Predictors of improvement in social support: Five-year effects of a structured intervention for caregivers of spouses with Alzheimer's disease
Introduction
An abundance of research shows that social support is important for physical and mental health (Berkman, Glass, Brisette, & Seeman, 2000; Cohen & Syme, 1985; Seeman, 2001; Stansfeld, 1999). Recently, social support has become the target of interventions to improve mental health outcomes (Burgio, Stevens, Guy, Roth, & Haley, 2003; Pillemer & Suitor, 2002; Turner, 2003; Turner & Marino, 1994). When social support is provided by counselors and support groups, individuals have a chance to express positive and negative emotions when coping with hardship (Spiegel & Kimerling, 2001). Discussing problems with counselors or other caregivers allows them to improve family communication, identify needs, and adjust to a disease that progresses and changes (Mittelman et al., 1997; Spiegel & Kimerling, 2001).
Social support is especially important for the spouses of those afflicted with Alzheimer's disease (AD), as these caregivers are generally elderly themselves, and often have their own physical limitations (Haley et al., 1996; Mittelman et al., 1997; Pinquart & Sörensen, 2003; Worcester & Hedrick, 1997). Many spouse-caregivers become isolated from family members and former friends because they may be reluctant to leave their spouses at home alone, or to go out with them. Moreover, their own health limitations may also preclude them from leaving the house regularly (Mittelman et al., 1997; Worcester & Hedrick, 1997). Finally, the behavioral problems associated with the middle to late stages of AD, such as night wandering, incontinence, agitation, and increasingly impaired memory that ultimately causes the patient to be unable to recognize his or her caregiver, can be especially difficult for caregivers (Mittelman, Roth, Haley, & Zarit, 2004; Roth et al., 2003). An intervention that helps caregivers mobilize and utilize their social support networks may help them adapt better to their caregiving responsibilities.
Being satisfied with one's social support network means not only that there is an availability of social support, but that it is at the desired level (Stansfeld, 1999; Stokes, 1983). Feeling as though one has the right amount of social support in one's circumstances leads to a well-balanced life (Cohen & Syme, 1985). Satisfaction with social support goes beyond received social support and taps into perceived support, which is important in predicting other mental health outcomes such as depression and overall sense of well-being (Turner & Marino, 1994; Wethington & Kessler, 1986).
The New York University Silberstein Aging and Dementia Research Center (NYU-ADRC) has been conducting a pioneering randomized trial of a psychosocial intervention for spouse-caregivers of people with AD since 1987. The intervention was designed to provide counseling and enhance social support over the entire course of the disease, and through the bereavement period. A total of 406 spouse-caregivers were enrolled over a period of 9.5 years and randomly assigned, after a comprehensive baseline interview, to receive an enhanced multicomponent treatment, or to participate in a usual care control condition. These caregivers enrolled in two successive cohorts: 206 subjects were recruited between 1987 and 1991 and an additional 200 subjects were recruited over a 5.5-year period between 1991 and 1997.
Previously published findings of this study show that the intervention improved many aspects of well-being for spouse-caregivers. The multidimensional treatment has been shown to delay nursing home placement (Mittelman, Ferris, Shulman, Steinberg, & Levin, 1996), and reduce depressive symptoms (Mittelman et al., 1995). The effect on depression was sustained for at least 3 years (Mittelman, Roth, Coon, & Haley, 2004). The severity of spouse-caregivers’ reactions to the troubling behavior problems associated with AD was also alleviated by the intervention, and this effect continued for at least 4 years (Mittelman, Roth, Coon, & Haley (2004), Mittelman, Roth, Haley, & Zarit (2004)). Finally, the intervention improved satisfaction with social support, which in turn predicted better mental health outcomes for the treatment group (Roth, Mittelman, Clay, Madan, & Haley, 2005).
