Effects of Social Networks On Medical Comorbidity Among People with Serious Mental Illness

Background and Purpose: People with serious mental illness (SMI) have higher rates of physical illness when compared to the general population. Among the general population, social networks have been found to be a significant predictor of physical health. However, little is known about the effects of social networks on medical comorbidity, which refers to the presence of physical illness in an individual with mental illness. Also, prior research has not looked at both the structural and transactional characteristics of social networks simultaneously in one study. Guided by Berkman's theoretical model on social network and health (Berkman et al., 2000), this study examined the extent to which both the network structure and network transaction characteristics affect medical comorbidity among people with SMI.

Methods: This study utilized interview data collected from a probability sample of 231 adults with SMI who resided in supported housing in Philadelphia, Pennsylvania. Pennsylvania Medicaid Claims and Managed Care data were matched with the survey data to obtain treated diagnoses of physical illnesses for the study sample. Two aspects of medical comorbidity, objective health and subjective health, were examined. Structured interviews provided data on socio-demographic and clinical characteristics, social network structure characteristics (i.e., size, composition, intensity of contact, and density), social network transaction characteristics (i.e., types of support rendered, perceived support, and reciprocity of support). To compare those with medical comorbidity and those without medical comorbidity, chi-square tests for categorical variables and independent samples t-tests for continuous variables were conducted. Path analyses were conducted for each of the following outcome variables: (1) Charlson Comorbidity Index (CCI) score (objective health), (2) the SF-12 PCS (subjective health), and (3) perceived health status (subjective health).

Results: The results from bivariate statistics showed that individuals with medical comorbidity were more likely to have contact with their network members and had a higher level of reciprocal positive tangible support when compared to those who did not have medical comorbidity. The results of the path analyses revealed that none of the social network characteristics were associated with medical comorbidity. A psychiatric diagnosis of schizophrenia and a high level of income were positively associated with objective health, but not with subjective health. By contrast, severity of psychiatric symptoms was negatively associated with the subjective health status of the individual, but not with the objective health status.

Conclusions and Implications: The findings from this study suggest that social network characteristics may not be relevant predictors of physical health among people with SMI. Considering the discrepancy found between objective health and subjective health, special attention is required to examine measurements on self-reported interview data and administrative data. The implications for future research are discussed.