Frequency of Attendance at Religious Services, Hypertension, and Blood Pressure: The Third National Health and Nutrition Examination Survey R. FRANK GILLUM, MD, MS,
AND
DEBORAH D. INGRAM, PHD
Objective: To test the hypothesis that frequency of attendance at religious services is inversely related to prevalence of hypertension and blood pressure level. Methods: In the Third National Health and Nutrition Examination Survey (NHANES III), 14,475 American women and men aged 20 years and over reported frequency of attendance at religious services, history of hypertension treatment, and had blood pressure (BP) measured. Results: The percentage reporting attending religious services weekly (52 times/yr) was 29 and more than weekly (⬎52 times/yr) was 10. Prevalence of hypertension (systolic BP ⱖ140 or diastolic BP ⱖ90 mm Hg or current use of blood pressure medication) was 21% in never at attenders, 19% in those attending less than weekly (1–51 times/yr), 26% in those attending weekly, and 26% in those attending more than weekly (p ⬍ .01). After controlling for sociodemographic and health variables, religious attendance was associated with reduced prevalence compared with nonattendance, significantly so for weekly ( ⫽ ⫺0.24; 95% confidence interval [CL], ⫺0.37 to ⫺0.11; p ⬍ .01) and more than weekly ( ⫽ ⫺0.33; 95% CL, ⫺0.60 to ⫺0.07; p ⬍ .05). No significant effect modification by gender or age was observed. Compared with never attenders, persons attending weekly had a systolic BP 1.46 mm Hg (95% CL 2.33, 0.58 mm Hg, p ⬍ .01) lower and persons attending ⬎52 times/yr had systolic BP 3.03 mm Hg (95% CL 4.34, 1.72 mm Hg, p ⬍ .01) lower. No significant effect modification by gender was observed; these estimates are adjusted for a significant interaction between age and less than weekly attendance (1–51 times) (p ⬍ .05). Conclusions: Compared with never attending, attendance at religious services weekly or more than weekly was associated with somewhat lower adjusted hypertension prevalence and blood pressure in a large national survey. Key words: hypertension, religion, blood pressure, epidemiologic methods. BMI ⫽ body mass index; CL ⫽ confidence limits; NHANES III ⫽ Third National Health and Nutrition Examination Survey; SBP ⫽ systolic blood pressure.
INTRODUCTION ypertension, a prevalent and well-established cardiovascular risk factor, is associated with social and cultural factors (1– 6). In 1999 to 2000, prevalence of hypertension in US adults aged 65 to 74 was 84% in women and 73% in men (3). Although prevalence is greatest among the elderly, recent increases in obesity prevalence might be expected to increase prevalence in younger and older adults. Hence, after declines between 1975 to 1980 and 1988 to 1994, by 1999 to 2000 hypertension prevalence had increased again (3). Religious affiliation, attendance at religious services, and other religious behavior is more prevalent in the United States than in any other industrialized nation (7–9). Yet few studies using population-based data have been published that examine the association between this important social factor and hypertension or blood pressure in the US population (10 – 19). Further, although several studies have shown an inverse association of attendance at religious services or church membership and blood pressure or hypertension (10 –16), others have not (10,17–19). For example, in a cohort of persons over 65 years of age, persons who attended church weekly and prayed or studied the Bible daily were 40% less likely to have diastolic hypertension (ⱖ90 mm Hg) than others and, if hypertensive, were more likely to be treated (11). It is important to examine the reported association in large, national samples. In order to test the hypothesis that frequent attendance at religious services is inversely associated with prevalence of hypertension and with systolic blood pressure (SBP) among adults of all ages, independent of sociodemographic variables and hy-
H
pertension risk factors in the American population, data on a large, multiethnic, national sample of adults from the Third National Health and Nutrition Examination Survey (NHANES III) were examined. METHODS The NHANES III was conducted in 1988 to 1994 on a nationwide multistage probability sample of the civilian, noninstitutionalized US population aged 2 months and over, excluding reservation lands of American Indians. Data were collected through personal interview, physical examination, and laboratory analyses. Details of the plan, sampling, operation, and response have been published, as have procedures used to obtain informed consent and to maintain confidentiality of information obtained (20 –23). This analysis includes persons who were 20 years and over at the time of interview. Of the 18,825 persons aged 20 years and over, we