Promoting Health from the Pulpit: A Process Evaluation of HIV Sermons to Reduce HIV Stigma and Promote Testing in African American and Latino Churches

Abstract

Embedding health messages into sermons is a potentially valuable strategy to address HIV and other health disparities in churches that predominantly serve racial and ethnic minorities. This study explores implementation of an HIV sermon as part of a multi-component intervention in three churches (Latino Catholic, Latino Pentecostal, and African American Baptist) in high HIV prevalence areas of Los Angeles County, California. Clergy were given an HIV sermon guide that included local public health data, stigma reduction cues, HIV testing messages, and a sample sermon. Findings are based on a process evaluation (i.e., reach, dose delivered, fidelity, and implementation) and in-depth content analysis to explore HIV frames and messages used by clergy. Sermons were audio-recorded, transcribed verbatim, and coded using an inductive approach. Complementary data were collected through systematic observation. Overall, five clergy delivered nine HIV sermons to majority African American or Latino audiences. On average, 174 congregants were reached per sermon. We found large variation in fidelity to communicating key HIV messages from the sermon guide. While promoting HIV testing from the pulpit seemed viable and acceptable to all the participating clergy, fewer embedded explicit stigma reduction cues. Most spoke about HIV using compassionate and non-judgmental terms, however, issue framing varied across clergy. Structured training of clergy may be necessary to implement the more theoretically-driven stigma reduction cues included in the sermon guide. More research is needed on the viability and acceptability of embedding specific health promotion messages into sermons.

Introduction

HIV-related mortality and morbidity rates are higher among African American (AA) and Latino populations compared to other racial and ethnic groups. HIV prevalence rates are estimated to be highest among AAs (1,018.1 per 100,000) and Latinos (350.8 per 100,000) compared to other groups (Centers for Disease Control and Prevention, 2015). The U.S.’s multi-pronged national HIV/AIDS prevention strategy includes reducing health disparities by collaborating with non-governmental partners, such as faith communities, to increase HIV testing and reduce stigma and discrimination (National AIDS Policy, 2015). Engaging AA and Latino faith-based organizations and clergy to communicate HIV prevention and stigma reduction messages is potentially important given the influence and authority of these institutions and faith leaders in minority communities (Nunn et al., 2013; Sutton & Parks, 2013).

In recent years, there have been a growing number of faith-based HIV interventions, nearly all in AA churches (Abara, Coleman, Fairchild, Gaddist, & White, 2015; Berkley-Patton et al., 2010; Berkley-Patton et al., 2016; Francis & Liverpool, 2009; Griffith, Campbell, Allen, Robinson, & Stewart, 2010; Lindley, Coleman, Gaddist, & White, 2010; Wooster et al., 2011). Many barriers to HIV programs in faith-based settings exist, including HIV-related attitudes and stigma, low HIV knowledge, lack of experience with HIV programming, lack of lay leadership and congregant support, and financial barriers (Coleman, Lindley, Annang, Saunders, & Gaddist, 2012; Cunningham, Kerrigan, McNeely, & Ellen, 2011; Derose et al., 2014; Mendel et al., 2014; Pichon, Powell, Ogg, Williams, & Becton-Odum, 2016; Pryor, Gaddist, & Johnson-Arnold, 2015; Smith, Simmons, & Mayer, 2005; Williams, Pichon, Latkin, & Davey-Rothwell, 2014; Wooster et al., 2011). AA faith leaders have also expressed concerns about HIV prevention strategies (e.g., discussing condoms and sexual education) that may be viewed as conflicting with doctrine (Nunn et al., 2012).

Despite these challenges, integrating HIV-related messages into clergy-delivered sermons is a potential health communication strategy to address health disparities because clergy members can increase the appeal and credibility of a health message and promote positive health behaviors among congregants (Lumpkins, Greiner, Daley, Mabachi, & Neuhaus, 2013). Clergy wield interpersonal influence and can shape congregants’ attitudes through their moral authority and visibility in the church (Baruth, Bopp, Webb, & Peterson, 2015; Campbell et al., 2007; Derose & Kanouse, 2011; Miller & Rubin, 2011). Previous multi-level HIV intervention studies have incorporated HIV sermon components (Abara et al., 2015; Berkley-Patton et al., 2010; Berkley-Patton et al., 2016; Griffith, Campbell, et al., 2010; Griffith, Pichon, Campbell, & Allen, 2010; Kaplan, Calman, Golub, Ruddock, & Billings, 2006). However, these studies do not describe the specific content or health messages in detail, nor do they analyze how sermons were delivered. An exploration of the content of this health communication strategy is needed (Muturi & An, 2010), especially on how clergy frame HIV and their openness to interventions that shape their framing of HIV (Derose & Kanouse, 2011). A detailed evaluation of how this works in the real world across a range of clergy members and congregational types is important for understanding whether enlisting clergy to deliver an HIV sermon is feasible and acceptable. Documenting implementation of intervention components is also important toward better understanding its reach, the social milieu of a setting, and the manner in which the content is delivered to inform future program implementation (Baranowski & Stables, 2000).

