Abstract
Informal caregivers suffer burden due to the responsibility of caring for an ailing family member. These caregivers are labeled as informal because they are not paid for their work and they are not trained to deal with the mental or physical health issues of the person in their care. As a result, they may suffer from mental and physical health problems of their own (Keith, Wacker, & Collins, 2009). This burden presents in the form of financial problems and poor physical health, as well as high levels of perceived stress (Vella & Pal, 2013). Studies have shown, however, that some aspects of religion may reduce the burden of caregiving. In a review of 83 empirical studies, Hebert, Weinstein, Martire, and Schulz (2006) found varied effects of religion on caregiver burden ranging from no effect to stress reduction. Of those 83 studies, only 13 utilized longitudinal data in their analyses. All other studies were cross-sectional, and only 28 used multidimensional measures of religiosity. One of their major recommendations was to use a valid instrument that represented the multidimensionality of spirituality and religion to improve future research. The authors also pointed out the importance of investigating the role of ethnicity/race in caregiving outcomes in subsequent studies.
The mental health of caregivers often reflects the burden of caring for a family member suffering from an illness that is often debilitating and requires countless hours of care, often at inconvenient times. Using a sample of almost 7,000 married couples in Norway 55 years and older, Ask et al. (2014) studied the mental health of a spouse caring for a partner with dementia. Caregivers had lower levels of life satisfaction and more symptoms of depression and anxiety than their elderly counterparts who were caring for a spouse without dementia. In a sample of 99 caregivers in Australia, mental and physical health were both compromised (Butow et al., 2014). This matches findings in other studies. According to O’Brien (2000), caregivers are at risk of suffering from chronic conditions ranging from depression to insomnia. In England, Smith et al. (2014) found that while mental health was affected negatively by the stress of being a caregiver, physical health in caregivers was no worse than noncaregivers. Cantwell, Muldoon, and Gallagher (2014) found that caregiver parents of children with developmental disabilities suffered from headaches and gastrointestinal and respiratory problems, as well as sleep disruptions.
This does not mean that a caregiver is destined to suffer from the burden of caregiving. Several researchers have found evidence that a strong religious or spiritual life can reduce burden due to informal caregiving. Weaver and Flannelly (2004) found that the most successful caregivers were those who had support from their religious faith. Murray-Swank et al. (2006), in their cross-sectional study of 83 caregivers, found that caregivers with strong religious or spiritual bonds tended to have higher self-esteem and greater self-care than those who were less religious whereas their less religious counterparts tended to be more depressed. In a cross-sectional analysis of their baseline data using structural equation modeling, Heo (2014) found that caregivers who used positive religious coping perceived an increase in social support, which led to lower levels of depression.
Unfortunately, the negative aspects of religion can be detrimental to the health of caregivers. Those who see God as having abandoned them or those who felt that their being a caregiver was punishment for their lack of spirituality often suffered from weight gain (Rabinowitz, Hartlaub, Saenz, Thompson, & Gallagher-Thompson, 2010). Similarly, Pearce, Singer, and Prigerson (2006) found that negative religious coping was associated with poorer quality of life and more stress among the 162 caregivers.
Rathier, Davies, Papandonatos, Grover, and Tremont (2015) found that while higher levels of religious coping working with God (i.e., collaborative and active surrender) were associated with decreased depression among caregivers, higher levels of religious coping working through God (i.e., passive deferral, pleading for direct intercession, and self-directed religious coping) were associated with increased depression in a sample of 191 caregivers. In their study, Herrera, Lee, Nanyonjo, Laufman, and Torres-Vigil (2009) found that intrinsic religiosity (i.e., internalized religion) and organizational religiosity (i.e., public practice of religious rites) were associated with lower perceived caregiver burden. However, nonorganizational religiosity (i.e., private practice of religious rituals) was associated with poorer mental health, and negative religious coping (i.e., viewing God as punishing and trying to work through the situation without God) predicted greater depression among 66 predominantly Mexican American Catholic caregivers.