Social support is a complex concept that has been defined in a multitude of ways (Vaux, 1988). The level of social support of an individual has commonly been examined with measures such as size of social network, density, frequency of contact and reciprocity (House, Kahn, McLeod, & Williams, 1985). Both social network measures and social support measures have been used in health outcome studies, and have been shown to lead to better physical and mental health outcomes (Stansfeld, 1999). In these studies, emphasis is typically given to the effectiveness of the members of the network in providing social support, which is often divided into instrumental support, emotional support and informational support (Wills, 1985). Instrumental support helps individuals with daily living needs and includes elements of physical help such as cleaning house, moving/bathing, and driving. Emotional support includes being available in times of need in order to have someone to confide in and lean on. Informational support includes advice and knowledge. This includes both formal advice from health professionals such as physicians and counselors; as well as informal advice from others who have experienced similar situations (Drentea & Moren-Cross, 2005). In this paper, we use these categorizations of social support, which were added to the assessment of the second cohort of participants in order to provide greater detail about the changes in social support that were observed in response to the intervention.
The multicomponent intervention was designed to improve the caregiving skills of and social support for the spouse-caregivers. Additionally, the intervention's goal was to have spouse-caregivers decide what type of support they wanted and then learn how to draw upon their networks to satisfy their needs. A major objective of the intervention was to promote communication among family members. The intervention included individual and family counseling, participating in a support group and ad hoc counseling (further counseling as requested). All counselings were done by experienced clinicians, with degrees in social work, counseling, psychology or gerontology. Individual and family counseling was custom-tailored to meet the needs of each family. In the first 4 months of the study the caregivers in the treatment group received two individual counseling sessions, one before and the second after a series of family counseling sessions. At the end of this 4-month period, spouse-caregivers were required to join support groups that met weekly. Additionally, counselors were available by telephone to caregivers and other family members to discuss any problems as long as they participated in the study. Caregivers also occasionally made appointments to speak to counselors face-to-face.
If this multicomponent intervention was efficacious in helping spouse-caregivers garner their social support to the level desired by the spouse-caregivers, then we should see improved levels of support satisfaction among the group that received the intervention. We formulated four hypotheses: (1) We hypothesized that caregivers who received the enhanced intervention would experience higher rates of satisfaction with social support over time compared to the usual care control group. (2) We expected that the growth in support satisfaction would be associated with increases in instrumental, emotional, and informational support, as well as network characteristics, including more frequent phone calls and visits by support network members, a larger support network and number of close network members. (3) We also expected that receiving the desired level of contact (i.e. not wanting more or less contact) would be associated with satisfaction with support measures. (4) Additionally, we hypothesized that improvements in these specific areas of social support would partially explain why the enhanced intervention group would have higher satisfaction with support than the usual care control group.
The participants in this investigation consisted of the 200 spouse-caregivers enrolled in the second cohort of the clinical trial conducted at the NYU-ADRC. Participants were eligible for the study if they were caring for a husband or wife with a diagnosis of AD living with them at home, regardless of the severity of dementia. Caregivers received regular follow-up assessments, using structured interview instruments, through nursing home placement and for 2 years after the death of the patient. The study has an unusually high retention rate; only 4.7% of caregivers refused to continue in the study while the patients were still living at home. The treatment group received the multicomponent intervention described above, while the caregivers in the usual care group received resource information and referrals on request, but no formal individual or family counseling sessions.
In the first year after enrollment, there were four assessments: baseline, 4 months, 8 months and 12 months. After the first year, spouse-caregivers were interviewed every 6 months. The data analysis reported herein was limited to assessments in the first 5 years after enrollment of caregivers of relatives still living in the community. We did not include data for caregivers after their spouse entered the nursing home because the nature of their needs for social support would be different than for those who still had spouses living in their home.
The social support assessment included measures from the Stokes Social Network List (Stokes, 1983). The assessment measured the size and composition (i.e. number of family members and friends) of the caregiver's network. Using the methods of Stokes (1983), caregivers were asked to list all of the “people who are significant in your life and with whom you have contact at least once a month,” up to a maximum of 20 individuals, and their relationship (family or friend) to each person. The detailed assessment developed in 1996 for the second cohort of the NYU study added the precise relationship of each family member (sibling, parent, friend, etc.). At baseline, caregivers were most likely to list children (43%), friends (26%), siblings (13%), and other family members (10%) as individuals who were significant in their lives. The remaining 8% of mentions at baseline were divided among grandchildren, neighbors, other, parents, household help, and counselors (in order of most to least). No caregiver specified support group members as a part of their social support network at baseline.
