The Role of Religion in the Well-Being of Older Adults With Schizophrenia

There were no significant differences between the group with schizophrenia and the group in the community by gender, race, religious preferences, or use of spiritualists or healers ( Table 1 ). There was a small difference in age (1.5 years) between the two groups that attained statistical significance. Both groups were in the same median personal income category ($7,000–$12,999).

Older persons with schizophrenia scored significantly lower on religiousness than their age peers in the community, although the absolute difference was only 1.3 points on a 25-point scale ( Table 2 ). Of note, 59% (N=115) of the schizophrenia group and 60% (N=68) of the community comparison group responded "a great deal" to the item asking how much religion or God was "a source of strength" to them (score of 3). Among the seven items in the religiousness scale, the two groups significantly differed only on frequency of attending religious services, with community residents participating more frequently than the schizophrenia group—that is, four times per year versus about once a month ( Table 2 ).

With respect to psychosis and religiousness among persons with schizophrenia, although there was an association between greater religiousness and lower PANSS positive scores, this did not attain significance. Thirty-two percent (N=64) of the sample with schizophrenia reported auditory hallucinations. Among those hearing voices, 27% (N=17) attributed them to God or other supernatural causes, 25% (N=16) attributed them to biomedical or clinical causes, 30% (N=19) attributed them to miscellaneous causes, and 20% (N=13) weren't sure. Nine percent (N=17) of the schizophrenia sample reported specific auditory hallucinations from God (N=13, 7%), the devil (N=2, 1%), or both (N=2, 1%). Among those hearing voices of God, 11 (85%) said they were pleasant, one (8%) said they were unpleasant, and one (8%) said they were both pleasant and unpleasant. In a hierarchical regression analysis, we found a significant inverse association between the PANSS positive scores and the QLI scores ( β =-.21, t=-2.91, df=180, p=.004). When religiousness was added to the analysis, there was essentially no change (that is, a .01 decline in the beta value) in the association between the PANSS positive scores and the QLI scores. When the interactional term (with centroid transformation) of religiousness × PANSS positive scores was added to the analysis, we found no significant association with the QLI scores.

Next, we examined the relationship between stressors, religiousness, and quality of life in the schizophrenia group. As shown in Table 3, in a hierarchical regression analysis, we found that four of the nine primary and intrapsychic stressors in the model—financial strain, acute stressors, depression, and self-esteem—were significantly associated with the QLI scores. (The PANSS positive score was no longer significantly associated with the QLI scores when entered with other variables.) When scores on the religiousness scale were added to the model, we found a significant association between religiousness and the QLI scores. However, the addition of scores on the religiousness scale did not diminish the beta values of the stress variables that had been significantly associated with the QLI scores in step 2. After two other mediating variables (cognitive coping techniques and number of confidantes) were added, the association of the religiousness scale scores with the QLI scores remained significant. Finally, we assessed the moderating effects of religiousness with each of the significant stress variables by separately testing the significance of the interactional term in the model. None of the following interactional terms were significant: Financial Strain Scale × religiousness scale, Acute Stressors Scale × religiousness scale, CESD × religiousness scale, and Self-Esteem Scale × religiousness scale.

In order to examine the potential role of religion to cope with alcoholism, we conducted a post hoc analysis and found no significant correlations between scores on the religiousness scale and current CAGE scores or past CAGE scores.

With respect to whether there would be differences in religiousness scores between older persons with schizophrenia and their age peers in the community, we found the latter had significantly higher levels of religiousness, although the absolute differences were very modest. Moreover, the differences could be attributed largely to formal religious attendance—that is, the community group had roughly three times more frequent attendance. Thus attendance differed between the groups, but older adults with schizophrenia and their age peers were equally likely to report using religion as a coping strategy and to view it as important in their lives.

Our findings were similar to those of a Swiss study with a broad age range of respondents in which persons with schizophrenia had about half the level of group religious attendance as their the counterparts without psychiatric disorders ( 32 ). Huguelet and Mohr ( 16 ) concluded that although a majority of persons with psychotic symptoms use religious coping, many pray alone and are often not involved in religious communities. Huguelet and Mohr ( 33 ) observed that such persons "repeat in religious settings what happens in other areas of their lives," presumably because of difficulties creating and maintaining social networks. Fallot ( 6 ) noted that religious communities can sometimes be rejecting for persons with schizophrenia because of their difficulties meeting expectations in behaviors, dress, or financial support.