In their review, Hebert et al. (2006) note several important issues that need to be considered in research on the relationship of religion with caregiver outcomes. First, there is a need for studies that use psychometrically tested measures of religion. Second, they note that the variability in results may be, in part, due to the multidimensional nature of the religious measures and that religion needs to be measured multidimensionally. Third, given research that suggests ethnic differences in religiosity (e.g., Dilworth-Anderson, Williams, & Gibson, 2002), they suggest that there is a need to examine ethnicity as a potential moderator of effects of religion on caregiver outcomes. Given that much of the research in this area examined both men and women caregivers and that women are generally more involved in religion than men (Levin, Taylor, & Chatters, 1994; Maselko & Kubzansky, 2006), it would also seem likely that gender should be evaluated as a moderator in studying the association of religion and caregiver outcomes. Finally, it is apparent that most research in this area is cross-sectional. There is a need to see whether earlier religious variables can predict caregiver outcomes in new caregivers.
Purpose of the Study
The purpose of this study was to use scales which have previously been psychometrically tested to determine whether religion measured multidimensionally before an individual became a caregiver predicts caregiver burden, mental health, and physical health after the same individual became a caregiver.
Method
Participants and Design
This study used a prospective cohort research design. Approval was obtained from the Institutional Review Board of Loma Linda University. To ensure confidentiality, all participant identifiers were removed. Data from the Biopsychosocial Religion and Health Study (BRHS; Lee et al., 2009), which is a substudy of the 96,000-person Adventist Health Study–2 (AHS-2; Butler et al., 2008), were used. AHS-2 targeted Black and White North American Seventh-day Adventist and, hence, the numbers of other ethnic groups in the study were typically too small to use. AHS-2 primarily sampled individuals attending Seventh-day Adventist churches and so the vast majority were Seventh-day Adventists. In the first wave of the study (2006-2007), a random sample of approximately 21,000 individuals were sent a 20-page survey on heath, cumulative risk, various intervening variables, and religion, and 10,988 individuals responded. All of these respondents were sent another questionnaire in 2010-2011 in the second wave of data collection. Of the 6,600 respondents, 585 became informal caregivers in the interval between the first and second wave and were used as the sample for this study. One was dropped because of missing data on caregiver burden. Of the remaining 584 caregivers, 460 were female and 124 male.
As for inclusion and exclusion criteria, the vast majority of study participants, 94.4% in the first wave of the BRHS, identified themselves as members of the Seventh-day Adventist denomination. As there were so few non–Seventh-day Adventists, they were dropped from the analysis. The participants who responded in the second wave also recorded information about caregiver burden. Data on religiosity were taken from the first wave, whereas caregiver burden and caregiver status data were taken from the second wave.
Measures
Caregiver status and burden
In the second wave of the BRHS (2010-2011), respondents answered the following question: “Not including paid work, have you been primarily responsible for giving direct care for a mentally or physically ill person for more than six months?” Individuals could respond to any or all of three time periods: in the last year, 2 to 3 years ago, or more than 3 years previous. If the respondents indicated they had been caregivers in the last year or in the 2 to 3 years previous (the period between the two waves of the study) but not in an earlier time period, they were considered new caregivers. Then they were asked, “If you are currently a caregiver please describe your caregiving experience in the last several months.” They responded on three, 10-point rating scales regarding the degree to which their experience had been straining, hard, or burdensome (Cronbach’s α = .95).
SF-12 physical and mental health composites
These were based on the short 12-question version (SF12v2) of the longer 36-item questionnaire and were assessed during the first wave of the study (2006-2007). Questions covered items such as general health, whether or not their physical functioning was limited, what their energy levels were like, and whether they were depressed. Internal consistency has been found to be .88 and .82 for physical health and mental health, respectively, by Cheak-Zamora, Wyrwich, and McBride (2009). The items were scored following the protocol in Ware, Kosinski, Turner-Bowker, and Gandek (2002).
Gender and ethnicity
These are discussed in the “Statistical Analysis” section of this article.
