A prevalence study of faith-based healing in the rural southeastern United States.

Background: Although prayer and other spiritual practices are common among residents of the rural south, the use of faith-based healers (FBH), or healers who use prayer as their primary healing modality, has not been explored in this population.

Methods: Secondary data analysis from a random digit dialing telephone survey of rural adults in eight southern states.

Results: Our overall response rate was 51% and 193 subjects (4.1%) had seen an FBH practitioner within the previous year. FBH use was significantly more common among younger respondents (OR 7.21, 95% CI 2.00, 25.94), women (OR 1.49, 95% CI 1.03, 2.14), those reporting poorer health (OR 1.83, 95% CI 1.19, 2.83), and those who believed in avoiding physicians (OR 1.82, 95% CI 1.24, 2.67). A relationship between FBH use and delayed or foregone medical care, and cost as a barrier to obtaining care was not statistically significant after controlling for other factors.

Conclusions: Prevalence of FBH use is low, but is significantly related to younger age, female gender, poorer health status, barriers to medical care and devaluing medical care. Clinicians may consider exploring FBH usage with their younger, female patients, and those in poorer health. Policy makers should consider how FBH usage is related to various indicators of health care services demand, utilization and access.

Key Words: access to care, religion and medicine, rural populations, prevalence

**********

Personal belief systems, faith traditions and spiritual practices are central to the health and illness experiences of many people. (1) For example, prayer for health concerns is the most commonly cited healing practice and popular polls report that 84% of Americans believe praying for the sick improves their chances of recovery from illness; 70% say they pray often for the health of a family member. (2) National surveys report that 43% of Americans acknowledged having prayed for their own health, 24% have had someone else pray for their health, 10% have participated in a prayer group, and 2% have participated in a healing ritual for their own health. (3,4) Beyond prayer, faith or spiritual healing involves specific practices that often take place in communal settings and involve healers who conduct prayer or ritual services. (5) Although faith-based healing (FBH) is generally considered to fall solely within the context of religious tradition or practice, (6) FBH might be related, as well, to attitudes and practices regarding medical care. (7) Therefore, in addition to its religious context, FBH use may have implications for health and healthcare services use.

Prior research on faith-based healing has examined the relationship of this practice with sociodemographic variables but not with medical care service demand, utilization or access. A survey using a convenience sample of patients in family physicians' offices in rural North Carolina found that African Americans and those with less than a high school education were more likely to participate in faith-healing services. (8) In addition, a population-based study of adults found that a strong belief in spiritual healing was associated with being female and African-American, and having poorer health status, lower education level, and low income. (9) The relationship between access to medical care, health services utilization, satisfaction with traditional medical care, and the use of FBH has not been explored. FBH has been viewed as a complementary or alternative medical (CAM) therapy by some (3,4,10) and patients may utilize CAM therapies for a variety of reasons, such as personal dissatisfaction with medical care, need for personal empowerment over their own health, and congruence with their existing belief structure. (11,12) If FBH is placed within a CAM framework, the relationship with these factors has not been explicitly studied to date. Perhaps the use of FBH affects one's use of medical care services and attitudes toward medical care, or, conversely, access and barriers to medical care and their attitudes about medical care impact one's use of FBH.

[FIGURE 1 OMITTED]

The overall goal of this study is to describe faith-based healing use and its relationship with medical care use among residents of the rural southeastern United States, a population with the highest reported utilization rates for prayer as a healing modality. (3) The principal aims are to determine the prevalence of FBH usage, to examine the association of FBH with patient demographic characteristics, medical service utilization, and attitudes about medical care, and to test the hypothesis that FBH substitutes for traditional medical care in this population.

Methods

Study Design

This study uses cross-sectional telephone survey data gathered as part of an evaluation of the Southern Rural Access Program (SRAP). Beginning in 1998, the Robert Wood Johnson Foundation sponsored this initiative to improve access to medical care in selected southeastern rural counties. (13) Underserved portions of eight states were targeted for the SRAP intervention, including Alabama, Arkansas, Georgia, Louisiana, Mississippi, South Carolina, Texas and West Virginia (Fig. 1).

As part of an evaluation of the SRAP, baseline data on access to health care and use of services were gathered by a telephone survey of residents in the rural counties targeted by the SRAP. Approximately 600 adults from each state were surveyed, with 15 to 40 surveys completed in each county. Between October 2002 and July 2003, Professional Research Consultants of Omaha, Nebraska (www.prconline.com) conducted the survey using computer-assisted telephone interviews with eligible adults who had lived in the targeted area for at least the previous 12 months and spoke either English or Spanish. Calls were made primarily in the evening, but those not answering on the first attempt were called again at all times of the day to minimize possible response rate bias due to omission of those involved in shift work or with other set daily schedules. At least ten attempts were made to each randomly selected number. A secondary randomization scheme was used to identify the specific individual in each household to be surveyed: if that specific individual refused to participate, the household was deemed a "nonrespondent." The protocol for this study was reviewed and exempted by the University of North Carolina School of Medicine's Committee on the Protection of Human Subjects.

