Prevalence and Predictors of Mental Health Programming Among U.S. Religious Congregations

Nearly one in five U.S. adults is affected by mental illness in any given year (1), and roughly 25% of these adults turn to religious congregations (for example, churches, synagogues, and mosques) for help with mental illness (2). Despite the substantial number of people who seek mental health support from congregations, little is known about the extent to which congregations are responding to the mental health needs within their communities. Prior studies have focused primarily on the role of clergy in addressing mental health needs and on factors that facilitate clergy referrals to mental health professionals (3–6). The relative dearth of studies examining congregational involvement in mental health programming is noteworthy given the extensive research on congregation-based social services and physical health–related programs (7–9).

The handful of studies assessing the prevalence of congregation-based mental health programming is mainly limited to studies involving African-American churches in specific regions in the United States (5,10). One recent exception is Frenk’s study (11), which analyzed data from the 2006 wave of the National Congregations Study (NCS) to estimate the prevalence of mental health programming among all types of U.S. congregations. In the 2006 NCS, congregations were asked to describe any social service programs, groups, meetings, classes, or events that they sponsored; responses were later coded to identify services targeting mental health–related issues. Only 8% of congregations reported sponsoring mental health programming, most of which were addressing substance use disorders (for example, Alcoholics Anonymous). Separate analyses of the same data set found that nearly two-thirds of congregations reported providing at least one health-related program (9); thus the proportion of congregations providing mental health programming was found to be a small subset of those providing health-related programming. Although Frenk’s study analyzed a nationally representative sample of U.S. congregations, its reliance on responses to an open-ended question most likely resulted in an underestimation of congregation-based mental health programming.

In addition to the need for a more accurate estimate of the prevalence of congregation-based mental health programming, there is a need to better understand which types of congregations are involved in providing such services. Congregations are organizations influenced by internal and external factors that affect their ability to provide various services (8,12). Internal factors, such as congregational resources (for example, the number of members and paid staff), characteristics (for example, theological orientation and clergy education), and member composition (for example, income levels), are associated with a congregation’s likelihood of providing social service programs (8,9,13,14). External factors, such as a congregation’s level of engagement with its surrounding community (for example, conducting community needs assessment) and community-level characteristics (for example, poverty level), also are associated with a congregation’s likelihood of sponsoring social services (8,13).

To our knowledge, Frenk’s is the only prior study that analyzed data from a nationally representative sample of congregations to identify factors associated with congregations’ provision of mental health programming (11). That study tested whether provision of mental health programming is related to a congregation’s internal factors but did not examine the influence of external factors. Yet external factors, such as interactions between a congregation and its surrounding community, may influence whether the congregation provides mental health programming. External engagement may foster increased knowledge, recognition, and capacity within congregations to respond to the mental health needs in their communities, which may be particularly critical considering that few clergy have formal training in mental health care (4,15). Moreover, the racial-ethnic and socioeconomic composition of the communities surrounding congregations may influence the demand for congregation-based mental health programming, given barriers to accessing professional treatment or preferences for faith-based mental health care (16).

In this study, we analyzed data from the most recent wave of the NCS (2012) to estimate the percentage of congregations providing programs to support people with mental illness and to identify internal and external factors associated with congregations’ sponsorship of mental health programming. Through its use of the 2012 NCS—the only wave to date that explicitly asked congregations whether they provided mental health–related programming—and its analysis of both internal and external factors, this study was an unprecedented national-scale examination of congregations’ role in providing mental health support.

Methods

Participants and Procedures

The NCS is a nationally representative survey of U.S. religious congregations that was initiated in 1998 and administered in two subsequent waves in 2006 and 2012 (17). Each of the three waves of the NCS used a sample derived from its respective 1998, 2006, or 2012 General Social Survey (GSS), a nationally representative survey of U.S. adults. GSS respondents who reported attending religious services at least once a year were asked to provide the names of their congregations. The congregations named by the respondents were then used to establish a representative sample of U.S. congregations (18). We analyzed data from the 2012 NCS, the first wave to ask congregations specifically about offering mental health programming. Key informants at each congregation (for example, clergy and staff) were interviewed in person or by phone on a wide range of topics, including congregational member characteristics, programming, and resources. A total of 1,331 congregations participated in the 2012 NCS (73% response rate). All study procedures were approved by the RAND Human Subjects Protection Committee, which determined that this study was exempt from further review given the use of publicly available, deidentified secondary data at the organizational level.

Measures

Our study’s dependent variable—whether a congregation provided mental health programming—was measured by using the following survey item: “Within the past 12 months, have there been any groups or meetings or classes or events specifically focused on the following purposes or activities?” Congregational informants were asked to provide responses for 26 distinct groups or activities, which spanned a wide range of areas (for example, “To discuss politics?” or “To discuss parenting issues?”). Mental health programming was assessed with the dichotomous item: “Support for people with mental illness?”