Religiosity
Fifteen variables measured in the first wave were used to assess religiosity. These variables were as follows:
Intrinsic religiosity
Intrinsic religiosity was measured using the last three items of the DUREL Index (Koenig, Parkerson, & Meador, 1997)—“I try hard to carry my religion over into all my other dealings in life,” “In my life I experience the presence of the divine,” and “My religious beliefs are what really lie behind my approach to life” all with possible answers ranging from not true to very true on a 7-point scale (Cronbach’s α = .71).
Private prayer and meditation
Time spent in personal religious activities was measured using two items created by splitting the second item on the DUREL Index (Koenig et al., 1997)—“How often do you spend time in private bible study” and “How often do you spend time in private prayer or meditation” both with responses ranging from never to more than once a day on an 8-point scale (Cronbach’s α = .74).
Church organization activity
Time spent in public or congregational religious activities was measured by three items from King and Hunt (1990)—“I keep pretty well informed about my congregation,” “I have some influence on the decisions of my congregation,” and “Church activities (meetings, committee work, etc.) are a major source of satisfaction in my life.” All responses range from not true to very true on a 7-point scale (Cronbach’s α = .77).
Positive religious coping
Positive religious coping was measured by five items from the RCOPE Scale for religious coping (Pargament, Koenig, & Perez, 2000)—The overarching question was “Thinking about how you have tried to understand and deal with major problems in your life, to what extent has each of the following been involved in the way you cope?” The five items were as follows: “Worked together with God as partners,” “Tried to find a lesson from God in the event,” “Looked to God for strength, support and guidance,” “Thought about how my life is part of a larger spiritual force,” and “Confessed my sins and asked for God’s forgiveness.” Responses range from not at all to a great deal on a 5-point scale (Cronbach’s α = .74).
Negative religious coping
Negative religious coping was measured using a 6-item scale taken from the RCOPE religion measurement tool with the same overarching question used to assess positive religious coping (Pargament et al., 2000). The six items were as follows: “Tried to make sense of the situation without relying on God,” “Felt punished by God for my lack of devotion,” “Decided that God was punishing me for my sins,” “Wondered whether God had abandoned me,” “Questioned whether God really exists,” and “Expressed anger at God for letting terrible things happen.” All responses ranged from not at all to a great deal on a 5-point scale (Cronbach’s α = .72).
Loving versus controlling God
Loving versus controlling God (Hill & Hood, 1999) was measured using a five-item semantic differential format. Two examples of word pairs used were saving versus damning and loving versus hating. Individuals were given seven choices between each pair of anchor words (Cronbach’s α = .78).
Spiritual meaning in life
To measure spiritual meaning in life, participants were asked to answer five questions from Mascaro, Rosen, and Morey (2004) about whether or not they “see a special purpose for myself in this world” and whether or not they thought their “life is meaningful.” Reponses ranged from not true to very true on a 7-point scale (Cronbach’s α = .77).
Confession prayer
Confession prayer was measured using a five-item scale (Luckow et al., 1996). Participants were asked to respond to five items that included “It is important to me to tell God about my sins or faults” and “When I feel guilty about something, it helps to tell God about it.” All responses ranged from definitely false to definitely true on the 8-point scale (Cronbach’s α = .78).
Habit prayer
Habit prayer was measured with four items (Luckow et al., 1996). Example items include “A morning prayer helps me cope with the world during the day” and “I pray daily.” The responses ranged from definitely false to definitely true on the 8-point scale (Cronbach’s α = .77).
Contemplative prayer
Contemplative prayer was measured using five items (Poloma & Pendleton, 1991). These included “Spend time just ‘feeling’ or being in the presence of God” and “Spend time reflecting on the Bible.” Responses ranged from never to very often on a 5-point scale (Cronbach’s α = .77)
Congregational sense of community
Congregational sense of community was measured using a 10-item scale (Hill & Hood, 1999). Examples of the items are “Members usually introduce themselves to new members” and “The clergy know most of the members by name.” Responses ranged from not true to very true on a 7-point scale (Cronbach’s α = .77).