Survey Items

The main outcome variable of this study was modified from an item included in the Survey on Disparities in Quality of Health Care. (14) Respondents provided a dichotomous (yes or no) response to the question, "In the past 12 months, have you been to see a healer who used prayer as their primary treatment method?" Selected sociodemographic indicators (age, gender, income, race, education, employment, marital status) were analyzed as potential correlates. The Community Tracking Survey (CTS), (15) National Evaluation of Rural Primary Health Care Programs (16) and the Medical Expenditure Panel Survey (MEPS) (17) provided general and specific measures of patient satisfaction with medical care. Several standard health service questions were included from these surveys as well as from the Behavioral Risk Factor Surveillance System (BRFSS) (18) and the National Health Interview Survey (NHIS). (19) Items exploring respondents' beliefs and values regarding medical care were utilized as well. (20)

Data Analysis

Analyses were limited to individuals who reported their race as African-American or white, since too few respondents (less than 4%) reported other races or ethnicity to enable appropriate analysis. A total of 184 individuals were not included from the original sample for this reason (111 reporting "other" races and 73 with missing data for this variable). To adjust for differential response rates across demographic groups and for over-sampling in smaller counties, all data were weighted with regard to race, gender, age, income and county sampling probabilities to proportionately represent the target adult population, per 2000 US Census data. The final weighted analysis included 4,695 subjects.

The sample was described in terms of sociodemographic measures, self-reported health status and state of residence, and along four domains related to medical care access (utilization of medical care, barriers to medical care, satisfaction with medical care and valuing of medical care). Income values were missing for 622 of the 4,879 (13%) respondents and were imputed based on a hotdeck procedure using education, age, sex, and race. Pearson's chi-square and unadjusted logistic regression were used to compare users and nonusers of FBH on sociodemographic characteristics. Multivariate logistic regression models were fit to test the associations between various indicators of the four domains of medical care access while adjusting for age, gender, income, race, employment status, and health status. The variables related to use of FBH from all of the previous models at or below the level of P [less than or equal to] 0.20 were combined in a final logistic model to derive a "best" predictive model of FBH use. Statistical significance was set using a two-sided significance level of P [less than or equal to] 0.05 and all statistical analyses were completed using Stata Version 8.2 statistical software. (21)

Results

The overall response rate for the parent SRAP survey was 51.0%, with 4,879 respondents. (22) Participation rates were lower among African Americans, those with household incomes less than $15,000, males and those aged 18 to 39 years, and we weighted the sample to adjust for these differences in response rates. (23,24) The study population was predominantly female, white, and had attained at least a high school education (Table 1). Over 75% of respondents reported their own health status to be "good" or "excellent." Of the final sample of 4,695 respondents, 186 subjects (weighted value = 193, weighted proportion = 4.1% [weighted numbers are provided in the remainder of this paper]) reported that they had seen a faith-based healer within the past year. Of those reporting FBH use, 121 users were female and 72 were male, while 120 whites and 73 blacks reported use. (Table 1) Relative FBH usage rates varied by over 100% across states, but absolute rate differences were small, varying between 2.5% and 6.5%, with residents of South Carolina, Texas and West Virginia having the highest prevalence (Fig. 2).

Bivariate analyses revealed no difference between users and nonusers of FBH in regards to self-reported race, educational attainment or marital status. Users of FBH, however, were more likely to be younger, women, have annual household incomes less than $50,000, be unemployed, and report poorer health. After adjusting for other sociodemographic variables, only those who were younger and those who reported fair or poor health status remained more likely to use FBH (Table 2).

Utilization Of Medical Care

In bivariate analyses, respondents who had either foregone or delayed needed medical care within the previous 12 months were more likely to use FBH. Logistic regression models controlling for demographic variables also supported these differences. Those who reported seeing a physician for four or more visits within the past year were more likely users of FBH in the bivariate analysis, but this association was no longer significant in adjusted analyses (Table 3).

Barriers to Medical Care

Respondents with a break in health insurance coverage in the previous 12 months, those who reported cost as a problem in obtaining care, and those who reported difficulties with travel to their primary medical care location were more likely to be users of FBH in bivariate analysis. After adjustment for sociodemographic factors, only those who reported cost as a barrier to obtaining care remained more likely to be users of FBH (Table 3).

