Health Programs in Faith-Based Organizations: Are They Effective?

Abstract

Objectives. We examined the published literature on health programs in faith-based organizations to determine the effectiveness of these programs.

Methods. We conducted a systematic literature review of articles describing faith-based health activities. Articles (n = 386) were screened for eligibility (n = 105), whether a faith-based health program was described (n = 53), and whether program effects were reported (28).

Results. Most programs focused on primary prevention (50.9%), general health maintenance (25.5%), cardiovascular health (20.7%), or cancer (18.9%). Significant effects reported included reductions in cholesterol and blood pressure levels, weight, and disease symptoms and increases in the use of mammography and breast self-examination.

Conclusions. Faith-based programs can improve health outcomes. Means are needed for increasing the frequency with which such programs are evaluated and the results of these evaluations are disseminated.

There is a sizable multidisciplinary literature describing the health-related activities of religious or faith-based organizations (FBOs). Studies have described the features of successful health promotion programs and partnerships in churches1,2 and the importance of the church as an ally in efforts to provide preventive health and social services to at-risk populations.3 In addition, the interconnections between public health, health education, and FBOs have been examined,4 and the possible contributions of FBOs to improved community health outcomes have been described.5

A development related to health programs offered by FBOs is the need for improving access to care for the 43 million nonelderly uninsured adults residing in the United States. It has been shown that uninsured individuals are more likely than those with insurance coverage to avoid seeking needed care, to have been hospitalized for a preventable condition, and to have been diagnosed with advanced-stage cancer.6 Proposals for expanding health insurance coverage focus on increasing the role of government7 and generally ignore the role played by nonfunded health care providers in providing access to care. Especially important for public health practitioners is whether faith-based health programs can, for example, provide predictable and measurable health benefits in the communities they serve.8

FBOs have a long history of independently and collaboratively9 hosting health promotion programs in areas such as health education,1,10 screening for and management of high blood pressure11 and diabetes,12 weight loss13 and smoking cessation,14 cancer prevention and awareness,15–17 geriatric care,18 nutritional guidance,19 and mental health care.20 However, little is known about the effectiveness of these programs. Nonfunded health programs are not part of an organized system of care and are sometimes considered “nonsystems of care.”21 However, if such programs provide consistent access to specific types of care for specific individuals, they may actually be delivering predictable—but unmeasured—community health benefits.

A study was undertaken to review the health programs in FBOs and to examine their effectiveness. The Working Group on Human Needs and Faith-Based and Community Initiatives notes that the current vocabulary surrounding discussions of “faith-based” organizations tends to “confuse and divide.”22 The term FBO evokes images ranging from store-front churches, to the YMCA, to the local chapter of Habitat for Humanity. In the present article, the term FBO is used as a catch-all category referring to health programs designed, conducted, or supported by groups affiliated with or based in a nonsecular setting.

The National Congregations Study revealed that about 57% of US congregations participate in various social service delivery programs, including food and clothing, housing and homelessness, domestic violence, substance abuse, employment, and health programs.23 In the present study, we examined the health activities of FBOs only or those activities specifically related to health promotion/disease prevention. Also, we examined the published literature on FBO health programs in an attempt to ascertain the effectiveness of these programs. Successful programs are likely to be overrepresented in such a review, which is consistent with our study intent: we were not concerned with presenting an exhaustive review of social service activities in FBOs; rather, we intended this study as a first step in determining the possible contribution of health programs to maintaining or improving the health of individuals in the communities they serve.

METHODS

Literature Review and Search Strategies

We conducted a systematic qualitative review of health-related databases for the years 1990 through 2000.24 This 10-year period was selected by consensus among the authors on the belief that a “faith and health movement”25 occurred in the 1990s. Another reason we selected this period is that faith–health collaborations represent a rapidly developing phenomenon, and the results of a preliminary search indicated the existence of a large body of literature available during the period. The purpose of the review was to identify all published English-language research articles reporting the health activities of FBOs. Our search strategies were guided by a preliminary review of the literature, and the searches were conducted by one of the authors, who is a professional research librarian (L. W.).

We chose MEDLINE as our major database and, because there were no existing medical subject headings specific enough for our topic, we devised a comprehensive search strategy. Our strategy involved the use of a set of indexing terms related to health service delivery, such as health promotion, health education, counseling, and screening. These terms were combined with a second set of text words (e.g., parish, congregation, faith based, community church) describing where the health services might be delivered.