The independent variables included measures of congregations’ internal and external factors expected to be associated with providing mental health programming, based on prior research examining factors associated with congregational provision of health and social services (8,9,12–14). Internal factors included variables related to member composition, congregational characteristics, congregational resources, and involvement with health and human services. Member composition was characterized along three dimensions: age (percentage of members under age 60), education (percentage of members with a four-year college degree or higher), and income (percentage of members with an annual household income greater than $140,000). Congregational characteristics included clergy’s age (age of head clergy), the congregation’s theological orientation (theologically conservative, 1; moderate or liberal, 0), congregational practices (speaking in tongues practiced: yes, 1; no, 0), and the congregation’s religious tradition (black Protestant, 1; Roman Catholic, mainline Protestant, conservative Protestant, or non-Christian, 0). The descriptive statistics provided the percentage of congregations affiliated with each religious tradition. Congregational resources were measured by the congregation’s size (number of regular adult participants) and staffing (had an employee who spends more than 25% of work time on social service programs: yes, 1; no, 0) and whether the congregation owned its building (yes, 1; no, 0). Involvement with health and human services was measured with two dichotomous items indicating whether the congregation provided health-focused education programs and whether the congregation sponsored social service programs.

External factors comprised variables related to the congregation’s external engagement and community characteristics. External engagement was measured with two dichotomous items indicating whether the congregation assessed the needs of its community and whether the congregation had a social service representative as a visiting speaker. Community characteristics were measured with four dichotomous items related to the racial-ethnic composition of the congregation’s surrounding neighborhood (census tract is at least 80% African American), the socioeconomic composition of the congregation’s surrounding neighborhood (census tract in which at least 30% live below the official poverty level), the congregation’s community context (census tract is predominantly urban), and geographic region (located in the South). The independent variables were all dichotomous, except for the member composition variables (clergy age and congregation size), which were continuous.

Missing values for the independent variables were multiply imputed by using the imputation by chained equations package in Stata 14.1 (19), which uses the distribution of the observed data to estimate a set of plausible values for the missing data (20). Ten complete imputed data sets were created and pooled by using Rubin combination rules (21). Item nonresponse diagnostics (that is, stabilization plots) indicated that the mean and variance estimates for the variables with missing values had stabilized and thus did not contain significant nonresponse bias (22). Four cases were missing values for the dependent variable. These cases were not included in analyses.

Analyses

To identify factors associated with congregations’ provision of mental health programming, we first conducted bivariate analyses on each of the variables measuring internal and external factors. To account for the multiple bivariate comparisons and protect against type I error, we applied a Bonferroni correction. We then performed a multivariate logistic regression incorporating all of the variables. All analyses were weighted to the congregational attendee levels, which is considered the appropriate method for studies aimed at assessing the social impact of congregational services (9,23,24). Additional analyses (not displayed) indicated that the cases with imputed values did not significantly affect the results of the multivariate analysis.

Results

Descriptive statistics at both the congregation and attendee level are presented in Table 1. Based on data from the 2012 NCS data, nearly a quarter (23%) of U.S. congregations provided mental health programming. Thus approximately 31% of all attendees belonged to a congregation that provided mental health programming.

Table 2 presents results of the bivariate and multivariate analyses. The multivariate analysis, which incorporated all the independent variables, indicated that congregations whose members were predominantly younger (odds ratio [OR]=1.01, p<.01) and relatively wealthy (OR=1.02, p<.001) were more likely to sponsor mental health programming. Among the congregational characteristics, sponsoring mental health programming was positively associated with the clergy’s age (OR=1.01, p<.05) and the practice of speaking in tongues (OR=1.43, p<.05). With respect to congregational resources, congregations with more attendees (OR=1.11, p<.05) and with staff dedicated to providing social services (OR=1.86, p<.001) were more likely to sponsor mental health programming. Providing health-focused education programs was also positively associated with providing mental health programming (OR=2.03, p<.001).

Regarding the external engagement indicators, congregations that assessed the needs of their communities (OR=1.98, p<.001) and those that had hosted a social service representative as a visiting speaker (OR=1.41, p<.05) were more likely to sponsor mental health programming. In addition, congregations in neighborhoods with predominantly African-American residents (OR=2.20, p<.05) were more likely to have mental health programming. The most notable correlations indicated that employing staff for social services, providing health-focused education programs, assessing the needs of the congregation’s community, and being located in a predominantly African-American community each increased the odds of a congregation providing mental health programming by approximately a factor of two.

Discussion

This study examined the prevalence of congregation-based mental health programming and factors associated with providing such services by analyzing data from the 2012 NCS. The 2012 NCS was the first survey conducted with a representative sample of U.S. congregations that specifically asked congregations whether they provided mental health programming. Approximately one-third of all congregational attendees were in a congregation that provided programming support for people with mental illness. At the congregation level, nearly a quarter of congregations reported providing such support, which is approximately three times greater than the estimate derived by a prior study that used the 2006 NCS (11).

Although the higher rate found in this study may partly reflect the general trend of congregations to become more involved in providing social services (14), it is also likely that this study was better able to detect the provision of congregation-based mental health programming because congregations were specifically asked about providing such services. Furthermore, the large majority of mental health–related services identified in the 2006 NCS study were related to substance use disorders (11). The 2012 NCS, which asked separate questions about programs for mental illness and substance use disorders, allowed for a more precise estimate of congregational programming focused specifically on supporting people with mental illness. We note that in the 2012 NCS data, the proportion of U.S. congregations providing programming for substance use problems (38%) was greater than that providing mental health programming.

