This Article Abstract Full Text (PDF) Services Download to citation manager Citing Articles Citing Articles via HighWire PubMed PubMed Citation

Age-Related Declines in Activity Level: The Relationship Between Chronic Illness and Religious Activities

¹ Population Research Center, Department of Sociology, The University of Texas at Austin.
Departments of ² Psychiatry and Behavioral Sciences
³ Sociology, Duke University, Durham, North Carolina.

Address correspondence to Maureen Reindl Benjamins, Population Research Center, 1800 Main Building, University of Texas at Austin, Austin, TX 78712. E-mail: reindl{at}prc.utexas.edu

	Abstract

TOP Abstract Theoretical Overview Methods Results Discussion References

 
Objectives. When they are faced with major life transitions such as worsening health, older adults may selectively withdraw from activities. Because of the importance of religion to a large proportion of the elderly population, research is needed to determine whether levels of religious involvement are affected by serious health problems such as the onset of a chronic disease.

Methods. Multiple waves of data from the Duke Established Populations for Epidemiologic Studies of the Elderly were used to analyze the effects of five different chronic conditions on two religious activities: service attendance and religious media use.

Results. Findings show that broken hip, cancer, and stroke were significantly related to levels of religious attendance. Furthermore, the combined conditions also significantly predicted religious attendance, with more conditions being associated with lower attendance. Neither the individual or summed conditions were significantly related to religious media use.

Discussion. The study finds some evidence to support the idea that older adults withdraw from social activities such as religious involvement when faced with declining health. In contrast, levels of religious media use remain stable following the onset of one or more new chronic conditions.

A LONG-STANDING interest in gerontological research is determining whether adults reduce activity levels in their later years, and, if so, why that occurs. Older adults sometimes withdraw from certain aspects of social life as a result of major life transitions, such as retirement or widowhood (Hochschild, 1975). Declining health has been shown to have a similar effect on activity levels. In the extreme case, some individuals withdraw from activities so completely that they become isolated from society (Strauss, 1975). However, it is less clear whether negative changes in health status have an effect on religious activity levels. Examining whether poor health inhibits religious activity among older adults is warranted for two reasons.

First, there is substantial evidence indicating the importance of religion to older adults. For example, some research suggests that many older adults use religion to find support for problems in later life (Koenig, 1994; Taylor & Chatters, 1986). Consequently, were older adults to stop attending church as a result of the onset of disease, such a change might compound the effects of the illness itself. Second, this research is also needed to help inform the larger religion and health literature that has recently emerged. As the research literature on religion and health has grown over the past two decades, some researchers (e.g., Sloan, Bagiella, & Powell, 1999) have questioned the general finding that higher levels of religious activity are associated with better health outcomes. A major critique posed by these researchers is that much of the research in this area is cross-sectional and thus cannot determine whether religion affects health or the reverse. More recent studies have tried to overcome this problem by using longitudinal data (Hummer, Rogers, Nam, & Ellison, 1999; Strawbridge, Cohen, Shema, & Kaplan, 1997). Findings continue to support the idea that religious activity is beneficial for health and mortality. Nonetheless, without adequate studies of the role health plays in religious activity, we cannot be certain of the directionality of the religion–health relationship.

Recent examinations of how health affects religious activity do exist. For example, Idler and Kasl (1997) showed that functionally impaired older adults attended services less frequently than those with no impairment. Similarly, Hays and her colleagues (1998) found that higher levels of impairment were associated with lower levels of attendance, but they found no association between health and religious media usage. Mixed results were also seen in a study by Ferraro and Kelley-Moore (2001), which found that religious consolation was not associated with any of the physical health measures but that both a summary measure of chronic conditions and cancer were negatively associated with religious service attendance.

Although some of these studies illustrate the process of declining religious activity, findings from several studies using other types of religion outcomes challenge this notion. In fact, it has been suggested that religion is more beneficial in some situations than others, specifically in times of higher stress (Pargament, 1997). Because the onset of a chronic disease in a person would presumably lead to high levels of stress, perhaps it also leads to an increase in the importance of religion to that individual. One study that supports this idea showed that individuals with health-related problems are more likely to use religious coping than those with other types of problems (Ellison & Taylor, 1996). Another study involving religious coping found that the presence of a serious health condition was associated with increased seeking of religious consolation (Ferraro & Kelley-Moore, 2000). Similarly, Musick and his colleagues (Musick, Koenig, Hays, & Cohen, 1998) found that individuals with cancer reported higher levels of religious activity compared with individuals who had another or no illness. This study focused on individuals who already had cancer (for an unspecified period of time); thus, it did not examine changes that may occur with the onset of a chronic illness, which we believe warrants special consideration. The examination of new conditions is needed because it allows us to better assess changes in activity that result from health problems.