We performed supplemental searches of the HealthSTAR, CINAHL, and PsycINFO databases. In the case of HealthSTAR, we created and combined 3 groups of terms: health service terms, religion terms and phrases, and diagnosis and therapy terms. The CINAHL search consisted of identifying articles including one of 3 phrases—faith based, church based, or parish based—or either parish nursing or congregational nursing. We used 2 alternative strategies in the search of the PsycINFO database. The first focused on the phrases faith based and church based, since the phrase parish based was not useful in this database; the second focused on a group of religion terms and a separate group of community mental health service terms.

All articles (n = 386) meeting the search criteria were reviewed by 1 of the investigators (M. J. D.) for possible inclusion in the present study. Titles and abstracts were examined for consistency with our objective of identifying health programs involving FBOs. In cases in which abstracts were not available, determinations were made on the basis of title alone. If the title did not provide a clear indication of the article’s content, the article was obtained before a determination was made regarding inclusion or exclusion. After evaluation of the search results, 106 articles1–4,9–20,26–115 were identified for formal review.

The formal review consisted of reading an article to ensure that it addressed a specific, identifiable health program that could be linked to a specific health benefit. The following types of articles were excluded: articles discussing the existence of a program without describing its features, articles discussing a “healing ministry” without describing a specific program, and review articles describing a collection of programs without providing details about individual programs. In addition, articles were excluded when the church building was being used for a multisite program developed as part of a broader public health strategy (however, articles were included if the church or congregation was an active member of a communitywide health coalition). Once these articles were excluded, 53 articles remained.

Data Gathering

Information was recorded about program features and outcomes, including location (city and state), scope (congregation, community, city, or region), number of congregations involved, target population (age and ethnicity), target conditions, and program objective (primary, secondary, or tertiary prevention). Objectives were coded as primary when the program was designed to increase awareness of disease, secondary when the goal was risk reduction, and tertiary if treatment was involved. When more than 1 type of prevention activity was involved, the objective of the majority of program activities was recorded. When a program qualified for more than 1 program scope area, the code for the largest geographic scope was entered.

Programs were categorized according to FBO level of involvement, whether program outcomes were measured, and number of participants. Almost all programs evaluated were based in a church or congregation, as opposed to an interfaith service organization, temple, or mosque, consistent with the finding of Chaves et al.116 that only about 3.5% of all social services are delivered in non-Christian settings. Determining level of church involvement was essential since most analysts agree that collaboration is necessary for the success of faith-based health and community programs.30,51,52,73 Church involvement was coded as “faith placed” if health professionals used the church to test an intervention and “faith based” if the program was part of the church’s health ministry. Programs were coded as “collaborative” if they combined faith-placed and faith-based features.

In instances in which no clinical outcomes were reported, we used process measures. When only number of client contacts was reported, we did not include this information in our measurements because it was not related to possible health benefits. Finally, we recorded total number of participants, including experimental controls and, in the case of multiple-year programs, individuals participating in all years of the program. When program outcomes were reported, articles were evaluated by 2 investigators, and disputes over coding content were resolved through discussion.

In the following, we report descriptive statistics, including percentages and measures of central tendency and dispersion. We conducted all analyses using SPSS version 10.0. We used χ² tests of independence in examining relationships between categorical variables.

RESULTS

Health programs were conducted in 30 distinct geographic locations, either counties or cities. Although most locations hosted 1 program, 5 cities accounted for approximately one third of the total number of programs: Chicago (n = 6; 11.3%), Baltimore (n = 4; 7.5%), Los Angeles (n = 4; 7.5%), Cleveland (n = 2; 3.8%), and Oakland (n = 2; 3.8%). Programs were located in 23 different states, but almost half (n = 26) were located in 5 states: California (n = 8; 15.1%), Illinois (n = 6; 11.3%), Maryland (n = 5; 9.4%), Ohio (n = 4; 7.5%), and Florida (n = 3; 5.7%).

The majority of programs were directed at congregation members (60.4%) or the surrounding community (24.5%) (Table 1 ▶). Although more than 40% of the programs involved a single congregation, the median number of participating congregations was 3 (range = 1–95), and the number of program participants ranged from 7 to 2519 (median = 238). Most programs focused on primary prevention (50.9%), usually patient education, in the area of general health maintenance (24.5%), cardiovascular health (20.7%), or cancer (18.9%). Approximately one third of the programs did not target a specific population (32.1%); however, when a population was targeted, it tended to be African American (41.5%) and adult (43.4%). The overwhelming majority of programs did not involve a specific target in terms of gender (75.5%).