excluded from the analysis 338 (1.8%) pregnant women, 938 (5.0%) persons who had a history of heart attack, 649 (3.4%) who had a history of stroke, and 757 (4.0%) who had a history of heart failure. Also excluded from this analysis were 1,668 (8.9%) persons with missing data for one or more of the analysis variables. After all exclusions, 14,475 persons aged 20 and over remained for analysis. At the home interview, participants were asked, “How often do you attend church or religious services?” Values ranged from 0 to 1,825 times per year. Values in excess of 365 (n ⫽ 7) were coded missing, as were four “don’t know” responses. For comparability with previous reports and because the large number of 0 responses made use of the continuous distribution unsuitable, number of religious services attended/yr was categorized. The four categories were 0 (never), 1 to 51, 52 (weekly), and ⬎52 services attended/yr. The large sample size permitted formation of a category for never attenders (a group unlikely to misrepresent attendance), as well as for persons reporting attending more than weekly, likely highly religious individuals. During the home interview, technicians obtained hypertension history and three blood pressure readings. During examinations carried out in a mobile examination center, physicians obtained an additional three readings (20 –23). The average of all available blood pressure readings was used. Hypertension was defined as SBP ⬎140 or diastolic blood pressure ⬎90 mm Hg or current use of blood pressure medication.
Statistical Analysis From the Centers for Disease Control and Prevention, Hyattsville, Maryland. Address correspondence to R. F. Gillum, National Center for Health Statistics, Centers for Disease Control and Prevention, 3311 Toledo Road, Hyattsville, MD 20782. E-mail
[email protected] Received for publication November 30, 2004; revision received December 20, 2005. DOI: 10.1097/01.psy.0000221253.90559.dd 382 0033-3174/06/6803-0382 Copyright © 2006 by the American Psychosomatic Society
Demographic and other characteristics of the participants were compared among attendance levels. A multivariate logistic regression model was developed to estimate the association of prevalent hypertension with frequency of attendance at religious services controlling for confounding variables. The model had hypertensive status (yes/no) as the dependent/outcome variable and dummy variables for frequency-of-attendance categories (with never attenders as the reference category) as independent variables. To examine the association between Psychosomatic Medicine 68:382–385 (2006)
RELIGIOUS SERVICE ATTENDANCE AND HYPERTENSION SBP and frequency of attendance, a multivariate linear regression model with continuous SBP as the outcome variable and dummy variables for frequency of attendance as independent variables was fit. For both models, age and gender interactions with attendance were examined, and a test for a linear trend was performed by fitting the logistic model with frequency of attendance entered as an ordinal variable. All models included age in single years, gender, non-Hispanic African American race/ethnicity (yes, no), Mexican American ethnicity (yes, no), education level (⬍12 years ⱖ 12 years), marital status (unmarried, married), body mass index (BMI) (kg/m2), cigarette smoking status (current smoker, nonsmoker), respondent-assessed health status (fair/poor, excellent/very good/good), region (South, other), metropolitan residence (yes/no). Region and urbanization were included because of their correlation with both attendance and hypertension prevalence in univariate analyses (not shown). The linear regression model also included current blood pressure medication use (yes, no). The analyses were performed using SAS-callable SUDAAN procedures (CROSSTAB, DESCRIPT, LOGISTIC, and REGRESS) to take the complex survey design and sample weights into account (24 –27).
RESULTS Table 1 presents means or percentages of the demographic and other characteristics of the study participants by frequency of attendance of religious services. As can be seen, 33% reported never attending services, 28% reported attending less than once a week, 29% reported attending weekly, and 10% reported attending more than once a week. Persons who reported attending religious services weekly or more often were older and had higher BMI than persons who attended less frequently, were more likely to be married females, and were less likely to smoke or live in a metropolitan area. Persons who attended services at least weekly were more likely to be hypertensive (thus, they are more likely to have elevated SBP and to currently be taking medication for hypertension) than persons who attend services less frequently. This result is likely due to confounding by age. TABLE 1.
