 Midway through the implementation of Faith in Action, we collected data examining implementation outcomes of a subset of churches participating in the physical activity intervention. 28 The findings suggest that pastor support, innovation-values fit, and resource availability were factors that impacted implementation effectiveness (average 6-month participation rates in PA classes at each church). In addition, churches with lower parishioner engagement had low support from church staff and leaders while churches with higher parishioner engagement reported high pastor support, high innovation-values fit, medium to high recourse availability, and medium to high parishioner engagement. 28 While these findings shed light on the organizational-level factors affecting implementation of Faith in Action, more targeted studies are needed identify determinants that impact implementation and implementation strategies that target these determinants among a wider range of religious denominations for wider scale-up and dissemination.
While churches are promising venues for health promotion programs, [29][30][31][32] including PA interventions, 8,11,[33][34][35][36] the lack of understanding of organizational context and determinants for implementation and sustainment limit dissemination capability for these programs. Few faith-based health promotion studies have been taken to scale and 37,38 most have targeted African-American populations; 37,39,40 and only a few have targeted PA as a main outcome, with limited effectiveness. 8, 35 Webb and colleagues explored faith leaders' perceptions of health and wellness; however, their sample was entirely Caucasian and majority Methodist, and did not focus on the implementation of a particular program. 41 Finally, few faith-based studies have systematically explored the perspectives of stakeholders regarding barriers and facilitators to implementation. Bernhart and colleagues administered surveys to pastors of churches participating in the Faith, Activity, and Nutrition (FAN) program, however the sample was comprised of churches with predominantly African-American membership. 42 The authors acknowledged the limitations of survey data and reported that in-depth interviews would have further elucidated the findings, explaining the mechanisms behind the barriers and facilitators to implementation.
The objective of this study is to examine factors associated with the implementation and sustainment of an evidence-based PA intervention (Faith in Action) through stakeholder interviews and identify implementation strategies specific to the church context to inform future program scale-up and dissemination.

Study Design
We conducted a qualitative study consisting of 18 semi-structured interviews (n=10 at six Catholic Churches and n=8 at eight Protestant Churches) with 22 stakeholders to identify barriers and facilitators to implementation and sustainment of an evidence-based, faithbased PA intervention in Latino churches. The study received approval from the Institutional Review Board at San Diego State University.

Theoretical framework
The semi-structured interview guide was developed using the Consolidated Framework for Implementation Research (CFIR). 43 This framework was chosen because it is comprehensive, examines domains influencing implementation effectiveness, and is well suited for complex, multilevel interventions. 44,45 Open-ended questions assessing all five CFIR domains (intervention characteristics, inner setting, outer setting, process, and characteristics of individuals) were adapted from those presented in the online interview guide tool (cfirguide.org) to assess implementation and the domains from Schell and colleagues 46 to assess sustainability. While the domains and constructs included in the CFIR are more oriented towards a healthcare setting, there are few options of frameworks aimed at examining best practices for health promotion in faith-based settings. 47 The questions were adapted to fit non-clinical settings, simplifying language to make it relevant for community-based settings and incorporating appropriate Catholic or Protestant vocabulary (e.g., parishioner, churchgoer, priest, pastor). A copy of the interview guide has been included in the supplemental files (see Additional file 1).

Sampling and Data Collection
To gather data on both the previous implementation of Faith in Action and potential for dissemination and scale up, we recruited pastors and staff from the 8 Catholic churches that participated in the Faith in Action intervention plus 10 Protestant churches in San Diego County serving large Latino congregations. Protestant is the second most common denomination for churchgoing Latinos. 48 These other churches were identified using online search engines (search terms "San Diego" and "Spanish church" or "iglesia") and compiled into a database of Protestant churches in San Diego County with at least one Spanish-language weekend service (n=53). We used phone, email and mail for initial contact and follow-up, stopping at 5 contact attempts per church. In addition, research staff visited 10 churches in-person to solicit interviews, prioritizing bilingual churches with large congregations and monolingual Spanish churches. This process continued until we recruited 10 churches that represented a variety of denominations within the larger category of Protestant (i.e., Evangelical, Baptist, Seventh-Day Adventist, nondenominational), demographics, and contexts to ensure variability across sites. During recruitment we described to potential participants the purpose of the research in identifying barriers and facilitators to implementation of health programs in churches.
Twenty-two pastors and staff agreed to participate.

Data collection
The majority of the interviews (n=13) were conducted by JH (a female MPH/MA researcher with experience implementing faith-based interventions) and EA (a female PhD health behavior research). The remaining 5 were conducted by MT (a female PhD researcher with extensive qualitative experience) and JS (a female graduate student with training in qualitative methods). Several co-authors (JH, EA, MT, JS) trained in qualitative methods, conducted the interviews in pairs. Interviews were conducted in-person at the churches in either English or Spanish according to the interviewees' preference and lasted an average of 45 minutes. In some cases, more than one individual participated in an interview, resulting in 18 interviews with 22 total stakeholders. Interviews were audio-recorded, transcribed verbatim, and reviewed by the interviewer to ensure accuracy. Field notes were collected during the interview and added to the data file. Identifying information such as pastor names were removed or abbreviated and Spanish transcripts were translated to English using standard protocols. 49 To ensure all aspects of the qualitative research was reported, the consolidated criteria for reporting qualitative studies (COREQ) 50 was used as a checklist (see Additional file 2).