TABLE 1—

Program Features (n = 53)

Feature	Sample, No. (%)
Program scope
Congregation	32 (60.4)
Community	13 (24.5)
Region	5 (9.4)
City	2 (3.8)
Not reported	1 (1.9)
Objective
Primary prevention	27 (50.9)
Secondary prevention	13 (24.5)
Tertiary prevention	7 (13.2)
Other	5 (9.4)
Target population
African American	22 (41.5)
Not specified	17 (32.5)
Low income	7 (13.2)
Hispanic	4 (7.5)
White	2 (3.8)
Other	1 (1.9)
Target conditions
General health maintenance	13 (24.5)
Cardiovascular health	11 (20.7)
Cancers	10 (18.9)
Mental health	6 (11.3)
Other/not specified	6 (11.3)
Nutrition/weight control	4 (7.5)
Smoking	3 (5.7)
Faith involvement
Faith placed	23 (43.4)
Faith based	13 (24.5)
Collaborative	16 (30.2)
Not specified	1 (1.9)
Outcomes measured
Yes	28 (52.8)
No	25 (47.2)
Target age group
Adult	23 (43.4)
Elderly	6 (11.3)
Not specified	24 (45.3)
Target gender
Not specified	40 (75.5)
Female	10 (18.9)
Male	3 (5.7)
No. of participants
7–46	9 (17.0)
55–187	9 (17.0)
238–668	9 (17.0)
743–2219	9 (17.0)
Not specified	17 (32.0)
Total	53 (100.0)

Faith-based programs developed as part of a congregation’s health ministry accounted for the smallest percentage of programs (24.5%), while faith-placed programs, usually developed by health professionals outside of a congregation, accounted for the largest percentage (43.4%). Although more than one half of the programs (52.8%) reported outcome measurements, such reports were significantly related (P ≤ .012) to type of church involvement (Table 2 ▶). Faith-placed programs were significantly more likely to report outcome data (75%) than either faith-based (30.8%) or collaborative (37.5%) programs.

TABLE 2—

Numbers of Programs, by Program Type and Published Measurement of Effects

Program Type	Outcomes Not Reported, No. (%)	Outcomes Reported, No. (%)	Total, No. (%)
Faith placed	6 (25.0)	18 (75.0)	24 (100)
Faith based	9 (69.2)	4 (30.8)	13 (100)
Collaborative	10 (62.5)	6 (37.5)	16 (100)
Total	25 (47.5)	28 (52.8)	53 (100)

Note. Outcome differences are significant at the P = .012 level of significance.

The characteristics and types of outcomes reported by programs with different levels of church involvement (n = 28) are reported in Table 3 ▶. The “results” column indicates whether a study reported a process evaluation (n = 8) or the effects of a program intervention (n = 20). Among the 18 faith-placed programs reporting outcomes, only 11 (61%) reported the effects of a program intervention. Effects were measured via self-generated33 or self-report18,39,43,53,97,106 instruments or via biological measures.12,13,84,112

TABLE 3—

Program Features and Outcomes of Programs at Different Levels of Church Involvement