This paper evaluates implementation of an HIV sermon in three intervention churches in Los Angeles County representing diverse race/ethnic groups and denominations: a Latino Catholic Church, a Latino Pentecostal Church, and an AA Baptist Church. Clergy from these churches were asked to deliver an HIV sermon at their weekly worship service(s), coinciding with or leading up to a church-based HIV testing event. Other components of the intervention were HIV education and peer leader workshops for congregants (Flórez, Payán, Derose, Aunon, & Bogart, 2017) and on-site HIV testing events conducted by a local health department (Williams et al., 2016). The HIV sermon was hypothesized to help reduce HIV stigma by promoting compassion for people living with HIV (PLHIV) and to increase HIV testing among congregants through clergy’s social influence. Specifically, the HIV sermon aimed to leverage clergy’s authority and reach within the church to promote positive norms around HIV, PLHIV, and HIV testing. A clergy’s framing of these issues from the pulpit can influence how congregants think about a health issue and behave. For instance, HIV frames that clergy could possibly use include HIV as punishment for sin, HIV as a call to compassion, or HIV as an opportunity for transformation (Derose & Kanouse, 2011), which could contribute to congregants’ perceptions of and attitudes towards PLHIV. To our knowledge, this is the first HIV stigma reduction intervention that has been evaluated in both Latino and AA churches and the first process evaluation of a sermon component of a church-based health intervention. Herein, we use process evaluation measures and an in-depth qualitative content analysis approach to explore clergy members’ willingness to implement an HIV sermon guide and to communicate specific stigma reduction and HIV testing messages to their congregants.

Method

A process evaluation was conducted to evaluate the extent to which various components were implemented at each intervention church. Key evaluation constructs of interest included reach, dose delivered, fidelity, and implementation (Linnan & Steckler, 2002; Yeary, Klos, & Linnan, 2012). Evaluation constructs and questions related to the sermon component are in Table 1.

Table 1.

Process Evaluation Constructs and Questions related to the Sermon Component

Construct	Process Evaluation Question
Reach	1. How many congregants were exposed to the sermons?
	2. Who heard the sermons?
Dose Delivered	3. To what extent did clergy participate in the intervention?
	4. Did clergy deliver the HIV sermon?
	5. How was the sermon delivered (i.e., language, style)?
	6. How long was the sermon?
Fidelity	7. Did clergy adhere to the objectives and strategies in the HIV sermon guide? Which ones did they adhere/not adhere to?
	8. How did clergy discuss and address HIV-related stigma?
Implementation	9. Did clergy encourage congregants to participate in any project activities?
	10. Were there any unintended consequences?

An in-depth qualitative content analysis approach was also employed to identify faith and HIV-related themes, explore HIV frames and messages used by clergy, and provide context for the process evaluation results.

Setting, Recruitment, and Pilot Outcomes

Six churches in and around Long Beach—a city with a relatively high cumulative AIDS incidence rate of 1,347 cases per 100,000 residents (Long Beach Department of Health and Human Services, 2013)—were invited to participate and matched on race-ethnicity, denomination, and congregation size. Five churches agreed, including two medium-sized AA Baptist churches, two small Latino Pentecostal churches, and a large Latino Catholic Church. Matched churches were randomized to the intervention or control; the Catholic Church was assigned to the intervention. Study design details are described elsewhere (Derose et al., 2014).

The multi-faceted, congregation-based intervention was developed collaboratively with faith and public health partners using community-based participatory research methods and extensive formative research (Bluthenthal et al., 2012; Derose et al., 2011). Analyses found the intervention decreased HIV stigma and mistrust in the Latino intervention churches (p<0.05) but not in the AA church. HIV testing increased across AA and Latino intervention churches compared to their matched control pairs (<0.001). Surveys of all congregants at baseline (73% response rate) and 6-month follow-up (79% response rate) were used to assess changes in HIV stigma, mistrust, and testing. A detailed description of the study outcomes and related measures (i.e., 12-item HIV stigma scale, 7-item HIV-related mistrust scale, and binary testing measure) is available elsewhere (Derose et al., 2016).