With respect to the second question concerning the relationship between psychotic symptoms and religiousness, although persons with schizophrenia who had greater levels of religiousness had lower PANSS positive scores, this association was not significant. We also found that religiousness had no mediating or moderating effects on the relationship between quality of life and the PANSS positive scores among persons with schizophrenia. These findings are somewhat perplexing because previous literature had indicated that a large proportion of persons with schizophrenia use religion to cope with their daily symptoms ( 5 ), although it may be less often used for coping with the psychotic symptoms per se. For example, Huguelet and Mohr ( 16 ) believe that religious coping works primarily by providing a positive sense of self, guidelines for interpersonal behavior, and resources to cope with their symptoms. Moreover, in our sample, levels of psychotic symptoms were fairly low (N=125, or 63%, scored below 4 on all of the PANSS positive items), and consequently, any relationship between religion and psychotic symptoms may have been attenuated.

With respect to the last question—the role religiousness plays in the Stress Process Model—we found that religiousness had no mediating or moderating effects on the four stressors found to be associated significantly with the QLI scores. However, religiousness was associated with a significant, albeit modest, 2.4% increase in the explained variance in the QLI scores. Even after the addition of other mediating variables, such as the number of confidantes and cognitive coping strategies, religiousness retained an independent effect on the QLI scores. Thus the benefit of religiousness in our sample was due to its independent, additive effects on life quality rather than its impact on stressors. Because our data are cross-sectional, this finding must be interpreted cautiously because it is possible that higher quality of life may promote greater religiousness.

Although the evidence for the independent effects of religiousness on the physical well-being of older adults has been overwhelmingly positive, the evidence for its independent effects on psychiatric well-being has been less clear. Strawbridge and coauthors ( 34 ) have pointed to evidence both for and against a buffering effect for religiousness on stressors that affect psychiatric well-being of older persons, which they attributed to the differences in the types of stressors that were involved. Our findings were consistent with an independent effect of religiousness on quality of life, and the absence of any buffering effects may also reflect differences from other studies in types of stressors examined, the measure of religiousness, or the outcome variable.

To our knowledge, this is the first study to examine the role of religion among older adults with schizophrenia. Our study has several strengths, including a large multiracial sample, a community comparison group, a psychometrically reliable measure of religiousness, and the use of a theoretical model to assess the role that religiousness plays in the Stress Process Model of quality of life. However, our study is limited by the use of cross-sectional data that precludes the determination of causality between religiousness and other variables, it is confined to one geographical area, there was a low participation rate in the community sample (52%), and the multiple comparisons increase the possibility of a type I error. Nevertheless, because this is an exploratory study, we wished to avoid type II errors, especially in the interpretation of the regression analysis. Finally, the religiousness measure, although having adequate reliability and content validity, will require further validation.

This study found that persons with schizophrenia have lower levels of religiousness than their age peers, although this was primarily due to differences in the rates of formal religious attendance. It also found that religiousness was not significantly associated with psychotic symptoms, nor did it have any buffering effects on the relationship between psychosis and quality of life. And finally, this study found that religiousness had a significant, albeit modest, independent additive effect on quality of life and that it did not have any buffering effects on the four stressors that were significantly associated with quality of life.

Mental health professionals have been found to be much less religious than their patients ( 1 ), and often they are not aware of their patients' religious involvement ( 32 ). Such blind spots may diminish cultural competence ( 1, 35 ). As our findings suggest, clinicians may overlook a therapeutically important agent.

This study was funded by grants SO6-GM-74923 and SO6-GM-54650 from the National Institute of General Medical Sciences. The authors thank Community Research Applications and Paul M. Ramirez, Ph.D., for their assistance.

The authors report no competing interests.

The authors are affiliated with the Department of Psychiatry, SUNY Downstate Medical Center, Box 1203, 450 Clarkson Ave., Brooklyn, NY 11203 (e-mail: [email protected]).