Additional items, developed at the NYU-ADRC, added in-depth information about social support provided by each member of the caregiver's social network who was listed in the Stokes Social Network List. These items included questions about the amount and adequacy of support received and whether the caregiver would go to the person in an emergency. The questionnaire included questions about whether caregivers wanted more or less contact, as it has been shown that negative support can be detrimental to one's well-being (Newsom, Nishishiba, Morgan, & Rook, 2003). Caregivers listed the amount of support they received from each person and how helpful it was in five key areas: emotional, physical, socializing, financial and advice, each on a four-point scale. These measures were recoded into fewer categories in order to make them both theoretically meaningful and methodologically viable. First, responses to questions indicating whether or not the caregiver believed that he/she was receiving an adequate amount of each type of support were dichotomized: “none,” “too much,” and “not enough” were coded to 0, and “enough” was coded to 1. Helpfulness of support was coded as “upsetting”=0, “somewhat helpful”=1, “very helpful”=2, to create an index with the highest number being most helpful. If amount of support was reported as none in a specific area, helpfulness was originally coded as not applicable in that area. For the five detailed dimensions of social support, only 1–3% of caregivers reported that the help they received from a support group member was “upsetting.” Therefore, “not applicable” was recoded to 0 along with “upsetting,” for parsimony (which precluded the possibility of explicitly examining the negative aspects of social support).
Because each spouse-caregiver could list up to 20 members in their social network, each measure on the detailed social support assessment could have up to 20 responses (i.e. one for each person to whom they were referring). Mean amount and helpfulness scores were computed for each of the five areas by summing the scores generated for each member of the social support network and dividing by the total number of members in the network. The mean amount and helpfulness of financial and physical support were summed to form an “instrumental” social support measure (range=0–6), while the mean amount and helpfulness of emotional and socializing support were summed to form an “emotional” social support measure (range=0–6). Finally, the mean amount and helpfulness of advice created the “informational” support measure (range=0–3). This coding reduced the number of measures to test in the random effects regression models and helped to avoid multicollinearity of predictors, while maintaining theoretical clarity of the three types of social support usually tested in social support literature. These summary scores were used in the analyses.
Caregivers were also asked, “Would you like more or less contact with this person?” Responses were 0=less, 1=same and 2=more for each person, and an average was taken. Amount of social interaction was measured using both the average frequency of phone conversations and the average frequency of people visiting in the social support network in the past 4 months; responses were coded using a Likert scale so that 0=never, .5=less than once a month, 1=once a month, 4=several times a month, 15=several times a week, and 30=daily. This coding produced continuous frequency measures of phone conversations and people visiting for each respondent.
The dependent variable, satisfaction with social support was created from three indicators. Caregivers were asked (1) “in general, how satisfied are you with your social network,” (2) “how satisfied are you with the assistance you get from your social network in daily activities (help with chores, patient care),” and (3) “how satisfied are you with the emotional support you get from your social network (feeling cared about, discussing problems).” All three items were answered on six-point Likert scales (6=very dissatisfied, 1=very satisfied; Stokes, 1983) and were asked concerning the entire network as a whole and not each individual member. This scale was reverse coded for comprehensibility such that higher scores reflected greater satisfaction. Cronbach's alpha revealed a high level of internal consistency between these three measures at baseline ; thus, the average was used as the composite satisfaction with social support indicator (range=1–6).
The aims of the analyses were (1) to determine if there were group differences in satisfaction with the social support network over time as a result of the intervention, (2) to examine multiple dimensions of social support and describe the specific factors that were affected by the intervention and associated with higher levels of satisfaction with social support, (3) to ascertain whether receiving the desired level of contact would be associated with higher levels of satisfaction with social support and (4) if there were group differences in satisfaction with social support, to identify factors that might explain this difference.