HIV Sermon Description

The research team and study’s Community Advisory Board (CAB) co-chairs (M.A.M. and C.W.O.) jointly developed an HIV sermon guide with key talking points, a visualization exercise, and sample sermon to promote destigmatizing messages, HIV screening, and positive religious norms for the treatment of marginalized individuals. The sermon guide emphasized framing HIV as a health issue (instead of a moral issue or sin) using compassionate, non-judgmental language toward stigmatized groups (Derose & Kanouse, 2011) and raising awareness about how HIV is affecting their community and the importance of HIV testing. Table 3 lists HIV themes and specific objectives from the sermon guide.

Table 3.

Sermon Guide Themes, Objectives, and Fidelity by Church Leader^* and Example Quotes

Theme	Specific Objective	Pentecostal Pastor	Baptist Pastor	Catholic Pastor	Catholic Assisting Priest 1	Catholic Assisting Priest 2	Example Quotes
Project Awareness	1. Discuss why the church is involved	√	√	√	√	√	“Since this summer, we started in our parish the FAITH project to educate and sensitize our community about the HIV.”
	2. Mention recent project activities**	--	√	√	√	--	--
HIV Awareness	3. Describe how HIV is affecting the AA or Latino community	--	√	√	--	--	“The rate of diagnosis of HIV in Long Beach is the double of the rate in the County of Los Angeles and triple the rate of California. In Long Beach, the new infections of HIV are the majority among Latinos and African Americans.”
Stigma Reduction Cues	4. Share a personal story about PLHIV	√	--	√	--	--	“In Kenya, I knew several people infected with AIDS. How they were rejected by society…living alone, abandoned by their own people. I visited some, sharing their suffering and bringing them the consolation of faith. I had a nice group of catechists who visited them, bringing medicine and food but especially love, attention, and care.”
	5. Describe the effect of HIV-related stigma on PLHIV	√	--	√	√	√	“A lady, a catechist told me about her encounter with a person who was just infected with this disease. How his own family threw him out of the house…becoming now a homeless and rejected person from society and his own family.”
	6. Contrast the prejudicial treatment of PLHIV with the value of love and compassion	√	√	√	√	√	“We as a community are participating because we want to create love and compassion towards our brothers and sisters suffering from AIDS and to reduce the stigma and to create more awareness. Because HIV is a virus, AIDS is a disease and neither is a sin.”
	7. Implement the visualization exercise	--	--	--	√	--	“And if the person to your left or to your right has cancer, will you help that person? Will you listen to that person? Will you take their hand? And will you offer this person help? Now please imagine that the person to your right or to your left has AIDS or is infected with HIV. Will you still love this person? Will you listen to this person? Will you offer them help? And will you take their hand? Very good!”
HIV Testing Prompts	8. Mention the importance of HIV testing	--	√	√	√	√	“And if you are HIV [positive], you want to know as soon as what? Possible. Yeah. Amen. Because it’s treatable if you get it, right? If you catch it, right, it’s what? Treatable. But if you let it do what it do, then your date gonna move up. Amen.”
	9. Encourage congregants to undergo testing	--	√	√	√	√	“We’re trying to be a socially conscious ministry, a socially sensitive church. Our prayer is that you won’t leave today without participating in what we’re trying to do today.”
	10. Mention the HIV testing event date	√	√	√	--	--	“I want to remind you of FAITH Health ministries and [organization] in conjunction with the Long Beach Health Department. They’re here today for the second phase for the HIV testing program.”
TOTAL SCORE	5/10	7/10	9/10	7/10	5/10

^*Fidelity was assessed for the sermon delivered on the day when on-site HIV testing was held at each church.

^**Church-related events were mentioned during the service announcements, which were not audio-recorded.

Stigma reduction cues were hypothesized to reduce HIV-related stigma by addressing discomfort interacting with PLHIV and reducing blame for PLHIV (Herek, Capitanio, & Widaman, 2002). The visualization exercise (or imagined/hypothetical contact scenario) consisted of having congregants imagine a meaningful social interaction with someone living with HIV, inasmuch as simulated contact or interaction with someone with a stigmatized characteristic can help to ameliorate negative attitudes toward stigmatized groups (Blair, Ma, & Lenton, 2001; Crisp & Turner, 2009). CAB clergy co-chairs developed a sample sermon as an example of how to integrate the visualization exercise with Biblical references. The sample sermon referenced the Biblical passage of the parable of the Good Samaritan (Luke 10:25–37) which encourages individuals to “love your neighbor as yourself.” The sermon was pre-tested during a CAB meeting and members provided feedback before it was finalized. Further, the guide was professionally translated to Spanish and reviewed by bilingual investigators and staff.