After controlling for sociodemographic and health variables, a significantly lower hypertension prevalence was found both for persons attending services weekly and persons attending more than weekly compared with never attenders (Table 2). Although a doseresponse pattern was seen across attendance categories, only the two highest categories attained significance. A test for linear trend was significant (p ⫽ .0005). There was no significant interaction of frequency of attendance with age (p ⫽ .10) or gender (p ⫽ .23). Blood Pressure Level In linear regression analyses, a significant negative association of frequency of attendance with SBP was seen after controlling for sociodemographic and health variables (Table 3). A test for effect modification by gender was not significant. However, the interaction of age with attendance 1 to 51 times/yr was significant (p ⫽ .03). Taking this into account, compared with never attenders, persons attending weekly had a SBP 1.46 mm Hg (95% confidence limits [CL], ⫺2.33 to 0.58) lower, and persons attending ⬎52 times/yr had an SBP 3.03 mm Hg (95% CL, 1.72– 4.34 mm Hg) lower. A test for linear trend across attendance categories was significant (p ⫽ .0001). DISCUSSION NHANES III data show that although crude rates of hypertension were higher among persons who attend religious services at least weekly compared with others, adjustment for multiple variables revealed an inverse association of attendance frequency and prevalence that was significant. Frequency of attendance also displayed a significant inverse association with SBP, after controlling for sociodemographic and health variables, and hypertension treatment. No
Level of Population Characteristics by Frequency of Attendance at Religious Services Among Persons Aged 20 yr and Over: NHANES IIIa No. Services Attended/yr
Population Characteristic
Age, mean, yr Systolic blood pressure, mean, mm Hg BMI, kg/m2, mean Gender, male Race/ethnicity European American, non-Hispanic African American, non-Hispanic Mexican American Other Unmarried Education ⬍12 yr Region (South) Metropolitan residence Hypertensiveb Current use of meds Current smoker Health status (fair/poor)
0 (n ⫽ 4,796)
1–51 (n ⫽ 4,037)
52 (n ⫽ 4,220)
⬎52 (n ⫽ 1,422)
42 122 26 45
41 120 26 48
48 124 27 40
48** 122** 27** 40
81 8 4 7 37 28 29 50 21 10 39 16
73 15 6 6 34 19 35 52 19 9 31 11
75 10 6 9 32 23 34 47 26 14 18 13
74** 15** 4** 7** 29* 21** 50** 42* 26** 14** 10** 14**
NHANES III ⫽ Third National Health and Nutrition Examination Survey. * p ⬍ .05, ** p ⬍ .01. a Values are percentages unless otherwise indicated. b Hypertensive: at examination SBP ⱖ160 or DBP ⱖ95 mm Hg or currently taking hypertension medication. Psychosomatic Medicine 68:382–385 (2006)
383
R. F. GILLUM AND D. D. INGRAM TABLE 2. Regression Coefficients From a Logistic Model Relating Hypertension and Frequency of Attending Religious Services, Controlling for Sociodemographic and Health Variables Among Persons Aged 20 yr and Over: NHANES III
Services attended/yr 1–51 services 52 services ⬎52 services Age, yr Male African American, non-Hispanic Mexican American Unmarried Education ⬍12 yr BMI, kg/m2 Current smoker Health status (fair/poor) Region (South) Metropolitan Pseudo R2 N

95% CL
⫺0.11 ⫺0.24** ⫺0.33* 0.08** 0.32** 0.63** ⫺0.11 0.08 ⫺0.10 0.10** ⫺0.14 0.37** 0.25* ⫺0.06 0.26 14,475
(⫺0.32, 0.10) (⫺0.37, ⫺0.11) (⫺0.60, ⫺0.07) ( 0.08, 0.09) ( 0.18, 0.46) ( 0.46, 0.80) (⫺0.27, 0.05) (⫺0.06, 0.22) (⫺0.29, 0.08) ( 0.09, 0.11) (⫺0.33, 0.04) ( 0.17, 0.57) ( 0.03, 0.46) (⫺0.23, 0.10)
NHANES III ⫽ Third National Health and Nutrition Examination Survey. * p ⬍ .05, ** p ⬍ .01.