Satisfaction with Medical Care

Bivariate analysis indicated that respondents who were dissatisfied with the quality of their medical care were more likely to use FBH. However, after adjusting for sociodemographic factors, there were no statistically significant differences with regard to any care satisfaction variable (Table 3).

Valuing of Medical Care

Those who agreed with the statement "a person should use doctors only as a last resort" were more likely to use FBH in both bivariate and adjusted analyses. Respondents who agreed with either this statement or another statement that said, "if a person waits long enough, they can get over almost any illness without medical care" were twice as likely to use FBH than those who did not agree with either statement (Table 3).

Final Regression Model

A multivariate logistic regression model was fit to describe factors that were independently associated with faith healer usage. Sociodemographic variables, including age, gender, race, income and education were included, along with variables that had shown statistically significant relationships with FBH use in prior regression models. The model demonstrated that younger age, female gender, lower self-reported health status and lower values placed on medical care use were independently associated with FBH use (Table 4). Using these four variables, we developed a predictive model to assess the relationships between age, belief status and use of FBH. Figure 3 depicts the predicted probabilities for FBH usage, with the younger subjects more likely to use FBH, after adjusting for gender, health status and belief about avoiding medical care. This figure also depicts differences among age groups regarding beliefs about avoiding medical care (adjusted for gender and health status): the effect of such beliefs on the use of FBH is largest in the 45 to 64 and 65 to 74 year age groups, is of lower magnitude but present among 18 to 44 year olds, and is no longer present in the oldest group.

Discussion

The purpose of this study was to define the prevalence of FBH use in the rural southeastern United States and to describe the relationship between FBH use and people's experiences with and attitudes about medical care services. We found that only 4.1% of residents in this population reported having received care from a practitioner who used prayer as their primary treatment modality and use was notably more common among women, younger individuals, and those who reported poorer health status. We also found that FBH use was more common among individuals who had experienced problems accessing traditional medical care, specifically those who had a break in health insurance coverage, those who encountered cost barriers and those who had foregone or delayed needed medical care. FBH use was also higher among those who believed in delaying medical care and avoiding seeing physicians.

If seeking care from a faith-based healer is equivalent to participating in a healing ritual for personal health, our results are similar to that reported in a national study. (4) However, the overall FBH prevalence is less than that reported for more global measures of the use of prayer for healing purposes. (3) Although residents of the rural southeast hold generally stronger religious beliefs as a population, they do not appear more likely to use FBH than residents of other regions, (4) but our measure of FBH use may not have captured all use of FBH in this population.

The relationship between age and use of FBH is consistent with demographic patterns of use for complementary therapies, (25) specifically mind-body therapies, (26) and prayer in health settings. The significantly higher prevalence of FBH use among those with fair or poor self-reported health may indicate that FBH is used more often among those with chronic health conditions and is consistent with previous studies that have shown that prayer and other CAM modalities are used more often by those with chronic conditions. These other studies also find that women are more likely to use prayer and CAM modalities, as they were more likely to use FBH in our study. (3,12,27) However, no relationship was found in our study between race or socioeconomic status and the use of FBH, which differs from some other studies regarding prayer and its use in health-related settings. (4,9,28) We also did not find evidence that FBH usage was independently related to dissatisfaction with medical care.

Studies show that users of CAM therapies in general have more physician visits and higher levels of unmet need for medical care. (27) While we did not find a relationship between FBH use and the number of physician visits, FBH was more commonly used by those experiencing problems with accessing medical care. Some studies have suggested a relationship between lack of health insurance and a higher likelihood of using CAM therapies. (29) Lack of insurance may be more directly tied to the use of complementary modalities that have lower out-of-pocket costs, such as FBH. In addition, having health insurance may be inversely associated with use of complementary therapies when the marginal economic cost of these therapies is high (30,31) or when satisfaction with medical care is low.

The decision by patients to use FBH as an adjunct to, or a substitute for, traditional medical care is a very complex one, and our data hint at a nuanced substitution effect. FBH use does not appear to be a matter of dissatisfaction with medical care in our study population, but rather seems to be tied to the cost of traditional medical care and to attitudes suggesting avoidance or delay of such care. While the absolute magnitude of the differences by age, economic barriers and beliefs about medical care are small, the relative differences by these characteristics are practically meaningful. For example, people less than 45 years of age are about five times more likely to be using a FBH than those over age 75 (Fig. 3). Recognizing high-use groups can help clinicians appropriately target inquiries about the use of FBH.

