| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLE |
From the Behavioral Medicine Research Center and the Department of Psychiatry and Behavioral Sciences (B.H.B., J.C.B., I.C.S., R.B.W.), Duke University Medical Center; the Outcomes Research and Assessment Group, Duke Clinical Research Institute (D.B.M., N.E.C.-C., B.L.L.), Duke University Medical Center; and Health Services Research and Development (H.B.B.), Durham Veterans Affairs Medical Center, Durham, North Carolina.
Address reprint requests to: Beverly H. Brummett, PhD, Duke University Medical Center, Box 2969, Durham, NC 27710. Email: brummett{at}acpub.duke.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
METHODS: Social isolation was examined as a predictor of mortality in 430 patients with significant CAD. More isolated patients were compared with their less isolated counterparts on factors that might help explain the association between isolation and survival.
RESULTS: The mortality rate was higher among isolated individuals. Those with three or fewer people in their social support network had a relative risk of 2.43 (p = .001) for cardiac mortality and 2.11 (p = .001) for all-cause mortality, controlling for age and disease severity. Adjustments for income, hostility, and smoking status did not alter the risk due to social isolation. With the exception of lower income, higher hostility ratings, and higher smoking rates, isolated patients did not differ from nonisolated patients on demographic indicators, disease severity, physical functioning, or psychological distress. Isolated patients reported less social support and were less pleased with the way they got along with network members, but they did not report less satisfaction with the amount of social contact received.
CONCLUSIONS: Patients with small social networks had an elevated risk of mortality, but this greater risk was not attributable to confounding with disease severity, demographics, or psychological distress. These findings have implications for mechanisms linking social isolation to mortality and for the application of psychosocial interventions.
Key Words: social support networks • cardiac mortality.
Abbreviations: CAD = coronary artery disease; CES-D = Center for Epidemiologic Studies Depression Scale; CI = confidence interval; DASI = Duke Activity Status Index; ISEL = Interpersonal Support Evaluation List; MOSS = Mediators of Social Support; PSS = Perceived Stress Scale; RR = relative risk; SF-36 = SF-36 Health Survey.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
Certain studies in community samples have found a graded effect for the relation between network support and health, whereas other studies have found that only those individuals who are most isolated are at increased risk for mortality (17). Understanding the functional form (eg, linear vs. threshold effect) of the association between mortality and network support is important, in part because such knowledge may be useful when designing potential interventions. The form of the association between network size and mortality has not been adequately examined in patient samples because many of the studies gathered only dichotomous support measures.
The purpose of the study reported here was to examine the association between social network support and mortality in a sample of patients with CAD who were well characterized with respect to demographic and psychosocial factors. Additional analyses explored 1) the form of the association between network support and mortality; and 2) the clinical, demographic, and psychosocial characteristics of isolated patients.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
75% stenosis of at least 1 coronary artery). Patients were excluded if they had prior angioplasty, congenital heart disease, primary valvular heart disease, substance dependence, history of an impairing psychological disorder, or an inability to give informed consent. Informed consent was obtained from all qualified patients who agreed to participate. Only patients initially treated medically for at least 90 days after enrollment are included in the analyses presented here. These patients underwent an extensive interview designed to assess the characteristics of their social support network; the interview was administered at the time of their admission for catheterization. Of the 433 participants interviewed, 3 had missing data on one or more of the medical control variables, resulting in a final sample of 430 with complete data. Patients were prospectively followed up at 3 and 6 months, 1 year, and annually thereafter. Follow-up for surviving participants lasted up to 76.3 months, with a mean of 47.3 months. There were 159 deaths; 120 were classified as cardiac deaths by an independent committee.
Measures
Network social support.