significant effect modification by gender was observed for hypertension or SBP. No significant effect modification by age was observed for hypertension; however, there was a significant interaction between age and attendance 1 to 51 times/yr. These findings are consistent in part with several populationbased studies that have reported negative associations of religious attendance and hypertension or blood pressure (10 –16). Effect TABLE 3. Regression Coefficients From a Linear Regression Model Relating Systolic Blood Pressure (mm Hg) and Frequency of Attending Religious Services, Controlling for Sociodemographic and Health Variables Among Persons Aged 20 and Over: NHANES III
 Services attended/yr 1–51 services 52 services ⬎52 services Age, yr Age ⫻ 1–51 services Male African American Mexican American Unmarried Education ⬍12 yr BMI, kg/m2 Current smoker Health status (fair/poor) Region (South) Metropolitan Current use of BP medication Pseudo R2 N
2.12* ⫺1.46** ⫺3.03** 0.54** ⫺0.07** 6.12** 2.94** 1.75** 2.35** 0.30 0.49** ⫺0.47 ⫺0.52 1.43** ⫺0.90* 9.61** 0.42 14,475
95% CL
( 0.46, (⫺2.33, (⫺4.34, ( 0.52, (⫺0.11, ( 5.44, ( 2.14, ( 1.10, ( 1.74, (⫺0.50, ( 0.43, (⫺1.22, (⫺1.49, ( 0.60, (⫺1.75, ( 8.33,
3.79) ⫺0.58) ⫺1.72) 0.57) ⫺0.03) 6.80) 3.74) 2.40) 2.95) 1.10) 0.55) 0.28) 0.45) 2.26) ⫺0.06) 10.89)
NHANES III ⫽ Third National Health and Nutrition Examination Survey. * p ⬍ .05, ** p ⬍ .01. 384
modification by age of the association of attendance with blood pressure was observed in some (11,15) but not reported on in other studies (12,14). For example, in an elderly sample in North Carolina, the association was stronger at age 65 to 74 than at 75 and above, consistent with the present study. Effect modification by gender was not reported on in one study that included both genders (11). Possible mechanisms for an inverse association of religiousness and blood pressure have been discussed at length elsewhere (10 –19, 28 –37). Religious participation may reduce effects of stress on the individual by aiding social integration, providing social support, and promoting avoidance of unhealthy behaviors. Religious belief systems may aid individuals in coping with stress by improving self-esteem and lessening depression and anxiety. Longitudinal designs are needed to exclude reverse causation (chronic illness and the need to take medication might induce more religious attendance). Strengths and Limitations NHANES III is the largest study to provide population-based data on the association of attendance at religious services and blood pressure or hypertension in nationwide representative samples of Americans. The large sample permits the estimation of effect sizes, of interest per se and in designing smaller studies. Several unavoidable limitations of the present study include possible bias arising from survey nonresponse and from missing values or some variables (38,39), although several special studies of NHANES III data have indicated little bias due to nonresponse (40). One study indicates that duration of hormone replacement therapy may be related to frequency of attendance (41), but adding this variable to the models for women had no effect here. Due to its cross-sectional nature, the study does not provide information as to the temporal sequence of religious attendance and hypertension prevalence or SBP level, making replication in longitudinal studies necessary. At least 12 dimensions of religiousness/spirituality have been defined and instruments developed that measure one or multiple dimensions (42– 46). Attendance at religious services is an indicator of organizational religiousness. As data on multiple religious dimensions were unavailable in NHANES III, the attendance variable was used because it is related to health outcomes, is correlated with other dimensions of public and private religiousness, and provided data that are directly comparable with a body of research data on this variable spanning many decades. This analysis might have yielded different results had other dimensions of religiousness been used. Modest overreporting of religious attendance is likely (43); however, the NHANES III variable should serve well to separate more frequent from less frequent attenders. However, attendance at religious services may be a poor measure of religiousness in the elderly due to physical limitations. Therefore, persons with a history of heart attack, heart failure, or stroke were excluded from the present study. The representativeness of the sample and the use of sample weights provide generalizability of the results to United States noninstitutionalized population of the same ages, but not necessarily to other nations or smaller population subgroups such as American Indians. Further research is needed, including longitudinal studies Psychosomatic Medicine 68:382–385 (2006)
RELIGIOUS SERVICE ATTENDANCE AND HYPERTENSION of religiousness, hypertension incidence, treatment, control and compliance with medical advice for blood pressure screening, follow-up, and medication.
21. 22.
CONCLUSION This analysis of data from a large national sample revealed inverse associations of hypertension prevalence and SBP with frequency of attendance that appeared only after adjusting for sociodemographic and health variables. Elucidation of the mechanism and clinical significance of these associations requires further research. We acknowledge the staff and contractors of the Division of Health and Nutrition Examination Statistics of the National Center for Health Statistics, Centers for Disease Control and Prevention, who conducted the survey and prepared the data for analysis, and Ms Catherine Duran, who assisted with computer programming.
