Prognostication of Mental Health Risk Clusters on Hospitalization and Mortality in Patients With Coexisting Diabetes and Kidney Failure: The Hidden Burden of Loneliness.

Rationale & Objective: Individuals with comorbid diabetes and kidney failure have poor clinical prognosis, often aggravated by psychological distress. Identifying individuals most at risk is crucial to improving service provision. This study aimed to identify psychosocial profiles in patients with diabetes and kidney failure, model their prognostic effects on hospitalization and mortality, and explore underlying mechanisms linking psychosocial health to clinical outcomes. Study Design: Prospective cohort study. Setting & Participants: A total of 221 participants with coexisting diabetes and kidney failure (median age: 59 years, 60.6% men) receiving hemodialysis were recruited from the National Kidney Foundation Singapore's dialysis centers. Exposures: Depression, anxiety, loneliness, and hopelessness alongside self-care indicators were measured using validated self-reported scales. Outcomes: All-cause hospitalization and mortality were ascertained from medical records. Analytical Approach: Latent profile analysis was used to identify psychosocial profiles. Associations of sociodemographic, clinical factors and psychosocial profiles with clinical endpoints were modeled with Negative binomial and Cox regressions (mean = 21.8 months). Casual mediation analyses modeled self-care as mediator. Results: Three psychosocial profiles emerged: resilient (37.6%; all below cutoffs), overwhelmed (30.3%; above cutoffs), and lonely (32.1%; above cutoff for loneliness only). The lonely group was more socioeconomically disadvantaged relative to the resilient group. The lonely and overwhelmed groups had increased hospitalization rates and more hospitalization days than the resilient group (incident risk ratio [IRR] range, 1.50-1.82; P < 0.05). No association with mortality was found. Better diabetes self-care and nutrition quality-of-life also predicted hospitalization (IRR range, 0.94-0.97; P < 0.05) and mortality (hazard ratio [HR] = 0.93 and 0.96). Mediation analysis indicated that diabetes self-care activities accounted for 18% of the associations between the lonely profile and hospitalization days. Limitations: Geographic generalizability of participants and sample size. Conclusions: Interconnected psychosocial burdens significantly affect disease management and hospitalization risk in patients with diabetes and kidney failure. Integrating psychosocial screening and interventions into clinical practice, particularly addressing loneliness and not just depression and anxiety, may be crucial.