The relationship between chronic illness and religious activities may depend on the type of illness as well as the cumulative burden of multiple conditions. It is important to look at each condition independently because it is likely that the physical and psychological repercussions of each condition differ and, consequently, the effect of that condition on religious activities may differ as well. In two studies that have used this approach, it has been shown that the presence of a specific illness, cancer, is associated with changes in religious activities (Ferraro & Kelley-Moore, 2001; Musick et al., 1998). However, it is also possible that individuals faced with more than one chronic condition have a cumulative burden of illness that may affect their religious participation above and beyond the effects of the individual conditions. One previous study supports this idea. The study found that a summary measure of chronic conditions is associated with religious attendance; however, it examined the presence of chronic conditions and not the onset of new conditions (Ferraro & Kelley-Moore, 2001).

The purpose of this article is to overcome this gap in our knowledge by considering the effects of a newly diagnosed chronic illness on levels of religious activity in a community sample of older adults. In pursuing this goal, we do the following. First, we discuss the ways that chronic illnesses (individually and combined) should affect religious activity and the mechanisms that underlie these relationships. Next, we test several hypotheses regarding these relationships by using data from a sample of older adults residing in the Piedmont area of North Carolina. Finally, we conclude with a discussion of the findings and implications for research on religion and health.

	THEORETICAL OVERVIEW

TOP Abstract Theoretical Overview Methods Results Discussion References

 
The Social Cost of Illness
The onset or diagnosis of a chronic illness is a major life transition that is usually accompanied by significant changes in status and roles (Bohachick & Anton, 1990). According to Parson's notion of the "sick role," those who are ill are allowed to temporarily forsake normal role obligations as a result of the illness (Parsons 1951, 1975; see also Gerhardt, 1989). When the illness is a chronic one, these changes in role obligations take the form of more permanently modified expectations. Compounded with the physical and emotional consequences of an illness, this acceptance of altered roles may make individuals more likely to withdraw. If this theory is correct, the onset of a chronic disease could significantly alter patterns of activity that existed before the illness. Previous research on chronic illness and social activity supports the idea that significant changes in activity do occur after the onset of a major illness (Bury, 1991). Although there are numerous possible mechanisms responsible for changes in activity level, two are prominent.

Physical Problems
Not surprisingly, the disease itself or its treatments may result in a multitude of physical repercussions such as reduced mobility, pain, and fatigue, as well as other symptoms such as nausea or shortness of breath. For example, among cancer patients, approximately 70% experience severe pain. Further, research has indicated that the treatment of pain among persons with cancer is inadequate, leaving many patients with the problem and the responsibility of alleviating it (Breitbart & Payne, 1998). Certain treatment regimens and medications can also cause the physical symptoms already noted. In sum, the physical toll that a chronic illness can take on individuals is often severe and can result in a limitation of social activities. It is therefore possible that any relationship we see between chronic conditions and religious activities may be mediated by physical limitations.

Psychological and Social Repercussions
Chronically ill adults appear to take on new negative evaluations of the self, which in turn affect social activities (Bury, 1991). In this regard, Charmaz (1983, p. 169) argued, "as they suffer losses of self from the consequence of chronic illness and experience diminished control over their lives and their futures, affected individuals commonly not only lose self-esteem, but even self-identity." She further argued that this loss of identity can result in the breakdown of social relationships and the lessening of social interaction. For some chronically ill adults, this withdrawal extends to social isolation from all but a few close family and friends. Bury (1982) also argued that self-confidence often wanes among chronically ill adults; consequently, they are less able, and therefore less willing, to manage and engage in social activities. Because a new chronic condition can have psychological repercussions, which may result in limited involvement in social activities, mental health status may also mediate the relationship between chronic conditions and religious activities.

Illness and Religious Activity
Given the consequences that ensue after the onset of disease, it is possible that older adults will reduce levels of social activity, including religious service attendance. However, the same cannot be said for private forms of religious activity. Indeed, in response to the disease, older adults might actually increase these types of activity, because older adults who value religious activity yet must limit their associational involvement may compensate by substituting more private forms of activity. For older adults who are physically impaired or face social repercussions because of their illness, private religious activity serves the dual function of providing religious involvement and giving the older adult a meaningful activity when few others might be available.