23. 24. 25. 26. 27. 28.
REFERENCES 1. Fields LE, Burt VL, Cutler JA, Hughes J, Roccella EJ, Sorlie P. The burden of adult hypertension in the United States 1999 to 2000: a rising tide. Hypertension 2004;44:398 – 404. 2. National Institutes of Health. Seventh Report of the Joint National Committee for the Prevention, Treatment and Control of Hypertension (JNC7). National Institutes of Health; 2003. NIH Publication No. 03-5233. 3. Hajjar I, Kotchen TA. Trends in prevalence, awareness, treatment and control of hypertension in the United States, 1988 –2000. JAMA 2003; 290:199 –206. 4. Staessen JA, Wang J, Bianchi G, Birkenhager WH. Essential hypertension. Lancet 2003;361:1629 – 41. 5. Vargas CM, Ingram DD, Gillum RF. Incidence of hypertension and educational attainment: the First National Health and Nutrition Examination Survey (NHANES I) Epidemiologic Follow-up Study. Am J Epidemiol 2000;152:272– 8. 6. Steptoe A. Psychosocial factors in the development of hypertension. Ann Med 2000;32:371–5. 7. Kosmin BA, Lachman SP. One Nation under God: Religion in Contemporary American Society. New York, NY: Harmony Books; 1993. 8. Balmer R, Winner LF. Protestantism in America. New York, NY: Columbia University Press; 2002:1–91. 9. Christiano KJ, Swatos WH Jr, Kivisto P. Sociology of Religion: Contemporary Developments. Walnut Creek, CA: Altamira Press; 2002:91–122. 10. Koenig HG, McCullough ME, Larson DB. Handbook of Religion and Health. Oxford: Oxford University Press; 2001:250 – 63. 11. Koenig HG, George LK, Hays JC, Larson DB, Cohen HJ, Blazer DG. The relationship between religious activities and blood pressure in older adults. Int J Psychiatry Med 1998;28:189 –213. 12. Graham TW, Kaplan BH, Cornoni-Huntley JC, James SA, Becker C, Hames CG, Heyden S. Frequency of church attendance and blood pressure elevation. J Behav Med 1978;1:37– 43. 13. Larson DB, Koenig HG, Kaplan BH, Greenberg RS, Logue E, Tyroler HA. The impact of religion on men’s blood pressure. J Religion Health 1989;28:265–78. 14. Livingston IL, Levine DM, Moore RD. Social integration and black intra-racial variation in blood pressure. Ethn Dis 1991;1:135–149. 15. Hixson KA, Gruchow HW, Morgan DW. The relation between religiosity, selected health behaviors, and blood pressure among adult females. Prev Med 1998;27:545–52. 16. Walsh A. Religion and hypertension: testing alternative explanations among immigrants. Behav Med 1998;24:122–30. 17. Steffen PR, Hinderliter AL, Blumenthal JA, Sherwood A. Religious coping, ethnicity, and ambulatory blood pressure. Psychosom Med 2001;63:523–30. 18. Levin JS, Makarides KS. Religion and health in Mexican Americans. J Religion Health 1985;24:60 –9. 19. Strawbridge WJ, Shema SJ, Cohen RD, Kaplan GA. Religious attendance increases survival by improving and maintaining good health behaviors, mental health, and social relationships. Ann Behav Med 2001;23:68 –74. 20. National Center for Health Statistics. Plan and operation of the Third
Psychosomatic Medicine 68:382–385 (2006)
29. 30. 31. 32. 33. 34. 35. 36. 37.
38. 39. 40. 41. 42. 43. 44. 45. 46.