Limitations

Several limiting factors of this study should be noted. The response rate of 51% for this survey may introduce significant response bias. This response rate is consistent with that of similarly structured telephone health surveys, including the CDCs Behavioral Risk Factor Surveillance Survey after which this survey was patterned and for which the median state response rate in 2000 was 48.9%. (32) In a random-digit-dialed telephone survey, it is difficult to assess demographic differences between respondents and nonrespondents and we statistically weighted analyses to correct for these demographic differences. (33) Although lower socioeconomic status is associated with a lower likelihood of owning a telephone, the effect of these differences on individual response items has been found to be small (34) and our demographically weighted analyses adjusted for this. Furthermore, we found no relationship between racial or economic variables and use of FBH in our study, so we do not believe that response likelihood is related to use of FBH.

We were unable to more fully explore the context in which FBH was used by respondents, since the questionnaire included no measures of religiosity or items regarding the context of the faith-based healing interventions. In addition, the survey specifically did not ask about use of self-prayer or participation in prayer circles or prayer groups, which may be more common among all populations than seeing a practitioner who performs faith-based healing. (4)

This study was limited to those who self-identified as either black or white. Those who self-identified as being from another racial group constituted less than 0.5% of the total survey respondents, and the populations in each of these other groups were too small to permit valid statistical analysis. These analyses, therefore, do not tell us about FBH use among rural Asians and Native Americans.

Finally, because many variables were included describing access to medical care, a large number of statistical tests were performed in this study. Multiple statistical tests can lead to statistically significant results that arise due to chance alone. We did not lower the threshold of statistical significance, because we felt that it was more important in this study on a previously unexplored topic to potentially over-identify correlates than to possibly miss important correlates by making the threshold of significance too stringent.

Conclusions

In summary, we found that the prevalence of use of FBH in the rural southeastern United States is low, at only 4%. Younger adults, women, and those with poorer self-rated health are more likely to use FBH. We also found that the use of FBH was associated with individuals who experienced problems paying for, or placed less value in, the care of traditional medical practitioners. Economic barriers to medical care and patients' beliefs and attitudes about medical care may influence their decision to seek care from a practitioner of faith-based healing. Our findings also point out to physicians working in the rural southeast that approximately one in 20 patients will also be seeking care from a faith-based healer. Because of the importance of people's religious beliefs and its importance to their approach to health and illness, clinicians should consider querying their patients' use of faith-based healers, especially among those most likely to use FBH. Finally, the relationship between the use of FBH and patient attitudes about, use of, and barriers to medical care can potentially assist policy makers and health planners to better understand deficiencies in the US medical care system.

[FIGURE 3 OMITTED]

Acknowledgments

The authors would like to thank Dr. Edward Brooks and Dr. Sally Stearns of the University of North Carolina School of Public Health, Department of Health Policy and Administration, for their input and guidance on an earlier version of this manuscript.

References

1. Koenig HG, McCullough ME, Larson DB. Handbook of Religion and Health. New York, Oxford University Press; 2001.

2. Kalb C. Faith & Healing. Newsweek. 2003 November 10, 2003:44-56.

3. McCaffrey AM, Eisenberg DM, Legedza AT, et al. Prayer for health concerns: results of a national survey on prevalence and patterns of use. Arch Intern Med 2004;164:858-862.

4. Barnes PM, Powell-Griner E, McFann K, et al. Complementary and alternative medicine use among adults: United States, 2002. Adv Data 2004 May;27:1-19.

5. Glik DC. Psychosocial wellness among spiritual healing participants. Soe Sci Med 1986;22:579-586.

6. Glik DC. Symbolic, ritual and social dynamics of spiritual healing. Soc Sci Med 1988;27:1197-1206.

7. Swan R. Faith healing, Christian Science, and the medical care of children. N Engl J Med 1983;309:1639-1641.

8. King DE, Sobal J, DeForge BR. Family practice patients' experiences and beliefs in faith healing. J Fam Pract 1988;27:505-508.

9. Mansfield CJ, Mitchell J, King DE. The doctor as God's mechanic? Beliefs in the Southeastern United States. Soc Sci Med 2002;54:399-409.

10. Eisenberg DM, Davis RB, Ettner SL, et al. Trends in alternative medicine use in the United States, 1990-1997: results of a follow-up national survey. JAMA 1998;280:1569-1575.

11. Shumay DM, Maskarinec G, Kakai H, et al. Why some cancer patients choose complementary and alternative medicine instead of conventional treatment. J Fam Pract 2001;50:1067.

12. Astin JA. Why patients use alternative medicine: results of a national study. JAMA 1998;279:1548-1553.

13. Rural Health Policy Center PSU. Southern Rural Access Program. Available at: http://www.srap.org. Accessed July 4, 2004.

14. The Commonwealth Fund. Survey on Disparities in Quality of Health Care: Spring 2001. Available at: http://www.cmwf.org/usr_doc/qualitysurvey_2001_questionnaire.pdf. Accessed September 28, 2004.

