Social network composition of vascular patients and its associations with health behavior and clinical risk factors

Background This study aimed to explore linkages of patients' social network composition with health behaviors and clinical risk factors. Methods/Design This observational study was embedded in a project aimed at improving cardiovascular risk management (CRVM) in primary care. 657 vascular patients (227 with cardiovascular disease, 380 at high vascular risk), mean age 72.4 (SD 9.4) years, were recruited as were individuals patients considered important for dealing with their disease, so called alters (n = 487). Network composition was measured with structured patient questionnaires. Both patients and alters completed questionnaires to measure health behavior (habits for physical activity, diet, and smoking). Clinical risk factors (systolic blood pressure, LDL cholesterol level, and body mass index) were extracted from patients' medical records. Six logistic regression analyses, using generalized estimating equations, were used to test three hypothesized effects of network composition (having alters with healthful behaviors, without depression, and with specialized knowledge) on six outcomes, adjusted for demographic, personal and psychological characteristics. Results Having alters with overall healthful behavior was related to healthful patient diet (OR 2.14, 95% CI: 1.52-3.02). Having non-smoking alters in networks was related to reduced odds for patient smoking (OR 0.17, 95% CI: 0.05-0.60). No effects of presence of non-depressed alters were found. Presence of alters with specialized knowledge on CVRM was inversely related to healthful diet habits of patients (OR 0.47, 95% CI 0.24-0.89). No significant associations between social network composition and clinical risk factors were found. Discussion Diet and smoking, but not physical exercise and clinical risk factors, were associated with social network composition of patients with vascular conditions. In this study of vascular patients, controlling for both personal and psychological factors, fewer network influences were found compared to previous research. Further research is needed to examine network structure characteristics as well as the role of psychological factors to enhance understanding health behavior of patients involved in CVRM.