The Mannheim Social Support Interview (18) was used to provide a comprehensive assessment of network social support. Initially patients were asked to list the names of people who 1) they like to talk to and do things with; 2) have invited them out during the last 4 weeks; 3) they can borrow things from and ask favors of; 4) could help them make a very important personal decision, such as taking out a big loan, deciding to move, or choosing a new car; 5) they could talk with and who would understand them if a very close friend or relative was extremely ill or had just died; and 6) would talk to them, encourage them, and make them feel better about themselves when everything goes wrong and they feel down and discouraged and even doubt their own worth. The total number of distinct individuals mentioned in response to the above questions yielded the primary measure of network size. In addition, participants were asked how often they visit with each network member per month and whether they were happy with the way they got along with that person. Five additional measures were derived from these responses: Psychological Everyday Support summarizes the number of individuals who were listed for items 1 and 2 above who met the requirement of visiting with the participant at least once a month; Psychological Crisis Support reflects the total number of individuals listed for items 5 and 6 above; Instrumental Everyday Support is the sum of individuals who were listed for item 3 above who visited with the participant at least once a month; Instrumental Crisis Support indicates the number of individuals listed for item 4 above; and Relationship Satisfaction indicates the proportion of network members with whom the patient reported being happy with the way they got along.
Patients were also asked if they would like 1) more people to talk to and do things with, 2) to be invited out more, 3) more people who could help them make important decisions, and 4) more people to talk to about their feelings when things are going wrong. These items were summed to provide an index of Network Adequacy, with higher scores indicating greater perceived adequacy.
Confidants and living alone.
As in previous studies (5, 19), patients were asked how many people lived in their household and whether they had someone in whom they could confide.
Perceptions of social support.
The ISEL (20) was used to measure perceptions of the availability of social support. The full ISEL consists of 40 items. However, because of the large number of questions asked in the MOSS assessment battery, a 16-item abbreviated version (7) of the ISEL was used to limit the burden on the patient. Higher scores indicate increased availability of support.
Participation in religious activities.
The following questions were used to assess participation in religious activities: 1) How often do you attend worship services at a church/synagogue? and 2) How often do you participate in other religious activities, such as choir practice, committee or board meetings, social functions, etc? Responses were coded as the number of times per year patients reported attending or participating. Frequent attenders were defined as those who reported attending worship services once a week or more (41.4% of the sample). Frequent participators were defined as those who reported participating in other religious activities once a month or more (26% of the sample).
Depression.
The CES-D (21) is a 20-item self-report scale designed to assess depressive symptomatology in a general population. The questions refer to symptoms experienced during the previous week. The items were scored on a four-point scale, with the total score ranging from 0 to 60. Higher scores represent depressive responses, and a score of 16 or greater is generally considered to be suggestive of clinical depression.
SF-36 Health Survey: Functional status, health perceptions, and mental health.
The SF-36 Health Survey (22) contains eight subscales designed to measure quality of life in the domains of physical, mental, and social functioning. The physical functioning (10 items), bodily pain (2 items), and vitality (4 items) subscales were used to represent functional health. A composite measure of functional status was formed by computing factor scores from a principal components analysis of the three scales. That analysis indicated a one-factor solution that accounted for 68% of the variance.
The SF-36 subscale designed to assess mental functioning was used separately as a measure of mental health. In addition, the SF-36 also contains a question that assesses an individual’s perception of their general health, rated on a five-point scale ranging from excellent to poor. Raw scores for all SF-36 subscales were standardized to range from 0 to 100, with higher scores indicating better functioning.
Activities of daily living.
Activities of daily living were measured with the 12-item DASI (23). Higher scores reflect increased functional capacity over a 1-month period. Items represent personal care, ambulation, household tasks, sexual function, and recreation. The DASI has adequate internal consistency (
= 0.86) and has been shown to correlate significantly with clinical and angiographic end points (23, 24).
Perceived stress.
The PSS (25) assesses the degree to which individuals feel that events in their lives are unpredictable and uncontrollable. The 10-item version of the PSS was used in this study (26).
Hostility.
The Cook-Medley Hostility Scale (27) has been used extensively in health psychology and is generally thought to reflect primarily cynicism and mistrust (28). The 50-item scale has been shown to predict all-cause mortality and coronary events (29–31) as well as a number of other health outcomes (32). A rational analysis of item content (33) has led to the refinement of the full scale down to 27 items containing three subsets (cynicism, hostile affect, and aggressive responding) that have been shown to be better predictors of health outcomes in some studies (33, 34). The 27-item version was used in this study.