Research team members met with the head pastor from each intervention church to review the sermon guide and address questions or concerns. Pastors were encouraged to tailor their sermon according to their style, congregation, and religious traditions to facilitate implementation and promote congruency between the goals of the study and church. Pastors were also encouraged to participate in on-site HIV testing to serve as role models for congregants by demonstrating the desired behavioral outcome (i.e., to get tested for HIV), in line with the Health Belief Model and Social Cognitive Theory, which suggest cues to action can influence congregants to undergo screening (Bandura, 1986; Champion & Skinner, 2008).

In the Catholic Church, the head pastor elected to share the sermon guide with two assisting priests who were also scheduled to deliver sermons during the week of the intervention. While the two priests did not attend the intervention training and only received information from the head pastor, the priests were included in the process evaluation since they were exposed to the HIV sermon guide and agreed to deliver an HIV sermon, thus providing us with an opportunity to evaluate sermons delivered by clergy who were trained by another clergy.

Data Collection and Measures

We used two data sources: (1) sermon transcripts and (2) systematic observations of religious services. Our multidisciplinary team and clergy consultants provided input and iteratively revised the procedures and tools, which were reviewed and approved by RAND’s internal review board. Sermons were audio-recorded and transcribed verbatim by a data transcription company. A systematic religious service observation questionnaire was adapted from a previous study (Derose et al., 2011; Mendel et al., 2014) and gathered qualitative and quantitative data about the service and attendees on intervention sermon days. The questionnaire included a 10-item checklist of sermon objectives, similar to checklists commonly used in process evaluations (Baranowski & Stables, 2000), to assess fidelity to key messages in the sermon guide. There was also an open-ended section for other information such as cues to action (e.g., if a pastor mentioned they had been tested for HIV). A trained research team member completed the questionnaire to avoid having churches self-report program data, inasmuch as the latter can be inconsistent (Kaplan et al., 2006) and burdensome.

Data Analysis

Sermon transcripts and observation data were imported into Dedoose, a web-based software platform to manage and analyze qualitative data (SocioCultural Research Consultants LLC, 2015). Transcripts were entered in their original language (English or Spanish) because all the team members involved in coding are bilingual. Open-ended (textual or narrative) observational data were also entered into the project file and coded. Quantitative process evaluation data from the observation questionnaire was entered into an Excel database and the results are summarized in Tables 2 and 3.

Table 2.

Reach and Dose Delivered Process Evaluation Results by Intervention Church

	Latino Catholic	Latino Pentecostal	African American Baptist	Overall
Average Attendance and Range (# of congregants per service)*	313	55	95	174
	220 – 405	--	80 – 110	55 – 405
# of Participating Clergy	3	1	1	5
# of Sermons Delivered
English	2	0	3	5
Spanish	3	1	0	4
Average HIV Message Length and Range (minutes)*	12.5	75	7.5	23
	8 – 17	--	5 – 10	5 – 75

^*Based on observational data collected at five church services.

The team employed a content analysis approach to analyze qualitative data (Altheide, 1996; Krippendorff, 2004). The initial codebook included relevant process evaluation codes (i.e., dose delivered, fidelity, and implementation), HIV frames/messages (i.e., HIV as punishment for sin, call to compassion, and opportunity for transformation), and faith-related themes (i.e., references to scripture, perceived role of the church, and motivation for involvement in health interventions). Themes listed in Table 3 were included as fidelity sub-category codes. The research team conducted a pilot coding session to test the initial coding structure and to identify emergent themes using an inductive open-coding approach (Strauss & Corbin, 1990). An emergent primary code consisted of HIV clergy knowledge and beliefs which included HIV knowledge, misconceptions, risk factors, and transmission as sub-category codes. Two investigators independently coded the data using the finalized version of the codebook. Coding discrepancies were resolved by consensus. Thematic summaries were developed based on coded content and were reviewed by two investigators. Example quotes in Spanish were translated to English for inclusion in this paper.

Results

Reach

Sermons delivered at the Catholic Church reached the most congregants. An average of 313 individuals attended each Catholic Church service (range: 220 to 405 congregants per service) compared to an average of 95 individuals per service at the Baptist Church (range: 80 to 110 congregants per service). Reach at the Pentecostal church was the smallest—only 55 congregants were in attendance when the sermon was given. Table 2 provides reach and dose delivered results by church and overall.