Evidence on whether older adults with health problems will participate in higher levels of private religious activity is mixed. For example, Mindel and Vaughan (1978) argued that although older adults might withdraw from the more organized forms of religious activity, in part because of health problems, they would not do so for private forms of activity. They found no health effects on public religious activity; however, adults in worse health reported more private activity than adults in a healthier group. Similarly, Ferraro and Kelley-Moore (2000) found that religious consolation seeking increased for those with chronic conditions. Recall, however, that Hays and her colleagues (1998) examined this idea by using data on functional health and found no effects on religious media use.

Previous Conditions
For many, problems resulting from chronic illness are difficult to surmount, and therefore they will remain inhibitors of activity. In contrast, some can be overcome with proper preparation and socialization. Given these distinctions, there may be a substantial difference between those for whom a new chronic condition is a first versus those who have one or more previously existing conditions. Though both groups are likely to suffer the physical consequences of the new disease, the latter group, having already been socialized to the disease process by their extant conditions, may not suffer in a social sense as heavily as those who are experiencing disease for the first time and are thus unprepared.

Hypotheses and Analytical Design
Given the findings of previous research and our own distillation of the literature, we propose to test five hypotheses. The first two are our expectations regarding illness and religious activity. Following these, we test two hypotheses regarding the mechanisms underlying the association between illness and religious activity. Finally, we test a hypothesis in consideration of the effects of illness timing.

Hypothesis 1: Respondents reporting new chronic illness engage in less public religious activity than those reporting no new chronic illness.

Hypothesis 2: Respondents reporting new chronic illness engage in more private religious activity than those reporting no new chronic illness.

Hypothesis 3: The association between new chronic illness and religious activity (public and private) is mediated by functional limitations.

Hypothesis 4: The association between new chronic illness and religious activity (public and private) is mediated by mental health problems.

Hypothesis 5: The effects of new chronic illness on religious attendance and media use are strongest for respondents who report no previous illnesses.

Analytical Design
The hypotheses are tested by using two religion outcomes: attendance at religious services and religious media use. These two outcomes are regressed on (a) their respective baseline measures, (b) chronic conditions newly diagnosed since baseline (individually and combined), (c) chronic conditions existing at the baseline, and (d) a set of control factors. The controls include standard sociodemographic variables such as age, gender, and race. Because previous studies have linked social interaction and support to religion (e.g., Ellison & George, 1994), we also include those factors in our models. In addition, we test the mediation hypotheses by including measures of functional limitations and depression in our models. Finally, we test for interactions between the chronic conditions and baseline conditions to determine if the effect of a new condition differs by the presence of a previous one.

	METHODS

TOP Abstract Theoretical Overview Methods Results Discussion References

 
Data
Data from the Established Populations for Epidemiologic Studies of the Elderly (EPESE) project at the Duke University Medical Center were used for this study. This project was part of a multicenter, collaborative epidemiologic investigation of physical, social, and cognitive functioning of persons 65 years of age and older living in one of four sites—East Boston, Massachusetts; Iowa and Washington counties, Iowa; New Haven, Connecticut; and five counties in the north central North Carolina Piedmont area. The North Carolina study is a stratified, random household sampling of noninstitutionalized elderly respondents in five contiguous counties (one predominantly urban and four predominantly rural). In the sample design, the clusters were smaller units within these counties. The sample includes approximately equal numbers of urban and rural residents and an oversample of African Americans (Cornoni-Huntley et al., 1990). Baseline interviews were conducted in person during 1986 with follow-up interviews 3 years later (1989). Telephone interviews took place during each of the intervening years (1987 and 1988). The data used for this study come from the two in-person interviews in 1986 and 1989 and include information about incident conditions from both of the telephone interviews.

The actual sample size is 2,958 participants. The sample size was reduced from the baseline sample size of 4,162 as a result of the elimination of proxies, those respondents not participating in all waves, and those with missing data in the dependent variables. With the exception of religious activity, new chronic conditions, and mediators, all variables were measured at baseline. For the cases with missing data, we imputed the mean value. The variables with imputed values are as follows (with number missing): education (n = 17), social interaction (n = 3), and perceived social support (n = 19). The number of missing for functional impairment (n = 124; 6% of the sample) was significant enough to warrant special attention. Several methods of handling these missing cases were tried, including imputation and exclusion. Analyses (not shown here) showed that imputation was the better approach; however, rather than assuming that all missing cases were at the mean for impairment, we coded them as having the highest level of impairment. Note, however, that the effects found with other forms of imputation were very similar to those found with this method. We did not impute values for those with missing data on religious activities (at baseline or at Time 2).