Income.
Participants indicated their household income by choosing one of the following income ranges: 1) $10,000 or less;, 2) $10,001 to $20,000; 3) $20,001 to $30,000; 4) $30,001 to $45,000; 5) $45,001 to 60,000; or 6) $60,001 or greater. This categorical measure of income was rank ordered so that it could be treated as an interval variable.
Disease severity.
The following measures were used to control for disease severity in all analyses: number of diseased vessels, left ventricular ejection fraction, presence or absence of congestive heart failure, age, and comorbidity. A comorbidity index was constructed from a count of the number of organ systems with major comorbidity. This information was abstracted from patient medical records.
Smoking and exercise status.
Patients were categorized as smokers at baseline if they reported smoking 10 or more cigarettes per day in the past 2 weeks. Patients were classified as actively participating in physical activity if they reported "exercising hard enough to work up a sweat" two or more times per week.
Data Analysis
Cox proportional hazard models were used to estimate the relation between the number of network members and survival. Treatment status was controlled as a time-dependent covariate. To describe the form of the relationship, the network size variable was divided into groups containing 10% to 20% of the total sample and compared in a categorical model. The most isolated group, patients who reported three or fewer network members, served as the reference category. There were too few people with only one or two network members to make meaningful comparisons within that category. Likewise, there were too few people with more than nine network members to make adequate comparisons above that level. Death due to cardiac disease was the outcome of interest for the analyses reported here. However, results of analyses for all-cause mortality were also examined.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
2(1) = 11.8, 95% CI = 1.52–3.89, p = .0006) for individuals reporting three or fewer network members. Addition of a continuous variable representing the total number of network members did not add predictive information beyond that provided by the dichotomous variable (2(1) = 0.00). Analyses for all-cause mortality yielded similar results, with a RR due to social isolation of 2.11, (2(1) = 10.9, 95% CI = 1.39–3.19, p = .0009). Additional tests evaluating the interactions of social isolation with gender were nonsignificant for either outcome (2 values < 0.5).
|
|
Demographic and medical characteristics of isolated patients.
Of the sociodemographic factors examined, only income and smoking status at baseline discriminated isolated patients from the others ( Table 2). Although there was a trend for income to predict survival in this sample (p = .09), inclusion of income as a covariate did not weaken the RR due to isolation for cardiac (RR = 2.84) or all-cause (RR = 2.47) mortality. Furthermore, smoking status did not predict survival in this sample (p = .71), nor did it significantly reduce the RR due to isolation for cardiac (RR = 2.64) or all-cause (RR = 2.21) mortality.
|
2(4) = 3.73, p = .44), indicating that isolated patients did not have more severe disease with respect to these medical indicators. Furthermore, measures of CAD severity were included in the survival models, demonstrating that the effects of social isolation on mortality were independent of the extent of illness. Finally, the absence of group differences on measures of functional status and health perceptions lends further support for the notion that the import of social isolation is independent of disease severity (Table 2).
Psychological characteristics of isolated patients.
The psychological characteristics of isolated and nonisolated patients are listed in Table 2. Isolated patients did not report more depression or stress, nor did they report poorer mental health on the SF-36 scale. The only psychological indicator that appeared to be related to isolation was hostility, with isolated patients having higher hostility ratings (p = .05). However, hostility did not predict survival in this sample, and the inclusion of hostility as a covariate did not substantially alter the RR due to isolation for either cardiac (RR = 2.47) or all-cause (RR = 2.36) mortality.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
The patients who were classified as socially isolated (according to the number of network members reported) were also significantly more likely to be unmarried, to lack a confidant, to have fewer social visits, to attend fewer social types of religious functions, and to have insufficient psychological and instrumental support. The interview procedure used in this study provides a comprehensive measure of network support that might continue to be useful in the understanding of socially isolated patients.