Data Analysis
Data were analyzed and coded using an iterative-deductive approach. MT and JH coded the first two interviews to develop preliminary codebooks guided by CFIR constructs.
Transcripts were independently coded to categorize data, then discussed in person to reach consensus. [51][52][53] The initial interview codebook was used by the larger research team (JH, MT, JS, and 6 students) to code each subsequent interview transcript. Groups of three team members, including at least one of the lead researchers (JH, MT, JS), coded each transcript independently and then met in person to discuss discrepancies until consensus was reached. As necessary, the team revised the codebook, adding codes that emerged throughout the process and refining definitions in the codebook for greater clarity. 54 All changes were discussed and agreed upon in weekly team meetings to further ensure mutual understanding. Furthermore, consistency was ensured by repeated rounds of re-coding the same transcript after clarifying differences. After coding all transcripts, we identified sub-codes to categorize and re-categorize codes into relevant relationships. 55 A detailed audit trail of codebook drafts, coded and re-coded transcripts, meeting notes, codebook edits, and resolved discrepancies was kept throughout the process. We coded all transcripts by hand and then inputted the coded data to Dedoose software (SocioCultural Research Consultants, LLC version 7.5.9) to organize and sort the data.
Once the data was coded and organized, JH reviewed and examined the codes, merging similar codes and separating broader ones, sorted them into themes, and selected key quotes that represented the main themes. A group discussion involving EA and MT finalized the main themes and identified the salient barriers and facilitators (see Table 1).
After identifying the barriers and facilitators to implementation and sustainment, JT and EA used the ERIC strategies 56 to link and identify mechanisms of action for each implementation strategy. These were then validated by BR, a co-author with expertise in implementation science (see Table 2).

Description of Study Participants
Eighteen semi-structured interviews were conducted with 22 pastors and staff between April and August 2018 out of 61 churches invited. Differences were not observed between interviews with one participant and those with two participants. Most interviews were conducted in English (n=14) and the remainder in Spanish (n=4). Seventeen of the interviewees were male and 5 were female. Of the 22 stakeholders who participated in the interviews, all held formal positions within their churches: 11 were senior pastors or priests, 5 were associate or assistant pastors, and 6 were paid church staff members (business managers, secretaries, and deacons). One interviewee (assistant pastor) declined to be recorded and was not included in the final analysis.

Barriers and Facilitators to Implementing Faith in Action
The barriers and facilitators identified fell within three of the five CFIR domains (i.e., characteristic of individuals, inner setting, and outer setting). We summarized major findings according to CFIR domains and included seminal quotes as illustrations of each (Table 1).

Characteristics of Individuals
Pastors and staff lack self-efficacy for and knowledge of PA. In general, pastors and staff reported having low self-efficacy for PA and lacked knowledge of PA. While a few mentioned having personal PA habits (e.g., cycling, running), most reported struggling to find time to be active and unsure of how to encourage churchgoers to be active. One pastor described his hesitancy in encouraging PA, "Well probably I'm not the best person to talk about that because I need it badly. I lost already 24 pounds but I need to lose another 40 pounds." Some pastors mentioned that they would feel like "hypocrites" if preaching about health, given their own challenges in maintaining healthy habits. Pastors shared that their demanding and unpredictable work schedules contribute to unhealthy habits and a lack of priority of their own health. In addition, pastors' lack of self-efficacy for PA limits their ability to motivate healthy behaviors among churchgoers.

Pastors have influence over churchgoers' behaviors.
Church staff felt that pastors who were roles models for PA and healthy eating had a strong influence on churchgoers. Some pastors said they had opportunities to counsel churchgoers, but while they felt equipped in mental health counseling, they were sure about how to best encourage individuals to be physically active. One pastor described his influence over his congregation in the following way, "If I open my mouth, they'll hear. They'll listen."

Inner Setting
Churches can support a culture of overeating and unhealthy behaviors. Pastors and church staff described the important role of food in fellowship in the church.
Stakeholders stated that food is often served at churches events, and lacks healthy options. One pastor told us, "There's a little saying in our Spanish 'Don't trust a skinny Pastor'" when describing the gifts of food given him by churchgoers.
Churches implement programs aligned with their mission and values. While many of the pastors and staff we interviewed acknowledged the importance of health and the need for more PA opportunities for their members, many emphasized the need for any program to align with the church's mission and values. Given that a church's mission is to build faith and share the message of God, finding a way to frame PA within that mission is essential. Once pastor said, "I think our mission is to help people to have good health physically, mentally, emotionally, and spiritually. So, that is a complement to our mission." Pastors did see an opportunity in connecting physical health with spiritual and mental health, as one shared, "the church is sometimes too spiritual, but the people do not die due to the spiritual, instead it is physical. They get sick because of physical health." Finally, pastors had the view that any new program, to be successful, would have to "nestle itself into the homeostasis of the church, so it just becomes natural."