15. Center for Studying Health System Change. Design and Methods for the Community Tracking Survey. Available at: http://www.hschange.org/index.cgi?data=01. Accessed September 28, 2004.

16. Sheps CG, Wagner EH, Schonfeld WH, et al. An evaluation of subsidized rural primary care programs: I. A typology of practice organizations. Am J Public Health 1983;73:38-49.

17. Agency for Healthcare Research and Quality. The Medical Expenditure Panel Survey. Available at: http://www.meps.ahrq.gov/. Accessed September 28, 2004.

18. Centers for Disease Control and Prevention. CDC's Behavioral Risk Factor Surveillance System. Available at: http://www.cdc.gov/brfss/. Accessed September 28, 2004.

19. Centers for Disease Control and Prevention. National Health Interview Survey (NHIS). Available at: http://www.cdc.gov/nchs/nhis.htm. Accessed September 28, 2004.

20. Sharp K, Ross CE, Cockerham WC. Symptoms, beliefs, and the use of physician services among the disadvantaged. J Health Soc Behav 1983; 24:255-263.

21. Stata Corporation. Stata Statistical Software, Version 8.2. College Station, Texas, 2004.

22. The American Association for Public Opinion Research. Standard Definitions: Final Dispositions of Case Codes and Outcome Rates for Surveys. Available at: http://www.aapor.org. Accessed October 19, 2004.

23. Thomas R, Purdon S. Telephone Methods for Social Surveys. Social Research Update, Issue Eight. 1994. Available at: http://www.soc. surrey.ac.uk/sru/SRU8.html. Accessed October 19, 2004.

24. CyBulski KA, Sinclair M, Potter F, Ciemnecki AB. Adjusting for Nonresponse Among Medicaid Households That Could Not Be Located or Were Located but Did Not Participated in the Minnesota Managed Care Survey. Available at: http://www.jpsm.umd.edu/icsn/papers/cybulski.htm. Accessed October 19, 2004.

25. Kessler RC, Davis RB, Foster DF, et al. Long-term trends in the use of complementary and alternative medical therapies in the United States. Ann Intern Med 2001;135:262-268.

26. Wolsko PM, Eisenberg DM, Davis RB, et al. Use of mind-body medical therapies. J Gen Intern Med 2004;19:43-50.

27. Paramore LC. Use of alternative therapies: estimates from the 1994 Robert Wood Johnson Foundation National Access to Care Survey. J Pain Symptom Manage 1997;13:83-99.

28. Simile C, Ni H. Use of Spiritual Healing Therapy in Relation to Race and Ethnicity. 2002 Data Users Conference, July 15-17, 2002. Washington, DC, Centers for Disease Control and Prevention/National Center for Health Statistics, 2002.

29. Mackenzie ER, Taylor L, Bloom BS, et al. Ethnic minority use of complementary and alternative medicine (CAM): a national probability survey of CAM utilizers. Altern Ther Health Med 2003;9:50-56.

30. Wolsko PM, Eisenberg DM, Davis RB, et al. Insurance coverage, medical conditions, and visits to alternative medicine providers: results of a national survey. Arch of Intern Med 2002;162:281-287.

31. Ni H, Simile C, Hardy AM. Utilization of complementary and alternative medicine by United States adults: results from the 1999 national health interview survey. Med Care 2002;40:353-358.

32. Mokdad AH, Stroup DF, Giles WH, Behavioral Risk Factor Surveillance Team. Public health surveillance for behavioral risk factors in a changing environment. Recommendations from the Behavioral Risk Factor Surveillance Team. MMWR Recomm Rep 2003;52 (RR09):1-12.

33. Centers for Disease Control and Prevention. 2003 Behavioral Risk Factor Surveillance System Summary Data Quality Report. Available at: http://www.cdc.gov/brfss/technical_infodata/pdf/2003SummaryDataQualityReport.pdf. Accessed November 18, 2004.

34. Keeter S, Miller C, Kohut A, et al. Consequences of reducing nonresponse in a national telephone survey. Public Opin Q 2000;64:125-148.

Sharon K. Hull, MD, MPH, Timothy P. Daaleman, DO, MPH, Samruddhi Thaker, MBBS, MHA, and Donald E. Pathman, MD, MPH

From the Department of Medical Humanities and Department of Family and Community Medicine, Southern Illinois University School of Medicine; the Department of Family Medicine, University of North Carolina School of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, NC; the Cecil G. Sheps Center for Health Services Research, University of North Carolina at Chapel Hill, Chapel Hill, NC; and the Department of Health Policy and Administration, School of Public Health, University of North Carolina at Chapel Hill, Chapel Hill, NC.