Study	No. Subjects	Program Scope	No. Churches	Ethnicity	Study Focus	Method	Result	Statistical Significance of Results
Faith placed
Wiist and Flack (1990)112	348	Congregation	1	African American	Heart (cholesterol)	Intervention	Decreased cholesterol	Significant
Holschneider et al. (1999)64	98	Congregation	1	Hispanic	Breast cancer	Screening	Process evaluation only	No statisticsa
Fox et al. (1998)58	82	Community	1	Hispanic	Breast cancer	Screening	Process evaluation only	No statisticsa
Duan et al. (2000)16	813	Congregation	30	Not specified	Breast cancer	Intervention	Increased/maintained screening level	Significant
Flack and Wiist (1991)56	661	Congregation	6	African American	Heart (cholesterol)	Screening	Process evaluation only	No statisticsa
Smith et al. (1997)11	97	Congregation	17	African American	Heart (blood pressure)	Intervention	Decreased blood pressure	Significant
Campbell et al. (1999)39	2519	Region	50	African American	Nutrition	Intervention	Increased fruit/vegetable consumption	Significant
Voorhees et al. (1996)106	292	Community	21	African American	Smoking	Intervention	Increased readiness to change	Significant
Smith (1992)97	32	Congregation	3	African American	Heart	Intervention	Increased knowledge about hypertension	Significant
Wilson (2000)10	129	Congregation	3	Not specified	Heart	Screening	Process evaluation only	No statisticsa
Erwin et al. (1999)53	433	Community	11	African American	Breast cancer	Intervention	Increased breast self-examination	Significant
Collins (1997)43	30	Congregation	1	African American	Prostate cancer	Intervention	Increased knowledge	No statisticsa
Huggins (1998)65	1200	Community	3	Hispanic	General health	Screening	Process evaluation only	No statisticsa
Boehm et al. (1995)33	123	Congregation	.	African American	Prostate cancer	Intervention	Increased knowledge	Significant
Weinrich et al. (1998)108	743	Region	59	African American	Prostate cancer	Screening	Process evaluation only	No statisticsa
Oexmann et al. (2000)84	133	Congregation	8	African American	Heart	Intervention	Decreased weight and blood pressure	Significant
McNabb et al. (1997)12	39	Congregation	3	African American	Weight	Intervention	Decreased weight and changed eating habits	Significant
Davis et al. (1994)17	1012	Congregation	24	Underserved (low income)	Cervical cancer	Screening	Process evaluation only	No statisticsa
Faith based
Ruesch & Gilmore (1999)93	7	Congregation	1	White	Heart	Intervention	Increased knowledge of heart disease	No statistics
Toh & Tan (1997)104	46	Congregation	1	White	Mental illness	Intervention	Decreased symptoms and complaints	Significant
Toh et al. (1994)105	18	Congregation	1	Not specified	Mental illness	Intervention	Decreased symptoms and percentage complaints	Significant
Roque et al. (1999)92	30	Community	1	Underserved (low income)	Asthma	Intervention	Decreased hospital and emergency department visits	No statisticsa
Collaborative
Schorling et al. (1997)14	453	Region	14	African American	Smoking	Intervention	Found no change in quit rates	Nonsignificant
Turner et al. (1995)115	2212	Region	.	African American	Heart	Health promotion	Process evaluation only	. . .
Cowart et al. (1995)18	238	Congregation	4	African American	General health	Intervention	Increased overall health	Significant
Barnhart et al. (1998)19	30	Congregation	1	African American	Nutrition	Intervention	Increased fruit/vegetable consumption	Significant
Kumanyika & Charleston (1992)13	187	Congregation	22	African American	Weight	Intervention	Decreased weight and blood pressure	Significant
Rydholm (1997)94	966	Congregation	20	Not specified	General health	Intervention	Cost savings/costs averted	No statisticsa

^aStatistical analysis not reported or incomplete.

The areas addressed by the programs included heart disease (36.4%), weight/nutrition (18.2%), breast cancer (18.2%), prostate cancer (18.2%), and smoking cessation (9.0%). The programs focusing on these areas achieved statistically significant effects in terms of, respectively, reducing cholesterol and blood pressure levels, increasing fruit/vegetable consumption and reducing weight, increasing use of mammography and breast self-examination, increasing knowledge about prostate cancer, and increasing readiness to change regarding smoking cessation. The number of participants in these programs ranged from 30 to 2519 (median = 133), and almost all of the programs (91%) were targeted at African Americans.

All 4 of the faith-based programs included in the sample reported intervention effects, and these programs addressed heart disease (25%), mental illness (50%), and asthma (25%). In both of the studies demonstrating significant effects, validated instruments showed decreased mental illness symptoms.104,105 The number of participants was small, ranging from 7 to 46 (median = 24).

Of the 6 collaborative programs, 5 (83.4%) reported program intervention effects on general health (40%), weight/nutrition (40%), and smoking cessation (20%). Outcomes were evaluated via self-report and biological measures,18 validated instruments,19 and biological measures.13 Significant effects included improvements in overall health status, increases in fruit/vegetable consumption, and decreases in weight and blood pressure. These programs ranged in size from 30 to 966 participants (median = 133), and the programs were almost exclusively (80%) directed toward African Americans.

DISCUSSION

In this study, we reviewed FBO health programs and assessed their effectiveness. Our objective was to take a first step toward determining whether these types of programs can provide a measurable form of community-based care. The first conclusion offered by our review is that relatively little information exists on which to base assessments of the effectiveness of such programs. Although our literature search identified a substantial number of articles (n = 386) possibly related to our study objective, fewer than 1 in 3 (n = 106; 27.5%) were eligible for the review, and even fewer (n = 53; 13.7%) actually discussed a specific program. Finally, only a small number of articles presented outcome measures (n = 28; 7.25%) or outcome measures associated with a particular program intervention (n = 20; 5.4%).