Religious emotional support received and religious emotional support given
Religious emotional support received (Cronbach’s α = .78) and religious emotional support given (Cronbach’s α = .78) were measured using two, 3-question scales (Krause, 1999). Example items are “How often do people you worship with make you feel loved and cared for” to measure emotional support received and “How often do you make the people you worship with feel loved and cared for” to measure emotional support given. Responses ranged from never to always on 5-point scales.
Negative interactions at church
Negative interactions at church were measured using a three-item scale (Krause, 1999). One item was “How often are people you worship with critical of you and the things you do?” The 5-point scale responses ranged from never to always (Cronbach’s α = .78).
Anticipated church support
Anticipated church support was measured using a three-item scale (Krause, 1999; Pargament, 1999) and included the item, “If you needed to know where to go to get help with a problem you were having, how much would the people in your congregation be willing to help out?” Responses on the 5-point scale ranged from never to always (Cronbach’s α = .77).
Control variables
Age, gender, and ethnicity were the only control variables used. To determine age, data were gathered using the date of birth variable in the demographics section of the questionnaire and subtracting that from the date the questionnaire was returned. Gender and ethnicity were determined by self-report.
Statistical Analysis
As mentioned in the “Participants and Design” section, we ended up with 584 participants because individuals who were caregivers and missing more than 20% of the items used in analyses were dropped from the final total. All 584 participants answered the ethnicity question, 559 of the 584 participants answered the question about their ethnicity, and 580 of the 584 respondents gave information on their education level. Ten iterations of multiple imputation (Graham, 2009) were used to account for individuals missing less than 20% of the data on ethnicity and education from the remaining 584 participants in regression analyses. Multiple linear regression was used to determine associations of the religion variables with caregiver burden, mental health, and physical health. Age, gender, and ethnicity were controlled in all regression analyses. Gender and ethnicity were moderator variables. Each respondent replied as being either male or female and indicated their ethnicity. The ethnicities used in this study are Black and White because there were too few other ethnicities to be included.
Interactions of religion variables with gender and ethnicity in predicting caregiver outcomes were also performed; however, only statistically significant interactions are reported below. Individual regressions were run regressing caregiver burden on each of the 15 religion variables separately (with controls); then a combined regression model was run regressing caregiver burden on all of the religion variables simultaneously. This approach was also used in predicting mental and physical health. However, so we could examine whether religion variables predicted a change in mental or physical health, the Wave 1 values for these variables were controlled in these analyses. (Control of Wave 1 caregiver burden was not possible in that analysis as caregiver burden was not measured in Wave 1). No associations were found for change in physical health regressed on the religion variables so we do not report the regressions involving physical health below. All analyses were carried out using SPSS 22, and results were considered significant if they had a p value of ≤.05.
Results
Table 1 shows how the religion variables related to ethnicity and gender in the caregivers. Black men caregivers viewed God as less loving and more controlling than caregivers who were Black women, White women, or White men. When it came to finding spiritual meaning in life, Black female caregivers reported this more than White female caregivers. Positive religious coping was used more by Black female caregivers than White men and women caregivers. Black female caregivers reported engaging in confession, habit, and contemplative prayer more than White female and White male caregivers. Black male caregivers reported experiencing negative interactions at church more than Black female, White male, or White female caregivers.
Values of Religion Variables by Gender and Ethnicity (N = 584).
Note. Values in the same row not sharing the same subscript are significantly different at p < .05 in the two-sided test of equality for column means. Cells with no subscript are not included in the test. Tests assume equal variances. Tests are adjusted for all pairwise comparisons within a row of each innermost subtable using the Bonferroni correction.
Table 2 shows that there was no significant relationship found between the burden of being a caregiver and physical health outcomes; however, caregiver burden did predict poorer mental health outcomes. Older age predicted poorer physical health while being male was associated with better physical and mental health.
Regression of Caregiver Burden on Physical and Mental Health With Age, Gender, and Ethnicity Controlled (N = 584).
Note. CI = confidence interval.