A variety of explanations have been offered to account for the association between social isolation and mortality. One hypothesis is that isolated patients may be sicker and may delay seeking medical care. Alternatively, individuals who are sicker may become more isolated. These propositions were not supported because measures of disease severity did not discriminate between isolated and nonisolated patients. Another hypothesis is that socioeconomic status might account for the phenomenon because it is a predictor of mortality that may be confounded with social isolation. Yet controlling for income did not alter the risk due to social isolation in these data. Another possible mediator is psychological distress, but this hypothesis was not supported because there was no evidence that isolated patients in this study experienced excessive distress. Finally, although isolated patients reported more hostility, controlling for hostility left the isolation effect unchanged. Thus, none of the explanations we examined seem to account for the elevated risk of the socially isolated patients.
Despite the limited social behavior reported by the isolated patients, they did not differ from other patients in their reported levels of psychological distress or their desire for more social contacts. This may pose a barrier to psychosocial interventions designed to reduce the mortality risk of isolated patients. These individuals may be difficult to enroll in interventions because they do not seem to feel that they have a problem. Without the experience of need, motivation to change may be low. Previous investigators have had difficulty retaining such patients in psychosocial intervention programs (35). Recent research in the area of health behavior modification suggests that a technique referred to as motivation-based counseling may be beneficial for patients such as those in our study (36, 37). Motivational interviewing encourages the clinician to share information that might help the client, but the technique relies on the fact that decisions to change rest solely with the client. It is possible, however, that social support interventions may not benefit those who feel they do not need them and could potentially have adverse consequences.
There are several limitations of the present findings. The distress assessments were based on self-report measures, and clinical ratings may have yielded different results. In addition, the sample was composed exclusively of patients with CAD, and our results may not generalize to other patient samples and cannot be applied to questions about disease onset.
Despite the failure of these data to clearly point to a likely mechanism, they again demonstrate the adverse effects of social isolation. Other explanations for this phenomenon should be investigated. Perhaps the lack of contact with other people leads to failure to get necessary medical attention or lower access to medical care. In addition, socially isolated individuals, especially those who are hostile, may trust their physician less or may be less likely to adhere to medical regimens. More detailed characterizations of socially isolated patients are needed to evaluate other potential explanations for this phenomenon.
It is desirable to measure multiple components when attempting to assess a complex construct such as social support. However, in the present study a summary measure indicating whether patients had three or fewer individuals to provide support was associated with many dimensions of social support and was sufficient to identify individuals at increased risk. Additional prospective research will be necessary to confirm whether such a measure could be considered a core component of a measurement battery. If so, it could be the basis of a future social support measurement strategy.
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
Received for publication March 23, 2000.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  E. Y. Cornwell and L. J. Waite Measuring Social Isolation Among Older Adults Using Multiple Indicators From the NSHAP Study J Gerontol B Psychol Sci Soc Sci, June 9, 2009; (2009) gbp037v1. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Sabir, E. Wethington, R. Breckman, R. Meador, M.C. Reid, and K. Pillemer A Community-Based Participatory Critique of Social Isolation Intervention Research for Community-Dwelling Older Adults Journal of Applied Gerontology, April 1, 2009; 28(2): 218 - 234. [Abstract] [PDF]
|
|
|
|
 |
|