The 2012 NCS enabled a more accurate test of factors associated with congregations’ sponsorship of programs targeting mental illness because it did not comingle mental health services with services for substance use disorders. We found that several internal factors were associated with providing mental health programming. Congregations with younger and more affluent members had slightly higher odds of providing mental health programming. Given that younger adults and people with a higher socioeconomic status tend to hold less stigmatizing views of mental illness (25–27), it is possible that congregations with younger and wealthier members may foster a more supportive environment for people with mental illness. In addition, the odds of providing mental health programming increased slightly in congregations with older clergy and in congregations with the practice of speaking in tongues. Older clergy may have greater exposure to and experience with caring for people with mental illness. Speaking in tongues is a spiritual practice similar to other practices, such as prayer with the laying on of hands and memorizing scripture, that many Pentecostal Christians view as being effective methods for treating depression (28). Because congregations that practice speaking in tongues tend to correlate mental health with spiritual health, they may be more inclined to provide support for mental health issues (29).

Congregational resources, including the congregation’s size, having staff for social services, and providing health education and social service programs, were positively associated with congregation-based mental health programming. These findings are consistent with previous studies demonstrating a connection between congregational resources and the provision of health-related programs (8,24), but they counter findings from the 2006 NCS, which did not identify any significant relationships between a congregation’s resources and provision of mental health programming (11). This contradictory finding could be partly due to the fact that most mental health–related services identified in the 2006 NCS study targeted substance use disorders; many of the corresponding services involved activities such hosting Alcoholics Anonymous meetings, which do not require substantial congregational resources. Future studies that examine differences among congregations sponsoring mental health–related versus substance abuse–related services could provide a better understanding of how congregations view and address these needs and reveal potential avenues for collaboration with professional care providers.

We also found that external factors were associated with congregations’ sponsorship of mental health programming. Congregations that engaged with their external environment by assessing the needs of their surrounding communities and hosting speakers from social service agencies were more likely to sponsor mental health programming. This finding is consistent with prior research on congregations’ provision of health programs (8,13). A congregation that interacts with its surrounding community may become more aware of mental health needs and recognize its capacity to provide support for people with mental illness (8). In addition, congregations in predominantly African-American neighborhoods were more likely to sponsor mental health programming. Given that African Americans are less likely to access mental health treatment (30), it is possible that congregations in predominantly African-American neighborhoods are responding to this unmet mental health need by providing support services. Although many African Americans turn to congregations for help with personal needs (31,32), little research has been conducted on congregation-based mental health programming within African-American communities (33). A few studies, however, have suggested that building collaborations between African-American congregations and the mental health sector is a promising strategy to increase access to needed services (34–36).

Our findings should be considered in light of certain study limitations. Associations for certain congregational factors (that is, clergy age and congregational members’ age and income), although statistically significant, should be interpreted with caution given the small ORs, which suggest minimal effects in actual practice. In addition, although the 2012 NCS included a specific question about providing support for people with mental illness, it did not collect data on the mental health conditions addressed. Congregations may regard a wide range of programs as being mental health related—from issues that are more typical of clinical disorders, such as depression, to less clinically defined conditions, such as marital counseling. Similarly, little is known about the type of mental health programming that congregations offer, which could vary from more spiritual forms of support (for example, prayer) to more education-based support or counseling. In addition, the congregational factors associated with providing mental health programming may differ depending on the type of programming offered. Furthermore, because the 2012 NCS did not collect data on why congregations decided to offer particular services, we could not determine whether congregations offered mental health programming as a collaborative continuum with professional mental health care providers or as an option to rely exclusively on religious care (37).

Overall, more detailed information about congregation-based mental health programming is needed to better understand the role and functioning of congregations within the broader system of care for people with mental illness. For example, even though a sizable proportion of congregations offer programming to support people with mental illness, it is unclear whether this programming was complementing, supplanting, or filling in gaps with respect to professional mental health treatment. Currently, more than 60% of U.S. adults with a mental illness do not obtain the professional treatment they need (38). Furthermore, over half of U.S. adults who seek help from congregations for a mental disorder do so at the exclusion of other health care providers (2). For instance, although racial-ethnic minority groups and young adults are among the least likely to obtain professional mental health treatment (30,39), they are among the most likely to seek mental health care from religious congregations (2,32). Further research is needed to examine whether interventions designed to foster greater collaboration between congregations and the mental health service sector can improve access to treatment among groups with the greatest levels of unmet mental health needs (10,40–42).

Conclusions

Given that congregations are often the first and only point of contact for people seeking help for mental illness (2), they can play an important role in providing support and in facilitating connections to professional treatment. A growing number of mental health professionals view congregations as potential collaborators in helping to address unmet mental health needs (41,43,44). Greater coordination between mental health providers and congregations with programs that support people with mental illness could foster more integrated and holistic care, which in turn may lead to improved recovery outcomes.