The data were weighted to adjust for oversampling, missing values, and attrition. Weights came from 1989, the wave at which the outcome variables were measured. Because of the complex sampling designs used to collect the EPESE data, the variances of the estimates in the regressions models may be understated if one assumes a simple random sample, as is done in most statistical software. Thus, we adjusted for the sample design effect by means of Taylor series linearization procedures in STATA 7.0. The estimates shown come from these models.

Measures
Religious activity
Two measures of religious activity were used: service attendance and media use. Service attendance was measured by the respondent's self-reported frequency of going to religious services or meetings on a scale ranging from never (1) to more than once a week (6). Media use was measured with the following question: "About how often do you watch religious services or religious programs on TV or listen to them on the radio?" The same 6-point scale was used. Box–Cox transformations were tested on both outcome variables to choose a transformation that would correct for skewed distributions of error terms and unequal error variances within the regression models (Neter, Wasserman, & Kutner, 1989). With the use of the baseline models, the square root transformation minimized the error sum of squares for both variables, and consequently it is used in all subsequent models. The baseline religion variables are also transformed in this manner.

Chronic conditions
Chronic conditions were measured in two ways. First, each illness was assessed independently with the following question: "Since we spoke to you the last time in (month and year), has a doctor told you that you have (illness name)?" The following chronic conditions were included: broken hip, cancer, diabetes, heart attack, and stroke. Information about incident conditions was collected during the two telephone interviews (1987 and 1988) and the second in-person interview (1989). The incident condition variable was then dichotomized to represent respondents who reported a new condition in any of those three interviews versus those who reported no new conditions. Second, a summed conditions variable was created by adding all five of the individual chronic conditions.

Baseline conditions were calculated as the number of chronic conditions reported by the respondents at baseline. This variable was then dichotomized to distinguish between those with no baseline conditions and those with one or more.

Social resources
Social resources were measured by using two commonly used scales from the Duke Social Support Inventory (Landerman, 1994). The first, social interaction, measured the degree to which the respondent interacts with others, using a five-item index. The first three items involved how often the respondent saw the following people per month: (a) their children, (b) other relatives, and (c) close friends. The fourth item estimated how many times the respondent spoke on the telephone in the last month. For the final item, respondents were asked whether they belonged to a club or other association. The index was constructed by summing the number of contacts mentioned in these five items. The second measure of social support used was perceived support. This was measured with a two-item index that indicated how often respondents can (a) count on family and friends in times of trouble and (b) talk about their deepest problems with family and friends. For both indices, higher scores indicate more social resources.

Functional impairment
This item was measured by using the Rosow–Breslau scale of activity limitation (Rosow & Breslau, 1966). This includes three items that measure impairment based on whether the respondent can (a) do heavy work, (b) walk up a flight up stairs, and (c) walk a half mile. Higher values indicate greater functional impairment.

Depression
The standard 20-item Center for Epidemiological Studies–Depression (CES-D) scale was used to measure depression (Radloff, 1977). The items were then summed to create a depression score. Higher scores indicate a greater number of depressive symptoms. As already noted, both functional impairments and depression were measured at Time 2 (in 1989).

Control variables
Controls include gender (female: 0 if male; 1 if female), race (Black: 0 if non-Black; 1 if Black), age (in years), marital status (married: 0 if not married; 1 if married), and education (years of formal schooling).

	RESULTS

TOP Abstract Theoretical Overview Methods Results Discussion References

 
Sample characteristics are provided in Table 1. Small percentages of the sample reported any newly diagnosed chronic illness. The most common new condition was heart disease, with approximately 5% of the population reporting a new incidence. No individual had more than three new conditions. Although only 16% reported a new condition between 1986 and 1989, 40% of the sample had one or more condition at baseline (1989). The religion variables (before transformation) indicate that the respondents attended religious services between one and two times a month and watched religious programs once a week, on average.

In Table 2, we regress the transformed religious activity variables, measured in the second in-person wave, on the five new chronic conditions. There are three models for each outcome. Model 1 includes baseline religious activity and chronic conditions as well as other baseline controls. In addition to these variables, Model 2 examines the possible effect of mediating factors by including functional impairment and depression measures. Model 3 includes all these variables as well as the interaction terms between each new chronic condition and the presence of a baseline condition.