Church culture and norms are influenced by pastors and staff. A few pastors
reported having spoken about physical health from the pulpit, however most reported not having the knowledge or resources to do so. One pastor shared, "that's what I gotta get my hands on, some of these templates that tie in the scriptures with you know healthy eating and exercise and stuff like that." Churches have many competing priorities. While pastors and staff recognized the importance of physical health and acknowledged the potential role of the church in promoting health among churchgoers, they also expressed that churches have many other important priorities. Pastors talked about kids and youth programming, marriage counseling and retreats, spiritual formation, social justice campaigns, and many other ongoing programs. In addition, pastors in more urban and low-resource communities shared about their challenges getting churchgoers to be involved in and committed to church programming given the transitional nature of their communities.
Churches often lack sufficient space and personnel. Many of churches included in this study have small programming budgets, limited space, and few paid personnel.
Pastors expressed concern about adding more to already overburdened staff and volunteers, saying, "the people who kind of are the natural leaders are already pretty spread out." Most were also concerned about the cost of program implementation and said they have little funds to spend on programming outside of faith formation activities.
Others were concerned that a new program may have to compete with current programs for space. While no pastors said they would not accept an outside program, many expressed hesitancy in hosting a program from an outside organization and said they would want to hear from other pastors about their experience with Faith in Action before moving forward.

Outer Setting
Denominational support is critical for program success. In this study, pastors noted the importance of support from upper denominational leadership. Pastors of Catholic churches mentioned the importance of buy-in from the local diocese. Those in Protestant churches saw denominational support as an opportunity to disseminate the program more widely to other churches.

Promising Implementation Strategies for Faith-Based Health Promotion Programs
The following four organizational-level implementation strategies are recommended to engage pastors and church staff to improve implementation of faith-based health promotion programs. While these strategies are tailored to faith-based settings, each is based on ERIC implementation strategies 56 as noted in Table 2. We also identified the proposed mechanism of action to explain how we linked barriers and facilitators to proposed implementation strategies.

1.
Health Behavior Change Training for Pastors and Staff. Pastors described a lack of self-efficacy for PA and that they would feel like "hypocrites" if preaching about physical health and wellness given their lack of knowledge. They reported not know much about PA recommendations and cited very little if any education or training in health. Pastors shared that their demanding and unpredictable work schedules contribute to unhealthy habits and a lack of priority of their own health, statements that support the findings of Proeschold-Bell and McDevitt. 57 The stress and burnout pastors and church staff experience can lead to negative health consequences. 58 However, pastors have a strong influence over churchgoers who look to pastors as role models. 41,58,59 Some spoke about their role as counselors, particularly with issues related to mental health, but admitted lacking confidence in counseling churchgoers on physical health. Our findings support the literature describing the strong influence pastors have over churchgoers' behaviors. 58,59 Finally, pastors and church staff have a strong influence over the church norms and culture, which can, at times, support a culture of overeating and unhealthy behaviors. 60 This is supported in the literature where a number of studies have found higher rates of chronic disease among faith leaders. A national study of various denominations found higher rates of physical inactivity, obesity, poor dietary habits, and chronic disease among faith leaders compared to the general US population. 61 Social interactions around food have long been an important part of church culture, and often these foods are high-fat and not nutritious. 60,62 While churches can have cultures that are resistant to change and act as barriers to healthy behaviors, there is promising research that churches providing instrumental support for PA through church-based PA programs see increases in PA among members. 36 As important decision-makers in the church, pastors and staff have the ability to adopt programs, align them with their vision, and influence the culture and norms of the church. 14 Pastors and church staff have the potential to influence church culture and norms in ways that promote health. 59 Given these findings, we suggest that faith-based health promotion programs include training and education on health behavior for pastors and influential church staff be included in the intervention, even if the pastors and church staff are not the primary program implementers. We expect that if pastors and church staff are trained in health behavior change methods, they will have increased self-efficacy for PA, will be able to role model healthy behaviors including PA, will encourage churchgoers to be active and healthy, and will influence the norms and culture of the church to shift to include health-promoting policies (e.g., water to replace sugar-sweetened beverages at church events, establishing a health ministry).

Limitations & Strengths
While the sample size meets evidence-based guidelines, 73

Availability of data and materials
The data that support the findings of this study are available from the corresponding author upon reasonable request.

Authors' contributions
JH, EA, and MT conceived the papers' aims and drafted and edited the manuscript. JH, MT, JS, and EA conducted the interviews and analyzed the qualitative data. EA, BR, SH, and RB reviewed earlier versions of the paper and provided critical feedback. All authors read and approved the final manuscript.