Aim <p>Loneliness and social isolation are risk factors for chronic diseases including dementia, but the pathways of influence are not yet clear.</p> Subject and methods <p>Levels of risk for nine modifiable health risk factors were calculated by applying Australian risk reduction guidelines to cohort study survey data. Serial multiple linear regression models were used to test associations by risk level, and a partial correlation network was used to map the directness of significant relationships.</p> Results <p>Regression models indicated that higher social isolation and loneliness were associated with high-risk status on seven of the nine risk factors (mean age 63&#xa0;years, <i>n</i> = 6670). However, the network analysis suggested that the size and supportiveness of social networks directly influenced rates of physical and cognitive activity, smoking rates, and dietary adherence, while loneliness only directly influenced smoking and higher body mass index (BMI). We found evidence suggesting an indirect influence on the management of cardiometabolic risk factors and alcohol consumption.</p> Conclusion <p>The direct associations identified may present potential intervention targets, while the indirect relationships suggest that other factors are influential in the pathway linking social isolation and loneliness with chronic disease and dementia. These results can be used to inform longitudinal studies to investigate underlying mechanisms in the relationships between social isolation, loneliness, and health risk behaviors.</p>