Consistent with an initial assessment conducted prior to the study, a majority of congregants attending the Catholic and Pentecostal churches were Latino, whereas the Baptist Church congregation was almost entirely AA. An estimated 60% of congregants were female at the Pentecostal and Baptist churches compared to a more equal gender representation at the Catholic Church. Church attendees were, on average, younger at the Latino churches compared to those at the AA church. About three-quarters of Latino church attendees were under the age of 45 whereas only about 40% of the AA Baptist church members were under the age of 45.

Dose Delivered

Five clergy members delivered a total of nine HIV sermons on two Saturdays and three Sundays. At the Catholic Church, the pastor and two assisting priests delivered five sermons, while the Pentecostal pastor delivered one sermon and the Baptist pastor delivered three. Four of the six sermons delivered in the Latino churches were in Spanish. Dose delivered process evaluation results are available in Table 2.

The Pentecostal pastor delivered the lengthiest HIV sermon (75 minutes). A majority of the sermon framed HIV using compassionate terms and the primary message was to love your neighbor as yourself. The pastor repeatedly exhorted congregants to “love people how God loved the world and all people.” He contrasted God’s “good” and “merciful” character with humankind’s sinful nature and encouraged congregants to extend love and kindness towards marginalized members of society, such as PLHIV.

Other HIV messages ranged from an average of 7.5 minutes at the Baptist Church (range: 5 to 10 minutes) to 12.5 minutes in the Catholic Church (range: 8 to 17 minutes). All three Catholic clergy members framed HIV as a call to compassion and called on congregants to be supportive and loving toward PLHIV. The Baptist pastor, however, only briefly spoke about HIV at the beginning of his sermons and gave a non-health related sermon. While his remarks covered project and HIV awareness and emphatically promoted HIV testing, few stigma reduction messages were mentioned. Further, he used rhetoric attributing HIV as punishment for sin and gave outdated information about HIV transmission modes, which may have undermined the goals of the sermon component.

Fidelity and Implementation

Overall, the Catholic pastor completed nine of ten sermon objectives included in the HIV sermon guide and the first assisting priest and the Baptist pastor each completed seven on the day when on-site HIV testing was held. The second assisting priest and the Pentecostal pastor completed the fewest objectives (five). In terms of including specific HIV-related themes, the Baptist pastor only mentioned one of four stigma reduction messages and the Pentecostal pastor only mentioned one of three HIV testing messages from the sermon guide. Table 3 provides information on fidelity to specific project themes and sermon objectives by church leader and provides example quotes. Detailed results for each theme are presented below.

Project and HIV Awareness, Knowledge, and Beliefs

All five clergy members discussed the purpose of the church’s involvement in the project. The Baptist and two Catholic clergy members also mentioned recent project activities. In terms of raising HIV awareness, only two mentioned how HIV was affecting the community. The Baptist pastor made a general statement about the impact of HIV on AAs and Latinos while the Catholic pastor provided information on the magnitude of the HIV problem in the community.

Another facet of HIV awareness is providing accurate information on how HIV is transmitted and related risk factors. Three clergy members discussed HIV transmission modes. The mostly commonly referenced mode consisted of sexual behavior, mentioned by the Baptist pastor and both assisting priests. The Baptist pastor identified adolescents as a particularly vulnerable population and mentioned adolescent sexual behavior in two of three messages. Specifically, he mentioned adolescent sexual behavior within the context of encouraging parents to have their adolescent children tested for HIV:

Our youth infected [sic] by the droves. Hopefully you are not a self-righteous parent if you have a teenager. Hopefully you are conscientious enough and loving enough that if you have a sexually active teenager, that maybe this test could potentially save the life of the young person. The teenager doesn’t necessarily have to be promiscuous, sleeping around. They can just sleep one time with the wrong person.

The two priests also referenced drug use as a risk factor. The first assisting priest mentioned drug use but did not state how HIV was transmitted: “There are many people in our community who, due to drug abuse or other behavior, may have HIV or AIDS-related complications.” In contrast, the second assisting priest explicitly mentioned the role of infected needles: “We sometimes believe [HIV] was transmitted through relations, right? A relationship. No! It’s also transmitted through injections, right? The poor use of syringes.”

In addition to mentioning the transmission of HIV through sexual contact, the Baptist pastor made two outdated statements about transmission that were not in the sermon guide. Specifically, he said, “And then it’s not always sexually transmitted. It could be a blood transfusion. You could go to a dental office…and you find yourself infected with HIV.” These statements reflect incorrect beliefs about HIV transmission that were disseminated to his congregation.

Stigma Reduction Cues

All the clergy members contrasted the prejudicial treatment of PLHIV with the importance of compassion and love. The two stigma reduction objectives that most clergy members did not complete were sharing a personal story about PLHIV and the visualization exercise (only one clergy member implemented this exercise). The Baptist pastor had the lowest fidelity among the five clergy members in this area, as he only completed one of four stigma reduction cues.