Other variables that strongly affect service attendance are baseline attendance, race, age, education, and social interaction. Model 2 displays the effect of the two mediating variables, functional impairment and depression, on service attendance. Only one of these variables has a significant influence on service attendance, with higher levels of functional limitations being associated with lower levels of attendance. When these mediators are included, none of the chronic conditions remains a significant predictor of attendance. Consequently, the first hypothesis on meditating effects is supported by the data.

Model 3 displays the effects of the baseline interaction terms. Only the interaction between cancer and the presence of a baseline condition is significant. The findings indicate that, among respondents without a baseline condition, there is a negative effect of cancer on attendance (b = -.15). In contrast, for those with one or more baseline conditions, the effect of being diagnosed with cancer on attendance is positive (b =.15). In other words, for individuals with a previously existing condition, cancer actually leads to more frequent religious attendance.

The relationship between newly diagnosed conditions and religious media use is shown in the second part of Table 2. Unlike service attendance, no new chronic conditions are associated with levels of religious media use in any of the models. These findings regarding newly diagnosed chronic conditions and media use do not support our expectations: reporting a new condition was not associated with increased levels of media use in the follow-up wave for any of the five conditions included. As with the previous models, the lack of statistical power of the models precludes us from concluding that no relationships between each of the incident conditions and religious media use exists.

The primary predictors of religious media use in these models are baseline religious media use, race, and education. More specifically, reporting previous use of religious media, being Black, and having less education are all factors related to higher levels of media use. In addition, having a greater number of functional limitations is associated with higher utilization of religious media. However, the coefficients of the individual chronic conditions did not change significantly between Models 1 and 2. Therefore, our expectations of mediating effects were not supported by the data. Furthermore, none of the interaction terms between presence of a baseline condition and a new condition were significantly associated with media use. This also contradicts expectations that the effects of a new chronic condition on media use would differ by the presence of a baseline condition.

We have found that, on one hand, individuals face declining levels of religious attendance when they are confronted with certain chronic conditions, primarily broken hip, cancer, and stroke. On the other hand, individuals maintain stable levels of religious media use despite new chronic conditions. It is possible that other forms of social activity could be affected differently. In other words, there may be something special about religious activities that makes them relatively resilient to change. Or, conversely, chronic conditions may have a similar effect on other types of social activities as well. As a cursory test of this question, the analyses were repeated with a nonreligious form of social activity as the outcome variable. Because of data limitations, the best choice for this outcome was the social interaction variable that was previously included as a predictor. When regressed on its baseline equivalent and the other variables in the full model, social interaction had a significant association with only one of the chronic conditions. More specifically, respondents with a broken hip engaged in less social interaction over time (not shown; p <.01).

Because several individual chronic conditions are associated with a change in level of religious activities, it is reasonable to expect that being confronted with more than one new condition may also bring about changes in activity levels. In Table 3, we test this idea of cumulative burden by regressing service attendance and religious media use (both transformed) on the total number of incident conditions and the mediators and baseline controls.

In contrast to service attendance, the total number of conditions was not related to religious media use. This is consistent with the findings shown in Table 2. Likewise, transformed baseline media use, race, and education are still significantly related to media use. The interaction between total new conditions and a baseline condition was found to be a poor predictor of religious media use.

	DISCUSSION

TOP Abstract Theoretical Overview Methods Results Discussion References

 
Direct and Indirect Effects of Chronic Conditions on Religious Activities
We proposed that older adults may withdraw from certain social activities as a result of declining health. Our findings suggest that religious attendance decreases when individuals report an incidence of a broken hip, cancer, or stroke. Furthermore, when the number of new chronic conditions is summed, decreasing levels of health also significantly affect religious attendance. Thus, the first hypothesis receives support from the data. The second hypothesis, that poor health leads to higher levels of religious media use, received no support. We found that levels of religious media use remain stable despite new incidences of broken hip, cancer, diabetes, heart attack, and stroke, or all conditions combined.

The inclusion of potential mediating variables shows that functional impairments, but not depression, lead to significantly lower levels of attendance. Furthermore, the inclusion of the mediators alters the significance levels of the attendance effects. Therefore, the third hypothesis receives limited support from the data, whereas the fourth hypothesis receives no support. Finally, the effects of a new condition on religious activity were not significantly stronger for respondents who reported no previous conditions at baseline (with the exception of cancer). Thus, only two of our original hypotheses received support from the data.

