Self-neglect among hospitalized older adults based on the health ecological model: an explanatory sequential mixed-methods study

Abstract Background Research on self-neglect has focused on community-dwelling populations, with hospitalized older adults understudied. Existing studies rely on quantitative surveys, rarely using qualitative methods. Aim This study aimed to examine the prevalence and multilevel correlates of self-neglect among hospitalized older adults, and to explain these patterns through participants’ accounts, using the health ecological model (HEM). Methods We used an explanatory sequential mixed-methods design. Phase 1 was a cross-sectional survey of 219 hospitalized older adults recruited from a tertiary hospital in China, using the Scale of Elder Self-Neglect–Rural (SESNr), Patient Health Questionnaire-9 (PHQ-9), Activities of Daily Living scale (ADL), Tilburg Frailty Indicator (TFI), and Social Support Rating Scale (SSRS). Phase 2 involved semi-structured interviews with 16 participants with an SESNr score ≥ 1, purposively sampled by maximum variation and analyzed using conventional (inductive) content analysis. Findings were integrated in a joint display. Results Applying SESNr criteria, 16.0% of participants were classified as having self-neglect, 73.1% possible self-neglect, and 10.9% no evidence of self-neglect. Depressive symptoms (IRR = 1.10, 95% CI 1.06–1.14), social support (IRR = 0.92, 95% CI 0.90–0.95), male sex (IRR = 1.30, 95% CI 1.05–1.60), ADL function (IRR = 1.01, 95% CI 1.00–1.02), number of admission diagnoses (IRR = 1.04, 95% CI 1.00–1.09), and age (IRR = 0.98, 95% CI 0.96–1.00) were independently associated with self-neglect severity. Education, monthly income, and insurance type were also independently associated (Wald χ² = 8.46, df = 3, P = .037; χ² = 7.04, df = 2, P = .030; χ² = 7.18, df = 2, P = .028, respectively). The qualitative findings identified four themes: limited health literacy, declining physical and cognitive capacity, low sense of self-worth, and limited family and social support. Depressive symptoms converged with the themes of declining capacity and low self-worth, while limited health literacy and cost-related postponement of care emerged only qualitatively, expanding the model beyond the survey. Conclusions HEM-guided, nurse-led multilevel strategies — combining depression and functional screening, health-literacy support, and assessment of family and financial barriers—warrant development and evaluation in future intervention studies.