Associations of Psychosocial Factors with Multiple Health Behaviors: A Population-Based Study of Middle-Aged Men and Women.

Background: The health behaviors smoking, risky alcohol consumption, insufficient physical activity, and poor diet constitute the main contributors to non-communicable diseases. Clustering of risk behaviors is common and increases the risk of these diseases. Despite health benefits, it is difficult to change health behaviors. Psychosocial factors could play a role in health behavior change, with research showing relationships between unfavorable psychosocial factors and health risk behaviors. However, many studies only investigated one or two health behaviors at a time. The present study, therefore, aimed to investigate associations between a broad range of psychosocial factors and multiple health risk behaviors in a general middle-aged population in Sweden.
Methods: A cross-sectional design was used to investigate a random sample from the general population in Sweden (n = 1007, 45-69 years, 50% women). Questionnaire data on health behaviors (smoking, alcohol consumption, physical activity, and fruit/vegetable intake) and psychosocial factors, with both psychological and social resources (social integration, emotional support, perceived control, self-esteem, sense of coherence and trust) and psychological risk factors (cynicism, vital exhaustion, hopelessness and depressiveness), were analyzed. Logistic and ordinal logistic regression were used to analyze associations between psychosocial factors and multiple (0-1, 2 or 3-4) health risk behaviors.
Results: A total of 50% of the sample had two health risk behaviors and 18% had three health risk behaviors. After adjusting for age, sex, education, employment status, and immigrant status, eight out of 10 psychosocial factors (exceptions: social integration and self-esteem) showed significant odds ratios (ORs) in the expected directions; low levels of psychosocial resources and high levels of psychosocial risk factors were associated with multiple risk behaviors. The strongest associations with multiple risk behaviors were seen for vital exhaustion (adjusted (adj.) OR 1.28; confidence interval (CI) 1.11-1.46), depressiveness (adj. OR 1.32, CI 1.14-1.52), and trust (adj. OR 0.80, CI 0.70-0.91). When controlling for all psychosocial factors in the same model, only the association with trust remained statistically significant (adj. OR 0.89, CI 0.73-1.00, p = 0.050). Associations with individual health behaviors were fewer and scattered, with no psychosocial factor being related to all four behaviors. Conclusions: Examining associations between a broad range of psychosocial factors and multiple health risk behaviors revealed consistent and significant associations for almost all psychosocial factors. These associations were stronger compared to associations to single health risk behaviors. Our findings support the relevance of considering psychosocial aspects in interventions aimed at health behavior change, especially for people with multiple health risk behaviors.