It must be noted that the number of cases of incident conditions is small (i.e., 1–5% of the sample reported having an individual incident condition and only 16% of the sample reported having any of the incident conditions). As previously mentioned, it is possible that the poor distribution of this dependent variable might account for the lack of findings regarding the individual conditions for both outcomes or the total lack of findings for religious media use. In this case, the lack of significant findings may reflect insufficient statistical power rather than no relationship between incident health conditions and religious activities (Ferraro & Wilmoth, 2000).

Numerous past studies have supported this idea that infrequent religious attendance is associated with lower than average physical health (Hummer et al., 1999; Koenig et al., 1997; Oxman, Freeman, & Manheimer, 1995; Pressman, Lyons, Larson, & Strain, 1990; Strawbridge et al., 1997); however, none have examined the predictive power of specific chronic illnesses on attendance while controlling for the mediating effects of physical disability or mental health. The broad range of chronic conditions studied, as well as the inclusion of potential mediators, represents a unique contribution to the religion and health literature.

Furthermore, the use of two distinct types of religious involvement allowed us to test the effects of chronic conditions on both public and private types of activities. We hypothesized that a new condition would limit individuals' ability to attend church but would increase their use of religious media. Although a positive relationship was expected, the lack of findings regarding religious media use is not surprising in light of previous studies that also include measures of both organized and nonorganized religious activities. Hays and associates (1998) examined whether religious media use increased with declining health to compensate for decreased attendance. They found that attendance did decrease, but religious programming use did not change. Although the current study focuses on chronic conditions instead of functional ability, the stability of religious programming use through times of physical difficulty remains the same.

It is possible that health problems might lead to increases in other forms of private religious activity, such as prayer. Our data included a measure of private devotional activities along these lines. Unfortunately, the wording of the relevant question was problematic because it combined two forms of activity that can be differentiated in terms of the level of functioning needed to perform them. Although we opted not to use this variable, preliminary analyses showed no significant associations between chronic conditions and these private devotional activities.

Possible Moderating Effects
We predicted that a new chronic condition would result in larger changes in religious activities for those without baseline chronic conditions. When interactions were tested with baseline conditions, no strong pattern emerged to support our hypothesis. In fact, only two of the interaction terms were significant—the interaction between a new incident of cancer and a baseline condition and the interaction between the summed conditions and a baseline condition, both in the service attendance model. Apparently for those with a baseline condition, being diagnosed with cancer or with a combination of conditions does not create a cumulative burden. In fact, there is something about the cumulative nature of the conditions that improves religious attendance. We are unsure of the implications of this finding and believe that more research is needed in this area.

In contrast, possible explanations for the general lack of effect of previous conditions are easier to envision. For example, it is possible that previous conditions may have had an effect on the religious activity levels of the individual but this effect was short term. Therefore, the arrival of each new chronic illness initiates a new set of changes and all individuals are affected similarly, regardless of previous conditions.

Study Limitations and Implications
There are several limitations to this study. First, the vast majority of the sample reported belonging to a Protestant denomination. Therefore, these trends are not representative of populations with a more diverse distribution of religious affiliations. Second, because the study is located in a southern "Bible Belt" state, the results are only representative of this region of the country. In addition, the illnesses themselves, along with the majority of the other variables, were all self-reported. More accuracy could be obtained with a physician's diagnosis.

Despite these limitations, several possibilities are suggested by this study. The relationship between chronic conditions (both individual and cumulative) and religious attendance may indicate that more extensive social outreach programs for the sick are needed from religious groups. Although older adults may become less involved in activities in several spheres of their lives, it is especially critical to investigate both the causes and consequences of a withdrawal from religious activities because such a withdrawal may intensify the far-reaching effects of chronic illnesses in their lives. If older adults become less involved in public religious activities, they may lose the emotional and instrumental support that these activities provide. This trend of reduced activity could potentially be counteracted by hospital- or community-based programs that offer transportation to religious services, information about television or radio programs, or invitations to Bible study or prayer groups.