National Health and Nutrition Examination Survey, 1988 –94. Vital Health Stat 1994;27:32. Westat Inc. National Health and Nutrition Examination Survey III: Body Measurements (Anthropometry). National Center for Health Statistics; 1996. Burt VL, Cutler JA, Higgins M. Trends in the prevalence, awareness, treatment, and control of hypertension in the adult United States population: data from the Health Examination Surveys, 1960 –1991. Hypertension 1995;26:60 –9. Burt VL, Whelton P, Rocella EJ. Prevalence of hypertension in the US adult population: results from the Third National Health and Nutrition Examination Survey, 1988 –1991. Hypertension 1995;25:305–13. SAS Institute Inc. SAT/STAT Software: Changes and Enhancements, Release 8.2. Cary, NC: SAS Institute Inc; 2001:1–324. Freund RJ. SAS System for Regression. 3rd ed. Cary, NC: SAS Institute Inc; 2000:1–236. National Center for Health Statistics. Statistical Analytic Guidelines Recommended for the Third National Health and Nutrition Examination Survey, 1988 –94 [MRDF: CD ROM]. National Center for Health Statistics; 1996. Shah BV, Barnwell BG, Bieler GS. SUDAAN Software for the Statistical Analysis of Correlated Data User’s Manual Release 7.0. Research Triangle Park, NC: Research Triangle Institute; 1996. Gillum RF. Trends in cardiovascular diseases: an overview of evolving disparities. In: Livingston I, ed. Praeger Handbook of African American Health. 2nd ed. Westport, CT: Greenwood Publishing; 2004. Kumanyika SK, Charleston JB. Lose weight and win: a church-based weight loss program for blood pressure control among black women. Patient Educ Couns 1992;19:19 –32. Yanek LR, Becker DM, Moy TF, Gittelsohn J, Koffman DM. Project Joy: faith-based cardiovascular health promotion for African American women. Pub Health Reports 2001;115(suppl 1):68 – 81. Levin JS, Vanderpool HY. Is religion therapeutically significant for hypertension? Soc Sci Med 1989;29:69 –78. Seeman TE, Dubin LF, Seeman M. Religiosity/spirituality and health: a critical review of the evidence for biological pathways. Am Psychol 2003; 58:53– 63. Bernardi L, Sleight P, Bandinelli G, Cencetti S, Fattorini L, WdowczycSzulc J, Lagi A. Effect of rosary prayer and yoga mantras on autonomic cardiovascular rhythms: comparative study. BMJ 2001;323:1446 –9. Al-Kandari YY. Religiosity and its relation to blood pressure among selected Kuwaitis. J Biosoc Sci 2003;35:463–72. Krause N, Liang J, Shaw BA, Sugisawa H, Kim H, Sugihara Y. Religion, death of a loved one, and hypertension among older adults in Japan. J Gerontol 2002;57B:S96 –107. Wang PS, Bohn RL, Knight E, Glynn RJ, Mogun H, Avorn J. Noncompliance with antihypertensive medications: the impact of depressive symptoms and psychosocial factors. J Gen Intern Med 2002;17:504 –11. Willis AS, Wallston KA, Johnson KRS. Tobacco and alcohol use among young adults: exploring religious faith, locus of health control, and coping strategies as predictors. In: Plante TG, Sherman AC, eds. Faith and Health: Psychological Perspectives. New York, NY: The Guilford Press; 2001:213–39. Sloan RP, Bagiella E. Claims about religious involvement and health outcomes. Ann Behav Med 2002;24:14 –21. Barkan SE, Greenwood SF. Religious attendance and subjective wellbeing among older Americans: evidence from the General Social Survey. Rev Religious Res 2003;45:116 –29. Mohadjer L, Bell B, Waksberg J. National Health and Nutrition Examination Survey III: Accounting for Item Nonresponse Bias. Rockville, MD: WESTAT, Inc; 1996. Daniels M, Merrill RM, Lyon JL, Stanford JB, White GL Jr. Associations between breast cancer risk factors and religious practices in Utah. Prev Med 2004;38:28 –38. Koenig HG, McCullough ME, Larson DB. Handbook of Religion and Health. Oxford: Oxford University Press; 2001:17–23. Koenig HG, McCullough ME, Larson DB. Handbook of Religion and Health. Oxford: Oxford University Press; 2001:495–510. Peterson J, Atwood JR, Yates B. Key elements for church-based health promotion programs: outcome-based literature review. Pub Health Nurs 2002;19:401–11. Fetzer Institute. Multi-dimensional Measurement of Religion/Spirituality for Use in Health Research. Kalamazoo, MI: John E. Fetzer Institute; 1999. Hill PC, Hood R. Measures of Religiosity. Birmingham, AL: Religious Education Press; 1999.
385