The data presented here nonetheless demonstrate that faith-based health programs can produce positive effects; for example, they can significantly increase knowledge of disease, improve screening behavior and readiness to change, and reduce the risk associated with disease and disease symptoms. According to the Bureau of Primary Health Care Faith Partnership Initiative, which seeks to facilitate partnerships between FBOs and health providers, there are 43 million uninsured citizens in the United States, it is not known how to meet the health-related needs of this group, there are more churches per capita in the United States than in any other country, and faith communities are involved in public health and community development issues related to social justice.8 Our findings suggest a number of recommendations for future study if FBOs are to contribute to community health in the ways envisioned by the Faith Partnership Initiative.

Recommendation 1: Increase collaboration between FBOs and health professionals for the purpose of evaluating health activities and disseminating findings. Disproportionately more is known about the effectiveness of faith-placed programs than either faith-based or collaborative programs. In the present study, we found that 55% of the programs testing interventions were faith placed, 20% were faith based, and 25% were collaborative.

As many as 57% to 78% of congregations are involved in health activities.23,117 By increasing collaboration between health professionals and faith-based groups, it may be possible to introduce evaluation strategies into programs and to disseminate the results to a wider audience. Researchers and other health professionals should consider developing user-friendly workshops and tools for use by individuals associated with FBOs that are accustomed to delivering but not evaluating health-related programs. Since FBOs and churches are familiar community-based institutions, they frequently succeed when outside health professionals cannot.118 More thorough collaboration between researchers and FBOs will facilitate better understanding of the community on the part of these health professionals, contribute to building the credibility of their projects,3,119 and, we hope, promote increased program evaluation.

Recommendation 2: Place more emphasis on effectiveness studies as opposed to efficacy studies. Efficacy studies test the effects of interventions regardless of their practical application, whereas effectiveness studies test interventions in a way that is sensitive to what is practical in the real world. Efficacy studies generally require a more sophisticated study design, a greater amount of funding, and a greater degree of commitment and control than is typically available in most community-based settings. Consequently, they may be difficult to replicate in most congregations, especially in a way that could reliably contribute to a community’s health.

In the present study, 7 of the 15 intervention studies reporting significant findings involved either a quasi-experimental53,112 or an experimental12,16,39,104,106 design, and all but 1 of these interventions were classified as faith placed. We suggest the use of study designs that are concerned with the quality of the care delivery system as opposed to more sophisticated designs that may be beyond the expertise of local program planners and difficult to implement in their care setting. Continuous Quality Improvement efforts and “Plan–Do–Study–Act” cycles, with their emphasis on process of care, systematic methods, short cycles, and real-world application, offer more accessible and manageable approaches to evaluating programs in these community-based settings.120,121

Recommendation 3: Devote more attention to building relationships with the racially and ethnically diverse populations that increasingly characterize communities in the United States. When a target population was identified in the present study, it tended to be African American (41.5%), and most of the faith-placed intervention programs (91%) were directed toward African American populations. This finding is not surprising since, in a majority of African American communities, the church is considered the most important social institution36 and is the key community agent linking the African American community to the wider society beyond the congregation.51 In addition, African American churches can reach large numbers of individuals in the communities outside of their particular congregations114 and can sponsor community activities for all of those in need.73,103

It is important to both continue and to expand the work that is currently being done in African American communities among the many successful and progressive faith–health partnerships. However, we must also recognize that there are significant needs in other racial and ethnic groups, especially Hispanics. Although non-Hispanic Whites represent approximately half of all uninsured individuals, African Americans and Hispanics, respectively, are twice as likely and 3 times as likely as non-Hispanic Whites to be uninsured.122 As previously mentioned, uninsured individuals are more likely than those with insurance coverage (1) to forgo or postpone preventive care and skip recommended tests or treatments,123 (2) to be hospitalized for conditions that can be treated in outpatient settings (e.g., uncontrolled diabetes), and (3) to be diagnosed with late-stage colorectal cancer, melanoma, breast cancer, and prostate cancer.124 Given the types of health services offered through FBOs, increased collaboration between health professionals and FBOs serving Hispanic populations could potentially improve quality of life in this vulnerable group.

The present study and the recommendations offered help provide a better conceptualization and understanding of the extent of existing information, our need for more information, and possible directions for future collaboration between public health professionals and those providing health services through FBOs. Despite the different perspectives of these 2 groups, they tend to share a passionate commitment to improving the quality of life of vulnerable populations. If faith and health partnerships can help address the existing and expected health needs of vulnerable populations, more thorough information about their possible contribution is needed to make informed policy decisions. Only by increasing the evaluation component of faith-based programs and disseminating the information gained will it be possible to determine how these programs can contribute systematically to improving the health and quality of life of at-risk populations in our communities.