Future Research
More research is needed concerning possible mechanisms underlying the relationship between chronic conditions and social activities. One possible mechanism that has not received much attention is time constraints. Individuals with chronic conditions must adhere to an often time-consuming treatment regimen, such as self-administered medication, regular visits to a hospital or clinic, or other types of therapy (Bury, 1991). Whatever the treatment regimen, the time costs associated with trying to overcome the disease are significant. Consequently, activities not immediately associated with treatment are likely to be curtailed. Second, adults with a new chronic health condition face a number of other difficulties for which they are not normally prepared. The process of dealing with this new condition will necessarily involve the shifting of social activities and network ties, with more emphasis placed on these new disease-related contacts. The role of these factors in the relationship between chronic conditions and religious and nonreligious involvement requires further investigation.

The results found here also suggest other areas in which future research is needed. Studies are needed to examine the differing effects of chronic illnesses on individuals of all ages. It is likely that an adolescent or young adult faced with a chronic disease would react differently than an elderly person, perhaps by continuing to attend church with the help and insistence of a parent, or by being more likely to turn away from nonorganizational activities that had yet to become ingrained as a part of daily life. Additionally, researchers should compare and contrast the ways that chronically ill individuals of different denominations modify their religious activities after diagnosis. It is possible that certain denominations are better equipped or more motivated to deal with chronically ill members. In turn, this may affect how these individuals modify their involvement following the onset of a chronic condition.

In summary, our results indicate that having a broken hip, cancer, or a stroke negatively affects religious attendance. Furthermore, the cumulative effect of new chronic illnesses in older adults also results in decreased attendance at religious services. However, the burden of an individual chronic condition or cumulative conditions does not have any effect on religious media use. These findings indicate that older adults may withdraw from certain religious activities when faced with declining health, while their involvement in other types of religious activities remains stable. They also point to specific areas in which more research is needed.

	Acknowledgments

 
The research upon which this publication was based was performed pursuant to Contract NOI -AG-1-2102 with the National Institute on Aging, in support of the EPESE (Duke). The content of this publication does not necessarily reflect the views or policies of the U.S. Department of Health and Human Services. Partial support for this study was also provided by the National Institute of Health under Grants HD30442 (Dr. Gold) and AG 11268 (Drs. Gold and George), and by a grant from the American Association of Retired Persons/Andrus Foundation (Dr. Gold).

	Footnotes

 
Decision Editor: Charles F. Longino, Jr., PhD

Received for publication September 17, 2002. Accepted for publication June 2, 2003.

	References

TOP Abstract Theoretical Overview Methods Results Discussion References

 