Reprint requests to Sharon K. Hull, MD, MPH, Research Associate Professor, Department of Medical Humanities, Southern Illinois University School of Medicine, PO Box 19620, Springfield, Illinois 62794-9603. E-mail: [e-mail].

This study was supported by grants from the Robert Wood Johnson Foundation (#036829) and the Health Resources and Services Administration (#T32-HP14001). Dr. Hull's work on this paper was substantially completed as part of an NRSA Primary Care Research Fellowship at the Cecil G. Sheps Center for Health Services Research and the Department of Family Medicine, University of North Carolina School of Medicine. Results from this paper were presented in part at the American Public Health Association Annual Meeting, November, 2004.

RELATED ARTICLE: Key Points

* Prevalence of the use of faith-based healers (FBH) is 4.1% among residents of the rural southeastern United States.

* Younger adults, women and those with poorer health status are significantly more likely to use FBH than older adults, men and those with better health status.

* FBH use is also related to lack of health insurance coverage, cost barriers to receiving medical care, beliefs about avoiding medical care, and having delayed or foregone medical care.

* Providers should explore patients' use of faith-based healing practices in patient encounters.

* Policy makers should consider the growing use of alternative therapies, including the use of FBH, in the context of health care services demand, utilization and access.

Table 1. Demographic characteristics of survey respondents
(N = 4695) (a)

                                     Weighted number (%) of subjects

                                     Range 18-94 Mean
Age (years)                          (SD) = 49.9 (16.6)

  18-44                              1843 (39.3)
  45-64                              1851 (39.4)
  65-74                               592 (12.6)
  75 and older                        409 (8.7)
Gender
  Female                             3148 (67.1)
  Male                               1547 (32.9)
Income
  <$50,000                           3604 (76.8)
  [greater than or equal to]$50,000  1091 (23.2)
Race
  Black                              1383 (29.5)
  White                              3312 (70.5)
Education (highest attained)
  Less than college                  2758 (58.7)
  Some college or higher             1923 (41.0)
Employment status
  Employed                           2435 (51.9)
  Not employed                       2234 (47.6)
Marital status
  Married                            2591 (55.2)
  Not married                        2085 (44.4)
Self-reported health status
  Good or excellent                  3475 (74.0)
  Fair or poor                       1203 (25.6)
State of residence
  Alabama                             584 (12.4)
  Arkansas                            611 (13.0)
  Georgia                             580 (12.4)
  Louisiana                           581 (12.4)
  Mississippi                         595 (12.7)
  South Carolina                      583 (12.4)
  Texas                               565 (12.0)
  West Virginia                       596 (12.7)

(a) Less than 1% missing data for all variables except income. For the
13% of respondents who did not report a household income, this variable
was imputed using a hotdeck procedure based on respondents' reported
education, age, sex, and race.

Prevalence of Use of Faith-based Healers by State p=0.009

                Prevalence

Alabama         2.5%
Arkansas        3.7%
Georgia         3.5%
Louisiana       3.9%
Mississippi     2.6%
South Carolina  5.3%
Texas           4.9%
West Virginia   6.5%

Fig. 2 Prevalence of use of faith-based healers by state, P = 0.009

Note: Table made from bar graph.

Table 2. Use of FBH among sociodemographic groups (N=4695)

                                                           Adjusted
                                            Unadjusted OR  OR (a)
                              Use of FBH    ([CI.sub.95])  ([CI.sub.95])
                              weighted n    for use of     for use of
                              (%)           FBH            FBH

Age (years)
  18-44                       102 (4.6)     2.53 (0.88,    4.04 (1.3,
                                              7.25)          11.92) (c)
  45-64                        62 (4.0)     2.23 (0.75,    2.89 (0.94,
                                              6.66)          8.88)
  65-74                        21 (4.4)     2.45 (0.75,    2.45 (0.75,
                                              7.96)          8.00)
  75 and older (b)              6 (1.9)     1.0 --         1.0 --
Gender
  Female                      121           1.42           1.26
                                (4.8) (c)
  Male (b)                     72 (3.4)     (1.01,         (0.99, 1.80)
                                              1.98) (c)
Income
  <$50,000                    161           1.62           1.36
                                (4.6) (c)
  [greater than or equal to]   32 (2.9)     (1.03,         (0.84, 2.22)
    $50,000 (b)                               2.54) (c)
Race
  Black                        73 (4.5)     1.15           0.99
  White (b)                   120 (4.0)     (0.81, 1.65)   (0.66, 1.48)
Education (highest attained)
  Less than college           116 (4.3)     1.07           0.84
  Some college or higher (b)   75 (4.0)     (0.76, 1.51)   (0.59, 1.19)
Employment status
  Not employed                107           1.56           1.51
                                (5.1) (c)
  Employed (b)                 84 (3.3)     (1.08,         (0.98, 2.31)
                                              2.27) (c)
Marital status
  Married                     103 (4.3)     1.07           1.26
  Not married (b)              88 (4.0)     (0.80, 1.42)   (0.91, 1.75)
Self-reported health status
  Fair or poor                 73           1.89           1.77
                                 (6.4) (d)
  Good or excellent (b)       121 (3.5)     (1.29,         (1.20,
                                              2.78) (d)      2.61) (d)