The Pentecostal and Catholic pastors each shared a personal experience interacting with PLHIV, which was hypothesized to activate injunctive social norms of respectful and compassionate treatment of PLHIV among congregants. The Catholic pastor described his experience in East Africa where he observed societal marginalization of PLHIV. He used inclusive language to preach a message of love and compassion and emphasized the importance of providing social support. By referring to PLHIV as “brothers and sisters” multiple times throughout the sermon and asserting a key message from the sermon guide—namely that neither HIV nor AIDS were sins—the Catholic pastor reinforced a message of de-stigmatization and promoted HIV as a health issue rather than a moral one. He also emphasized that congregants needed to address stigma by becoming informed, undergoing testing, and showing compassion and love for PLHIV, as illustrated by the following quote:

We can love Christ in them by showing them love and compassion, treating them normally by overcoming the stigma that they have because of lack of information…being a spiritual support for them, treating them as brothers and sisters, educating ourselves as a community, knowing better the reality of this problem, participating in different workshops in our community…and especially supporting and encouraging people affected with HIV, accepting and welcoming them in our community. Finally, being a source of inspiration for others by getting tested for HIV. Our example can make a difference.

While most clergy talked about PLHIV and how PLHIV were marginalized and stigmatized, the first assisting priest was the sole clergy member to employ the hypothetical contact scenario of meeting someone with HIV. Specifically, he used diabetes and cancer as examples of non-stigmatized diseases and mentioned the role of education to combat HIV stigma:

Because the problem of HIV and AIDS is in our community. That is why we have to educate ourselves about this problem so that we can create a home, a community, a parish where these people can come, so that you can assist these people – to listen or provide help or take their hand.

While the Pentecostal pastor also compared HIV to diabetes and cancer, he did not employ the visualization exercise. Instead, he said these diseases had significant reach in the Latino community and congregants had a duty to reduce their own disease risk and that of others.

HIV Testing Prompts

All the clergy members prompted congregants to get tested for HIV by stating the importance of testing, encouraging congregants to get tested, or mentioning the church’s testing event date. The Baptist and Catholic pastors exhibited a high level of fidelity to the HIV testing prompts and accomplished all three of these items, whereas the Pentecostal pastor only mentioned the church’s HIV testing event date.

Four clergy members (the Baptist pastor and three Catholic priests) spoke about the importance of getting tested, however, they presented different motivations for doing so. The Catholic priest accentuated the importance of education as motivation for getting tested while the Baptist pastor listed several reasons why congregants should be tested for HIV (including the value of early treatment) and emphasized that results were confidential to persuade those concerned. The second assisting priest mentioned socio-cultural reasons for not getting testing and encouraged congregants to overcome these barriers. He noted denial and fatalistic cultural beliefs as barriers to HIV testing among Latino populations:

Sometimes we say no, why should I go to a doctor? To find out that I have an illness? That is very common among Latinos. We say I would rather die than find out. That’s how it is. But we do not think about the pain of our neighbors and of those around us who also suffer.

In terms of cues to action, the Baptist and Pentecostal pastors were each tested at their respective church’s HIV testing event. While the Pentecostal pastor announced the event during the sermon, he did not share that he planned to be screened. The Baptist pastor told congregants he underwent testing to serve as a role model:

“What about you pastor?” Yeah, I’m doing something, too. That’s why I took the test. Everybody is doing something, right? That could potentially get them what? Infected. Am I right about that? Don’t look at me like you don’t know what I’m talking about.

Later in the same message, the Baptist pastor repeated that he had undergone testing, however, this time, he equated being infected with HIV with dishonor:

I’ve been tested. I’m happy to report that I am a vessel of honor…I am vessel of what? Honor. Because in God’s house you have vessels of dishonor. But I ain’t one of the dishonored ones; so I’m grateful for that. So I want to encourage you today to get tested.

In a second church service, he also said he was a “vessel of honor” since he “passed the test.” By referring to PLHIV as “dishonored ones,” the Baptist pastor expressed an attitude of condemnation and judgment that was not aligned with messages in the HIV sermon guide. He may have undermined the impact of his other statements to address HIV-related stigma, such as: “If a person has HIV, it doesn’t mean that they are what? A bad person. Doesn’t mean that. We want to live that stigma down right now. Amen.”