Table 3 shows the association of religion variables with caregiver burden. Caregiver burden was lower for those who saw God as loving and not controlling, and for those who felt a sense of community with their congregation or church family. When all variables were added into the regression simultaneously, no religious variable predicted caregiver burden though belief that God was loving and not controlling was significant at p < .10. However, in general, religion did significantly improve prediction of caregiver burden as evidenced by the statistically significant R2 change (.103), F(18, 550) = 3.494, p < .0005, when the religion variables were added as a unit. Interactions of the religion variables with gender and ethnicity were tested; however, none were statistically significant.
Regression of Caregiver Burden on Religious Variables With Age, Gender, and Ethnicity Controlled (N = 584).
Note. R2 change for adding all religious variables to control variables was .103, F(18, 550) = 3.494, p < .0005. CI = confidence interval.
p < .10. *p < .05.
Then we looked at gender to determine whether religiosity affected the mental health of caregivers. The interaction term of gender and negative religious coping on the mental health of caregivers was statistically significant (B = 4.37, confidence interval [CI] = [0.36, 8.38], p = .033). This was the only significant interaction term. Negative religious coping predicted caregiver burden more strongly in women (B = −6.33, CI = [−8.27, −4.40], p < .0005) than in men. As with caregiver burden, in general, the full set of religion variables improved prediction of mental health (R2 change = .306), F(19, 467) = 10.839, p < .0005. There were no significant interactions between religion variables and ethnicity.
As this is a prospective study and as mental health was strongly related to reported caregiver burden, we decided to investigate whether or not there was a change in caregiver health over time and we discovered that there was a small but significant (p = .003) reduction in reported physical health of the caregivers between the first and second waves. The mean of physical health in Wave 1 was 48.45, and in Wave 2, it was 47.22. This shows a decline in physical health, which is expected as caregivers aged during this period. There was also a change in mental health from a mean of 52.26 to 52.03 but it was not significant (p = .575).
Finally, to see whether any difference between the waves could be predicted by the religion variables, we ran multiple generalized linear regressions controlling for mental health in Wave 1 with mental health in the second wave as the outcome variable against the religion variables. These results are outlined in Table 4. We found that improved mental health was influenced by seeing God as loving and not controlling and by emotional support given to church members. Mental health change was negatively affected by use of negative religious coping methods. When a combined regression was run, the only religion variable that remained significant was loving versus controlling God. There were no statistically significant interactions between mental health and measures of religiosity.
Regression of Wave 2 Mental Health on Religious Variables With Age, Gender, Ethnicity, and Wave 1 Mental Health Controlled (N = 584).
Note. R2 change for adding all religious variables to control variables was .306, F(19, 467) = 10.839, p < .0005. CI = confidence interval.
p < .10. *p < .05.
Discussion
According to the National Alliance for Caregiving and the American Assosication of Retired Persons (AARP), in 2015, there were more than 43.5 million informal caregivers over the age of 18, providing care for a family member with an illness or disability over the last 12 months. The majority of the caregivers are women. The breakdown of caregivers by race is 62% White, 17% Hispanic, 13% African American, and 6% Asian American . Table 5 shows the demographics of the study participants. Most of the female caregivers were between 55 and 64 years of age (M = 60.98, SD = 13.68) while most of the male caregivers were between 65 and 74 years old (M = 62.46, SD = 13.46). We had only 559 who identified their ethnicity. Of these 559, 50.3% were White females, 28.3% Black female, 16.6% White male, and 4.8% Black male. Because the parent AHS-2 study was designed to compare cancer rates in Black and White North American Seventh-day Adventists, the sample size of the 26 caregivers who identified as neither White nor Black was too small, and therefore, they were excluded from the analysis. The largest group of the women reporting their education, 78.7% (359/456), reported at least some college. All 124 men reported their education status, and the majority (83.8%) reported at least some college. To some degree, these statistics mirror the findings of the 2015 data regarding the more than 43 million caregivers in the United States. In both instances, there were more women caregivers than men and more White caregivers than caregivers of other ethnicities. The one notable difference is that in our study, there were more African American caregivers than Hispanic caregivers. This was likely the result of African Americans being oversampled in the dataset, however. In fact, the number of Hispanic and Asian American caregivers in our study was so low that they were excluded from the analysis. We also included information on caregiver education although this did not prove to be a significant factor in our analysis. The interaction between gender and ethnicity was tested also but there were no significant findings so they were not included in the table or discussion.
