Abstract
This study examined cross-sectional (Ising) and longitudinal (cross-lagged panel) networks of depressive symptoms and health risk behaviours in people with diabetes, compared to those without. Participants (N = 22,510) from the Survey of Health, Ageing and Retirement in Europe with (n = 1781) and without (n = 20,729) diabetes with full data on smoking, high alcohol consumption, low fruit and vegetable consumption, physical inactivity and depressive symptoms (EURO-D) were included. Network models indicated that physical inactivity was by far the most closely connected behaviour to depressive symptoms for both groups. Networks were more densely connected for people without diabetes. Differences in whether sleep problems and physical inactivity activated or were activated by depressive symptoms were observed between people with and without diabetes. Identifying structural, symptom-level differences between health risk behaviours and depressive symptoms associated with the presence of diabetes holds potential for informing precision medicine.