 R. E. Petrea, A. S. Beiser, S. Seshadri, M. Kelly-Hayes, C. S. Kase, and P. A. Wolf Gender Differences in Stroke Incidence and Poststroke Disability in the Framingham Heart Study Stroke, April 1, 2009; 40(4): 1032 - 1037. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Vicki Kafanelis, M. Kostanski, P. A. Komesaroff, and L. Stojanovska Being in the Script of Menopause: Mapping the Complexities of Coping Strategies Qual Health Res, January 1, 2009; 19(1): 30 - 41. [Abstract] [PDF]
|
|
|
|
 |
|
 E. Calvo, K. Haverstick, and S. A. Sass Gradual Retirement, Sense of Control, and Retirees' Happiness Research on Aging, January 1, 2009; 31(1): 112 - 135. [Abstract] [PDF]
|
|
|
|
 |
|
 S. K. Hull, L. F. DiLalla, and J. K. Dorsey Prevalence of Health-Related Behaviors Among Physicians and Medical Trainees Acad Psychiatry, February 1, 2008; 32(1): 31 - 38. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. A Schiffer, J. Denollet, J. W Widdershoven, E. H Hendriks, and O. R F Smith Failure to consult for symptoms of heart failure in patients with a type-D personality Heart, July 1, 2007; 93(7): 814 - 818. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. E. Williams, M. I. Chimowitz, G. A. Cotsonis, M. J. Lynn, S. P. Waddy, and for the WASID Investigators Gender Differences in Outcomes Among Patients With Symptomatic Intracranial Arterial Stenosis Stroke, July 1, 2007; 38(7): 2055 - 2062. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. S. Y. Thong, A. A. Kaptein, R. T. Krediet, E. W. Boeschoten, F. W. Dekker, and for the Netherlands Cooperative Study on the Adequ Social support predicts survival in dialysis patients Nephrol. Dial. Transplant., March 1, 2007; 22(3): 845 - 850. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. S. Lett, J. A. Blumenthal, M. A. Babyak, T. J. Strauman, C. Robins, and A. Sherwood Social Support and Coronary Heart Disease: Epidemiologic Evidence and Implications for Treatment Psychosom Med, November 1, 2005; 67(6): 869 - 878. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. M. Burg, J. Barefoot, L. Berkman, D. J. Catellier, S. Czajkowski, P. Saab, M. Huber, V. DeLillo, P. Mitchell, J. Skala, et al. Low Perceived Social Support and Post-Myocardial Infarction Prognosis in the Enhancing Recovery in Coronary Heart Disease Clinical Trial: The Effects of Treatment Psychosom Med, November 1, 2005; 67(6): 879 - 888. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S Vale Psychosocial stress and cardiovascular diseases Postgrad. Med. J., July 1, 2005; 81(957): 429 - 435. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Boden-Albala, E. Litwak, M.S.V. Elkind, T. Rundek, and R. L. Sacco Social isolation and outcomes post stroke Neurology, June 14, 2005; 64(11): 1888 - 1892. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. S. Knox, G. Weidner, A. Adelman, C. M. Stoney, R. C. Ellison, and for the Investigators of the National Heart, Lung, Hostility and Physiological Risk in the National Heart, Lung, and Blood Institute Family Heart Study Arch Intern Med, December 13, 2004; 164(22): 2442 - 2447. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Rutledge, S. E. Reis, M. Olson, J. Owens, S. F. Kelsey, C. J. Pepine, S. Mankad, W. J. Rogers, C. N. Bairey Merz, G. Sopko, et al. Social Networks Are Associated With Lower Mortality Rates Among Women With Suspected Coronary Disease: The National Heart, Lung, and Blood Institute-Sponsored Women's Ischemia Syndrome Evaluation Study Psychosom Med, November 1, 2004; 66(6): 882 - 888. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Mookadam and H. M. Arthur Social Support and Its Relationship to Morbidity and Mortality After Acute Myocardial Infarction: Systematic Overview Arch Intern Med, July 26, 2004; 164(14): 1514 - 1518. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Johnston, J. R. Goss, J. A. Malmgren, and J. A. Spertus Health status and social risk correlates of extended length of stay following coronary artery bypass surgery Ann. Thorac. Surg., February 1, 2004; 77(2): 557 - 562. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. Phillips Bute, J. Mathew, J. A. Blumenthal, K. Welsh-Bohmer, W. D. White, D. Mark, K. Landolfo, and M. F. Newman Female Gender Is Associated With Impaired Quality of Life 1 Year After Coronary Artery Bypass Surgery Psychosom Med, November 1, 2003; 65(6): 944 - 951. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. M. Leyden Social Capital and the Built Environment: The Importance of Walkable Neighborhoods Am J Public Health, September 1, 2003; 93(9): 1546 - 1551. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. S. House Social Isolation Kills, But How and Why? Psychosom Med, March 1, 2001; 63(2): 273 - 274. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