Demographics of Study Participants (N = 584).
Note. Due to missing data, totals for age, ethnicity, and highest education obtained do not each add up to 584.
The relationship between caregiver burden and religion still needs further study; however, current research on caregiving and religion shows that caregivers use religion to reduce stress and give them strength to face stressful events (Hebert, Koenig, Arnold, & Schulz, 2006; Klassmann, Kochia, Furukawa, Higarashi, & Marcon, 2008). Yeh and Bull (2009) found that belief in a higher being and a prayer life were associated with better mental health for caregivers while Newberry et al. (2013) in their longitudinal study of 50 caregivers discovered that over an 8-month period, caregivers reported higher levels of depressive symptoms if the care recipient became more ill. However, these symptoms occurred far less in caregivers who reported high levels of spirituality. Religion is not always a benefit to caregivers as Rabinowitz et al. (2010) found that negative religious coping was associated with a weight gain among the 256 Latina and Caucasian female caregivers, which can affect physical health adversely. This may be related to our finding of an association of negative religious coping with a reduction in mental health. Mannan, Mamun, Doi, and Clavarino (2016) in a meta-analysis of 21 studies found that those who were obese had an increased risk of being depressed, and Hemmy Asamsama, Lee, Morton, and Tonstad (2015), using our dataset, found a positive relationship between body mass index (BMI) and later depression. Others have found that religion influences how caregivers approach their work and can help them deal with the challenges they face when caring for ill family members (Nightingale, 2003).
The purpose of this study was to learn more about the relationships between caregiving burden and religiosity, and how caregivers can reduce burden through forms of religion that they already use. We were unable to find a link between the burden of being a caregiver and impaired physical health. Mental health, however, was predicted by the burden of being a caregiver. In individual regressions, caregivers who viewed God as more loving than controlling had lower burden. The same is true for those caregivers who felt a sense of community with their church. When we attempted to investigate whether there was any interaction between these religion variables and gender in predicting mental or physical health, we found only one significant result. Caregiver women who used negative religious coping methods, such as questioning God’s love, voicing “anger that God didn’t answer [their] prayers,” or felt that God was punishing them, experienced a greater loss of mental health. The same was not true for men, however. There were no significant interactions between religion variables and ethnicity.
Further analysis indicated that there was a statistically significant decline in physical health over the 3-year interval between Waves 1 and 2, but this did not obtain for mental health overall. We also discovered that religion played a role in the change of the mental health between the two waves. Positive aspects of religion, specifically viewing God as a loving and not controlling God and giving emotional support to other church members were associated with improved mental health between Waves 1 and 2 of the survey.
Shah, Snow, and Kunik (2001) found that caregivers of Alzheimer’s patients had higher levels of spirituality and greater use of religious coping mechanisms when dealing with caregiver burden compared with an age matched control group of noncaregivers. Caregivers who questioned their faith and felt anger toward God or felt that He was distant displayed more depressive symptoms and had more perceived burden. Our study confirmed that negative religious coping contributes to caregivers experiencing greater loss of mental health in an already difficult situation but primarily for women. Another finding regarding negative religious coping was that it can negatively affect the mental health of caregivers over time. We were also able to discover that a sense of community fostered in their church helps caregivers deal with the burden of caregiving. This is new information that can be added to the literature as these expand on church support. We now have definable areas of church support that can aid caregivers in the future, and this includes support given to other church members. Conventional thinking would suggest that receiving support would help to protect mental health, but in this study, giving support to others was beneficial to later mental health. This aspect of support has not been explored in the literature so future research should be done to determine whether this can be replicated and whether it can be applied to other religious and ethnic groups.