We used random effects regression models to see if the treatment group enjoyed higher levels of satisfaction with social support compared to the control group even after adjusting for baseline differences. The use of random effects regression models, also called multilevel modeling, allowed us to examine how each person changed over time as well as to describe how these changes differed across people (Singer & Willett, 2003). This technique has the advantage that it is possible to utilize data from all subjects with at least two assessments. Another advantage of using random effects regression models is that they do not assume that the interval between assessments is fixed. This enabled us to include data from participants with missing data at certain occasions, or who missed their appointment by 1 or 2 months in the analysis (Rogosa, 1995). These models also focus more closely on individual history (as they examine each individual unit) than traditional analytic models, making it possible to answer questions about individual rates of change, the correlates of that change (including individual difference variables), and group mean rates of change (Francis, Fletcher, Stuebing, Davidson, & Thompson, 1991).
Section snippets
Results
The results are based on the 183 participants from the second cohort who provided baseline data and at least one post-intervention assessment while still providing care in the home (94 of these individuals were in the treatment group and 89 in the usual care group). By the 11th assessment, approximately 4.5 years from baseline, 58 of these 183 (32%) were still providing care in the home and active in the study. In addition, there was no difference in attrition from the analysis between the
Discussion
In examining the effect of a counseling and support intervention on the growth of social support satisfaction over time, we found that the group that received the intervention experienced consistently higher levels of support satisfaction than the usual care control group beginning at the first follow-up, 4 months after baseline; thus, the first hypothesis was supported. All components of the intervention used in this study taught spouse-caregivers to mobilize their social support. Four months
Acknowledgments
We would like to thank the NYU Counseling Team and the caregivers who participated in this project. This research was supported by grants from the National Institute of Mental Health (R01 MH42216) and the National Institute on Aging (R01 AG14634). Additional resources were provided by the NYU Alzheimer's Disease Center (P30 AG08051).
References (33)
- et al.
From social integration to health: Durkheim in the new millennium
Social Science & Medicine
(2000) - et al.
Religion and coping
- et al.
The economic value of informal caregiving
Health Affairs
(1999) - et al.
Impact of two psychosocial interventions on white and African American family caregivers of individuals with dementia
The Gerontologist
(2003) - et al.
Caregiving outside of the home: The effects of race on depression
Ethnicity & Health
(2006) - et al.
Social capital and social support on the web: The case of an internet mother site
Sociology of Health and Illness
(2005) - et al.
Analysis of change: Modeling individual growth
Journal of Consulting and Clinical Psychology
(1991) - et al.
Appraisal, coping, and social support as mediators of well-being in black and white family caregivers of patients with Alzheimer's disease
Journal of Consulting and Clinical Psychology
(1996) - et al.
Measures and concepts of social support
Effects of a multicomponent support program on spouse-caregivers of Alzheimer's disease patients
A comprehensive support program: Effect on depression in spouse-caregivers of AD patients
The Gerontologist
A family intervention to delay nursing home placement of patients with Alzheimer disease: A randomized controlled trial
Journal of the American Medical Association
Sustained benefit of supportive intervention for depressive symptoms in Alzheimer's caregivers
American Journal of Psychiatry
Effects of a caregiver intervention on negative caregiver appraisals of behavior problems in patients with Alzheimer's disease: Results of a randomized trial
Journals of Gerontology, Psychological Sciences
The relative importance of three domains of positive and negative social exchanges: A longitudinal model with comparable measures
Psychology and Aging
Cited by (105)
Social support moderates the effect of stress on the cortisol awakening response in dementia family caregivers
2021, Physiology and BehaviorCitation Excerpt :Nevertheless, these findings hold much promise for future caregiver interventions, as social support is not only a modifiable factor, but one that can be relatively easily enhanced. It is also likely to be acceptable to caregivers, who report dissatisfaction with the level of support received in the absence of intervention [20]. Social support is particularly important for dementia caregivers as they may be reluctant to leave the care recipient at home alone, or conversely go out with them [20], and therefore become isolated from family members and friends.
Social connections and participation among people with mild cognitive impairment: barriers and recommendations
2023, Frontiers in PsychiatrySocial Prescription Interventions Addressing Social Isolation and Loneliness in Older Adults: Meta-Review Integrating On-the-Ground Resources
2023, Journal of Medical Internet ResearchTypology of family caregivers of older persons: A latent profile analysis using elder mistreatment risk and protective factors
2023, Journal of Elder Abuse and Neglect