References to Scripture

Clergy generally indicated in their sermons that they were motivated by biblical principles and referenced parables or content from the New Testament as justification for their involvement in the intervention. We found clergy referenced similar scripture and content to address stigma and promote testing. Two pastors compared the HIV epidemic to leprosy and mentioned the historical role of marginalization in society and the church. The Baptist pastor framed HIV as modern-day leprosy and used the comparison to promote HIV testing:

AIDS is like our modern day what? Leprosy. So Jesus would go to the HIV ward in the hospital. Maybe you won’t go, but he’ll go. And he’ll pull up one of his [testing vans] right there in front of the synagogue. Amen. And have his disciples get what? Tested. Amen.

Similarly, the Pentecostal pastor mentioned leprosy as an “illness that marginalized people” and compared HIV stigma with the social isolation of people with leprosy. Near the end of his sermon, the Pentecostal pastor implored congregants to show compassion for PLHIV:

Maybe some 20 or 30, 50 [people] will come here with HIV, brothers and sisters, what will you do? Are we going to close the church? What did Israel do back then, with their legalism, brothers and sisters, toward those who had leprosy? “Get them out of here, take them over there.” And they marginalized a lot of people, brothers and sisters. We cannot do that. Never, never.

Three clergy members from two churches also referenced the Bible’s 25^th chapter of Matthew (which was the assigned lectionary reading in the Catholic Church that day). Two priests emphatically promoted compassion for marginalized individuals and HIV awareness and education using Matthew 25:31–46 (“Whatever you did to the least ones, you did to me…”).

Discussion

This study explores clergy members’ implementation of an HIV sermon intervention and the sermon content in a real-world setting. Clergy members were generally found to be receptive to the HIV sermon intervention and embedded HIV messages into their sermon while framing HIV using compassionate and non-judgmental terms. However, we found significant variation across clergy in the ways in which they implemented the sermon as well as fidelity to specific objectives. Some differences could be due to the varied faith traditions included (Roman Catholic, Pentecostal, and Baptist). For example, the wide range in length of the sermon is likely due in part to faith tradition norms. However, several differences could be related to differing levels of comfort with HIV and HIV-related attitudes and knowledge among clergy members.

Similar to findings from our extensive case study research among diverse religious congregations, the sermon content reflected a range of HIV norms and attitudes (Bluthenthal et al., 2012). Theological principles, institutional policies, understanding of the local community and related public health needs, individual agency, and personal experience with PLHIV shape clergy members’ attitudes and beliefs and can influence the content of an HIV sermon (Cunningham et al., 2011). It is encouraging that all five clergy members delivered an HIV message in this pilot study. Similarly, other studies with AA faith leaders have found a high level of interest in learning about HIV and discussing it with their congregants (Berkley-Patton, Thompson, et al., 2013; Nunn et al., 2012; Pichon, Williams, & Campbell, 2013).

Our findings shed light as to the feasibility and acceptability of asking clergy to deliver an HIV sermon. Although the visualization exercise was an innovative aspect related to a key stigma reduction theory, low fidelity to this objective suggests that clergy members may not have found it appropriate for a sermon. It is also possible that the study personnel did not sufficiently prepare clergy members to facilitate the visualization exercise or it was too difficult to integrate stylistically and theologically. In contrast, less theoretically-driven stigma reduction cues were found to be acceptable by nearly all clergy members. Future intervention studies that aim to incorporate imagined contact scenarios into a sermon may need to spend additional time discussing the value of the strategy or provide more training. It may also be useful to more effectively frame the value of the intervention in terms of spiritual goals that clergy hold for their congregants instead of health promotion goals that are more salient from a public health perspective. Finding other ways to incorporate “contact” components or personal narratives into church-based HIV interventions (e.g., PLHIV testimonials) should also be investigated.

Exploring the content of HIV sermons and clergy members’ communication strategies provides a valuable contextual understanding of the observed study outcomes. The Catholic pastor was particularly effective at integrating inclusive and compassionate rhetoric to describe PLHIV and framed HIV as an opportunity for transformation, which was likely due to his personal experience interacting with PLHIV (Derose & Kanouse, 2011). Other clergy at Latino intervention churches framed HIV as a call to compassion, which could have contributed to decreased HIV stigma and mistrust at the church level in these sites. The non-significant stigma reduction finding in the AA Baptist church compared to its control pair (Derose et al., 2016) may be partially explained by the pastor’s implementation and framing, inasmuch as he only briefly mentioned HIV-related themes and exhibited low fidelity to stigma reduction cues. Further, he disseminated outdated HIV transmission information and used condemning language toward PLHIV. His rhetoric and focus on HIV testing (and not stigma reduction) coupled with not describing the negative effects of HIV stigma or sharing a personal story about PLHIV may have contributed to the null stigma reduction finding. It is possible that the pastor’s own level of comfort on specific HIV topics (Pichon et al., 2012), stigma toward PLHIV, or lack of HIV knowledge could have influenced the content and brevity of the message.