Strengths and Limitations of the Study
One strength of this study is that it is a prospective study. We were able to follow the participants over time and establish temporal relationships between physical health outcomes, mental health, and caregiver burden. This point is strengthened because we were able to follow only those individuals who became caregivers for the first time during the interval between our two waves of measurement. Another strength is that we were able to sample a much wider range of religion-related variables than is typically the case. We were also able to sample a relatively large population of Whites and more importantly Blacks who are often underrepresented in research. A third strength of this study is that religiosity was measured in the first wave of the study and caregiver status and burden were measured in the second wave, thus reducing the likelihood of response bias. Assessing religiosity before gathering information on caregiving was important in reducing the risk that the responses of new caregivers would be generated based on desired or expected religiosity results.
This study was not without its limitations. The fact that we were able to examine so many aspects of religion simultaneously meant that in the combined regressions, some associations were likely to be weakened because of the intercorrelation among the religious variables. This suggests that we should probably place more reliance in the regressions of the individual religious variables on potential outcomes than on the regressions combining all religious variables together. However, in predicting both burden and mental health, the addition of all the religious variables as a set significantly improved prediction. Also, although we were able to sample a relatively large population of Whites and Blacks, which is a strength of this study, we did not have close to as many Black respondents as White respondents. There were no respondents from other ethnicities so these results cannot be applied to them. Future research should be used to explore whether or not these findings are similar in other ethnicities. Another limitation is the homogeneity of the population regarding religion. All study participants were Seventh-day Adventist, and this may have increased the risk of participants responding based on what they felt was socially acceptable to conform to Seventh-day Adventist norms. However, this homogeneity makes it more likely that the religion items are interpreted similarly across participants. In addition to all the respondents being Seventh-day Adventist, the largest group of women and the majority of men had at least some college education which could limit generalization of results by excluding to individuals who may not have received as much exposure to education. Although unintended, this may have led to omission bias, as the responses of those with less than some college education are not as greatly represented in this study. As with all self-report surveys, there may be noteworthy differences between those who chose to respond and those who did not. In addition, we did not have information on whether the care recipients were children or adults, spouses or parents, nor did we have information about the reason the care recipient needed care. Therefore, no conclusions can be drawn as to how caregiver burden may differ or be similar when providing care for people of different ages or need-for-care conditions.
Finally, it is important to note that religiosity can and might have changed for caregivers during the interval between the collection of data in the first and second waves. This could then affect the outcome of mental and physical health of caregivers. Although this was outside the scope of this article, this relationship should be explored in future research.
Conclusion
We looked at how specific religion variables could play a role in helping caregivers cope with the pressures they face. We found that there are aspects of religion that can help with the mental strain caregivers experience. To reduce burden, religious social support in the form of feeling a sense of community proved helpful. Another helpful aspect of religion is the caregiver’s view of God. Those who see God as a loving God are less likely to experience burden due to being a caregiver than those who see God as controlling. In addition to burden reduction, seeing God as loving, forgiving, and saving and giving emotional support to other church members are both protective against loss of mental health over the 4 years of the study. Of note is the fact that giving emotional support to other church members even while dealing with the burden of being a caregiver actually had a positive impact on mental health. Previous research has shown that support from a religious community acts as a buffer for caregiver burden (Marquez-Gonzalez, Lopez, Romero-Moreno, & Losada, 2012), and we did find that a congregational sense of community was associated with less burden. However, support given by religious individuals to their religious community has not up to this point been explored in the literature about religion and caregiving and suggests that altruism and being outward focused can be beneficial to the caregiver.
Footnotes
Acknowledgements
The authors acknowledge the Biopsychosocial Religion and Health Study (BRHS) committee for supporting this research and sharing the dataset for analysis.
Authors’ Note
This article was completed as a part of the dissertation of Carla R. A. Fider for the degree of doctor of public health (DrPH) at Loma Linda University.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Funding for the original Biopsychosocial Religion and Health Study (BRHS) was provided by the National Institute on Aging (1R01AG026348-01), and funding for the parent study—Adventist Health Study–2—was provided by the National Cancer Institute (5R01CA094594).