(a) Adjusted for age, gender, income, employment status, health stanus,
and race (Where one of these is the dependent variable, it is not
included as a covariate in the regression model.).
(b) Referent group
(c) Statistically significant at P < 0.05.
(d) Statistically significant at P < 0.01.

Table 3. Faith-based healer (FBH) use and access to medical care (N =
4695) (a)

                                            No. of
                                            (% of all)   Use of FBH n
                                            subjects     (%)

Utilization of Medical Care

In the past 12 months, was there any time
    that you did not get the medical care
    you thought you needed?
  Yes                                        586 (12.7)   45 (7.7) (e)
  No (c)                                    4044 (87.3)  148 (3.7)
In the past 12 months, was there any time
    that you delayed getting the medical
    care you thought you needed?
  Yes                                       1415 (30.5)   85 (6.0) (e)
  No (c)                                    3222 (69.5)  107 (3.3)
No. of physician visits in past year
  4 or more                                 1827 (40.7)   96 (5.2) (e)
  Fewer than 4 (c)                          2660 (59.3)   88 (3.3)

Barriers to Medical Care

Lives in whole- or part-county HPSA (2002)
  Yes                                       3587 (77.2)  161 (4.5)
  No (c)                                    1061 (22.8)   32 (3.0)
Usual source of care is ER
  Yes                                        380 (8.2)    23 (6.2)
  No or not known (c)                       4268 (91.8)  170 (4.0)
During the past 12 months, was there a
    break in health insurance coverage?
  Yes                                       1388 (30.2)   78 (5.6) (e)
  No or not known (c)                       3202 (69.8)  113 (3.5)
Ease of traveling to place of care
  Not easy                                   538 (11.6)   34 (6.3) (d)
  Easy (c)                                  4087 (88.4)  160 (3.9)
Cost is a problem in obtaining care
  Somewhat or great problem                 1111 (23.7)   67 (6.0) (e)
  Minor or not a problem (c)                3528 (75.1)  124 (3.6)
Feels that race is a barrier to care
  Agrees                                    1443 (33.7)   57 (3.9)
  Does not agree (c)                        2832 (66.3)  124 (4.4)

Satisfaction with Medical Care

Overall satisfaction with care
  Not satisfied                              366 (8.0)    18 (4.9)
  Satisfied (c)                             4239 (92.0)  172 (4.1)
Satisfaction with quality of care provided
  Not satisfied                              307 (6.7)    21 (6.9) (d)
  Satisfied (c)                             4301 (93.3)  167 (3.9)
Confidence in provider ability
  Somewhat to not confident                  944 (20.5)   42 (4.5)
  Mostly or very confident (c)              3658 (79.5)  149 (4.1)
Satisfaction with concern shown by
    provider
  Not satisfied                              298 (6.5)    15 (5.0)
  Satisfied (c)                             4310 (93.5)  177 (4.1)
Satisfaction that questions were answered
  Not satisfied                              291 (6.3)    13 (4.6)
  Satisfied (c)                             4314 (93.7)  177 (4.1)
Satisfaction with feeling welcome and
    comfortable
  Not satisfied                              300 (6.5)    14 (4.5)
  Satisfied (c)                             4322 (93.5)  178 (4.1)

Valuing of Medical Care

Believes that a person should use doctors
    as a last resort
  Agrees                                    1226 (27.1)   68 (5.6) (e)
  Neutral or disagrees (c)                  3303 (72.9)  120 (3.7)
Believes that, if a person waits long
    enough, one can get over almost any
    illness without medical care
  Agrees                                     980 (21.6)   46 (4.7)
  Neutral or disagrees (c)                  3551 (78.4)  140 (3.9)
Believes one should delay or avoid using
    physicians (factor combination of
    previous two responses)
  Yes                                        818 (17.7)   53 (6.5) (e)
  No or not known (c)                       3806 (82.3)  138 (3.6)

                                            Unadjusted OR  Adjusted OR
                                            ([CI.sub.95])  ([CI.sub.95])
                                            for use of     for use of
                                            FBH            FBH (b)