Future intervention studies using HIV sermons may benefit from conducting more structured assessments of clergy participants’ HIV-related knowledge, attitudes, beliefs, and comfort with HIV topics prior to implementation. In our pilot, we invited pastors to participate in HIV education and peer leader workshops offered to congregants in addition to the individual meetings with pastors about the HIV sermon guide, but none did. We did, however, provide all workshop material to pastors for approval prior to implementation. It is likely that religious leaders’ wide range of responsibilities and activities limited their involvement (Bopp, Baruth, Peterson, & Webb, 2013; Bopp & Fallon, 2013). It is also possible that the workshop format—which was highly interactive and involved role playing—did not seem appropriate for them to do with other congregants, and an individualized or clergy-to-clergy approach may have been preferred (McNeal & Perkins, 2007). With a larger set of churches, clergy workshops could be feasible. Future studies should examine the dimensions and effects of clergy members’ HIV-related discourse (e.g., presentation style, tone, expressed empathy, and attitude) (Campbell & Babrow, 2004; Muturi & An, 2010) as these factors can shape congregants’ behaviors related to HIV prevention (Williams et al., 2014).

In the outcomes study, intervention churches demonstrated significantly higher HIV testing rates compared to their matched pairs (38% vs. 7 % in Latino Pentecostal churches; 32% vs. 13% in AA Baptist churches) (Derose et al., 2016). Clergy members’ high level of fidelity to communicating HIV testing prompts likely contributed to high testing rates, since nearly all the clergy members mentioned the importance of HIV testing and encouraged congregants to do so. The Baptist pastor’s high level of adherence to HIV testing prompts paired with his screening testimony may have contributed to the observed higher rate of HIV screening at his church, inasmuch as support from faith leaders can help promote testing through a “trickledown effect” (Stewart, 2015). Having faith members preach about their personal experience with testing can serve as an influential cue for congregants because role-model stories may increase audiences’ receptivity to a health message (Berkley-Patton, Goggin, Liston, Bradley-Ewing, & Neville, 2009). The magnitude of the effect of pastoral role modeling is unclear (Abara et al., 2015), however, and further research is needed to disambiguate the effect of a standalone HIV screening testimony from its inclusion in a multi-component intervention.

Limitations and Strengths

Our study’s findings are limited by the pilot nature and small sample size. It would be important to collect similar data on a full scale RCT, examine implementation by different types of faith leaders (e.g., non-clergy faith leaders), and include non-Christian faith-based organizations. Another limitation is related to the data collected from the religious service observation questionnaire, which is subject to observer bias. Because of limited resources in this pilot study, we did not evaluate two process evaluation measures (dose received and maintenance). Further, it would have been useful to collect data on clergy members’ comfort discussing certain health behaviors before and after the intervention (Pichon et al., 2012). Future health interventions with a sermon component should assess congregants’ reactions to message content, whether clergy members continue to embed health messages into sermons over time, and whether the framing of these messages is consistent after a study has ended.

Despite these limitations, this paper makes several important contributions. We are not aware of any publications that explore the content of health-related sermons, despite the frequent mention of this type of strategy as part of church-based health interventions. The qualitative exploration of the sermon component provides rich data and specific examples of how clergy can implement HIV sermons. Moreover, even with a small sample, we found considerable variation in implementation and collected data reflecting a range of attitudes, beliefs, and knowledge on HIV which helped us identify contextual factors and potential explanations for the outcome data. Lastly, the inclusion of Latino and AA churches, English and Spanish-speaking clergy, and multiple Christian faith traditions are also strengths, since most studies focus on churches that predominantly serve a single race/ethnic group or Christian faith tradition.

Conclusion

Embedding health messages into sermons is a potentially valuable strategy to address HIV and other health disparities in churches that predominantly serve racial and ethnic minorities. This study provides an example of how clergy members implemented an HIV sermon guide in AA and Latino churches across different Christian faith traditions. Overall, five clergy implemented nine HIV sermons in a real-world setting, indicating high acceptability of this strategy. However, fidelity varied widely across clergy members. While promoting HIV testing from the pulpit seemed viable and acceptable to all the clergy members, some exhibited low fidelity to the stigma cues, and issue framing varied. Future work is needed to fully optimize implementation of theory-based components in health-related sermons and to examine the viability and acceptability of embedding other health promotion messages into sermons.