• Bohachick, P., Anton, B. B. 1990. Psychosocial adjustment of patients and spouses to severe cardiomyopathy. Research in Nursing and Health, 13,385-392.
• Breitbart, W., Payne, D. K. 1998. Pain. In J. Holland (Ed.), Psycho-Oncology (pp. 450–465). New York: Oxford University Press.
• Bury, M. R. 1982. Chronic illness as biographical disruption. Sociology of Health and Illness, 4,167-182.
• Bury, M. R. 1991. The sociology of chronic illness: A review of research and prospects. Sociology of Health and Illness, 13,451-468.
• Charmaz, K. 1983. Loss of self: A fundamental form of suffering in the chronically ill. Sociology of Health and Illness, 5,168-195.
• Cornoni-Huntley, J., Blazer, D. G., Lafferty, M. E., Everett, D. F., Brock, D. B., Farmer, M. E. 1990. Established Populations for Epidemiologic Studies of the Elderly (Volume II): Resource data book (NIA Publication No. 90-495). Washington, DC: National Institute on Aging.
• Ellison, C. G., George, L. K. 1994. Religious involvement, social ties, and social support in a Southeastern community. Journal for the Scientific Study of Religion, 33,46-61.
• Ellison, C. G., Taylor, R. J. 1996. Turning to prayer: Social and situational antecedents of religious coping among African Americans. Review of Religious Research, 38,111-131.
• Ferraro, K. F., Kelley-Moore, J. A. 2000. Religious consolation among men and women: Do health problems spur seeking? Journal for the Scientific Study of Religion, 139,220-235.
• Ferraro, K. F., Kelley-Moore, J. A. 2001. Religious seeking among affiliates and non-affiliates: Do mental and physical health problems spur religious coping? Review of Religious Research, 42,229-251.
• Ferraro, K. F., Wilmoth, J. M. 2000. Measuring morbidity: Disease counts, binary variables, and statistical power. Journal of Gerontology: Social Sciences, 55B,173-189.
• Gerhardt, U. 1989. Ideas about illness: An intellectual and political history of medical sociology. Washington Square, NY: New York University Press.
• Hays, J. C., Landerman, L. R., Blazer, D. G., Koenig, H. G., Carroll, J. W., Musick, M. A. 1998. Aging, health, and the "electronic church.". Journal of Aging and Health, 10,458-482.[Abstract/Free Full Text]
• Hochschild, A. R. 1975. Disengagement theory: A critique and proposal. American Sociological Review, 40,553-569.
• Hummer, R. A., Rodgers, R. G., Nam, C. B, Ellison, C. G. 1999. Religious involvement and U.S. adult mortality. Demography, 36,273-285.[Medline]
• Idler, E. L., Kasl, S. V. 1997. Religion among disabled and nondisabled persons II: Attendance at religious services as a predictor of the course of disability. Journal of Gerontology: Social Sciences, 52B,S306-S315.[Abstract]
• Koenig, H. G. 1994. Aging and God: Spiritual pathways to mental health in midlife and later years. Binghamton, NY: Haworth Pastoral Press.
• Koenig, H. G., Cohen, H. J., George, L. K., Hays, J. C., Larson, D. B., Blazer, D. G. 1997. Attendance at religious services, interleukin-6, and other biological parameters of immune function in older adults. International Journal of Psychiatry and Medicine, 27,233-250.
• Landerman, R. 1994. EPESE composite measures and other commonly used measures: Description and documentation of original variables and variables with missing values imputed. Durham, NC: University Medical Center.
• Mindel, C. H., Vaughan, C. E. 1978. A multidimensional approach to religiosity and disengagement. Journal of Gerontology, 33,103-108.
• Musick, M. A., Keonig, H. G., Hays, J. C., Cohen, H. J. 1998. Religious activity and depression among community-dwelling elderly persons with cancer: The moderating effect of race. Journal of Gerontology: Social Sciences, 53B,S218-S227.
• Neter, J., Wasserman, W., Kutner, M. H. 1989. Applied linear regression models (2nd ed.). Homewood, IL: Irwin.
• Oxman, T. E., Freeman, D. H., Jr., Manheimer, E. D. 1995. Lack of social participation or religious strength and comfort as a risk factor for death after cardiac surgery in the elderly. Psychosomatic Medicine, 57,5-15.[Abstract/Free Full Text]
• Pargament, K. I. 1997. The psychology of religion and coping. New York: Guilford Press.
• Parsons, T. 1951. The social system. New York: Free Press of Glencoe.
• Parsons, T. 1975. The sick role and the role of the physician reconsidered. Milbank Memorial Fund, Health and Society, 53,257-278.
• Pressman, M. A., Lyons, J. S., Larson, D. B, Strain, J. J. 1990. Religious belief, depression, and ambulation status in elderly women with broken hips. American Journal of Psychiatry, 147,758-760.[Abstract/Free Full Text]
• Radloff, L. S. 1977. The CES-D scale: A self-report depression scale for research in the general population. Applied Psychological Measurement, 1,385-401.
• Rosow, I., Breslau, N. 1966. A Guttman health scale for the aged. The Journal of Gerontology, 21,556-559.
• Sloan, R. P., Bagiella, E., Powell, T. 1999. Religion, spirituality, and medicine. The Lancet, 353,664-667.[Medline]
• Strauss, A. L. 1975. Chronic illness and the quality of life. St. Louis, MO: Mosby.
• Strawbridge, W. J., Cohen, R. D., Shema, S. J., Kaplan, G. A. 1997. Frequent attendance at religious services and mortality over 28 years. American Journal of Public Health, 87,957-961.[Abstract/Free Full Text]
• Taylor, R. J., Chatters, L. M. 1986. Church-based informal support among elderly Blacks. The Gerontologist, 26,637-642.[Medline]

This article has been cited by other articles:

				T. A. Balboni, L. C. Vanderwerker, S. D. Block, M. E. Paulk, C. S. Lathan, J. R. Peteet, and H. G. Prigerson Religiousness and Spiritual Support Among Advanced Cancer Patients and Associations With End-of-Life Treatment Preferences and Quality of Life J. Clin. Oncol., February 10, 2007; 25(5): 555 - 560. [Abstract] [Full Text] [PDF]

				N. Agahi, K. Ahacic, and M. G. Parker Continuity of leisure participation from middle age to old age. J. Gerontol. B. Psychol. Sci. Soc. Sci., November 1, 2006; 61(6): S340 - S346. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Services Download to citation manager Citing Articles Citing Articles via HighWire PubMed PubMed Citation