Utilization of Medical Care

In the past 12 months, was there any time
    that you did not get the medical care
    you thought you needed?
  Yes                                       2.19           1.76
  No (c)                                    (1.47,         (1.11,
                                              3.28) (e)      2.79) (e)
In the past 12 months, was there any time
    that you delayed getting the medical
    care you thought you needed?
  Yes                                       1.87           1.65
  No (c)                                    (1.33,         (1.14,
                                              2.63) (e)      2.39) (e)
No. of physician visits in past year
  4 or more                                 1.61           1.21
  Fewer than 4 (c)                          (1.14,         (0.83,
                                              2.27) (e)      1.75)

Barriers to Medical Care

Lives in whole- or part-county HPSA (2002)
  Yes                                       1.51           1.50
  No (c)                                    (0.90, 2.52)   (0.90, 2.49)
Usual source of care is ER
  Yes                                       1.58           1.46
  No or not known (c)                       (0.86, 2.91)   (0.76, 2.81)
During the past 12 months, was there a
    break in health insurance coverage?
  Yes                                       1.64           1.41
  No or not known (c)                       (1.14,         (0.92, 2.14)
                                              2.35) (e)
Ease of traveling to place of care
  Not easy                                  1.65           1.36
  Easy (c)                                  (1.03,         (0.81, 2.28)
                                              2.63) (d)
Cost is a problem in obtaining care
  Somewhat or great problem                 1.71           1.47
  Minor or not a problem (c)                (1.25,         (1.05,
                                              2.36) (e)      2.07) (d)
Feels that race is a barrier to care
  Agrees                                    0.89           0.91
  Does not agree (c)                        (0.61, 1.29)   (0.61, 1.34)

Satisfaction with Medical Care

Overall satisfaction with care
  Not satisfied                             1.21           1.05
  Satisfied (c)                             (0.66, 2.22)   (0.56, 1.97)
Satisfaction with quality of care provided
  Not satisfied                             1.83           1.55
  Satisfied (c)                             (1.03,         (0.85, 2.84)
                                              3.23) (d)
Confidence in provider ability
  Somewhat to not confident                 1.11           0.95
  Mostly or very confident (c)              (0.71, 1.74)   (0.59, 1.53)
Satisfaction with concern shown by
    provider
  Not satisfied                             1.23           1.04
  Satisfied (c)                             (0.58, 2.63)   (0.48, 2.26)
Satisfaction that questions were answered
  Not satisfied                             1.12           0.91
  Satisfied (c)                             (0.52, 2.42)   (0.40, 2.10)
Satisfaction with feeling welcome and
    comfortable
  Not satisfied                             1.10           1.00
  Satisfied (c)                             (0.55, 2.21)   (0.49, 2.05)

Valuing of Medical Care

Believes that a person should use doctors
    as a last resort
  Agrees                                    1.56           1.63
  Neutral or disagrees (c)                  (1.12,         (1.17,
                                              2.17) (d)      2.27) (e)
Believes that, if a person waits long
    enough, one can get over almost any
    illness without medical care
  Agrees                                    1.19           1.26
  Neutral or disagrees (c)                  (0.79, 1.81)   (0.85, 1.89)
Believes one should delay or avoid using
    physicians (factor combination of
    previous two responses)
  Yes                                       1.86           2.06
  No or not known (c)                       (1.30,         (1.43,
                                              2.67) (e)      2.97) (e)

(a) Less than 4% missing data for all variables except "Feels that race
is a barrier to care" (8.5% missing for this variable).
(b) Adjusted for age, gender, income, race, employment status, and
health status.
(c) Referent group.
(d) Statistically significant at P < 0.05.
(e) Statistically significant at P < 0.01.
FBH, faith-based healing; HPSA, health professional shortage area.

Table 4. Final multivariate regression model adjusted odds ratios for
use of faith-based healers (FBH) (N = 4695)

                                           Adjusted odds
                                           of using a FBH  [CI.sub.95]

Age (years)
  18-44                                    7.21            2.00, 25.94
  45-64                                    5.89            1.52, 22.13
  65-74                                    6.20            1.58, 24.43
  75 and older (a)                         1.00            --
Female gender                              1.49            1.03, 2.14
Black race                                 0.88            0.58, 1.33
Income <$50,000                            1.42            0.88, 2.29
Less than college education                0.77            0.54, 1.11
Self-reported health status fair or poor   1.83            1.19, 2.83
Did not get needed health care (foregone   1.22            0.78, 1.89
  care)
Delayed getting needed health care         1.29            0.89, 1.87
Break in insurance coverage                1.08            0.69, 1.69
Cost is a problem in obtaining care        1.36            0.97, 1.91
Satisfied with quality of health care      1.23            0.68, 2.22
Believes one should delay or avoid seeing  1.82            1.24, 2.67
  physicians

(a) Referent group