Psychosocial heterogeneity among hospitalized patients with chronic diseases without a primary caregiver: a multicenter cross-sectional latent profile analysis of anxiety, depression, sleep quality, and perceived social support

Abstract Background Hospitalized patients with chronic diseases who lack a primary caregiver may show heterogeneous psychosocial configurations across emotional symptoms, sleep, and perceived support. This study identified latent profiles of anxiety, depression, sleep quality, and perceived social support and examined demographic and clinical correlates of profile membership. Methods A multicenter cross-sectional study was conducted at four healthcare institutions in Hubei Province, China, from February to June 2024. The analytic sample included 1,830 hospitalized adults with chronic diseases and no primary caregiver. Latent profile analysis used standardized Patient Health Questionnaire-9, Generalized Anxiety Disorder-7, Pittsburgh Sleep Quality Index, and reverse-standardized Multidimensional Scale of Perceived Social Support scores. Model selection considered information criteria, classification quality, profile size, stability, parsimony, and interpretability. Associations with profile membership were examined in 1,807 participants using multinomial logistic regression pooled across 200 posterior-probability-based pseudo-class draws, with Benjamini-Hochberg correction for multiple omnibus tests. Results A three-profile solution was retained (entropy = 0.891; mean maximum posterior probability = 0.953): relatively low anxiety and depressive symptoms (n = 1,351; 73.83%), depressive symptoms-sleep disturbance (n = 96; 5.25%), and high anxiety and depressive symptoms (n = 383; 20.93%). Profile differences were driven mainly by anxiety, depressive, and sleep symptoms; perceived social support was similar across profiles and was not a major differentiating dimension. Educational attainment showed a nominal association before multiple-testing correction, but no demographic or clinical factor remained stably associated with profile membership after correction. Sensitivity analyses supported the stability of the two larger profiles, whereas the depressive symptoms-sleep disturbance profile was more sensitive to model specification and sample perturbation. Conclusions Hospitalized patients with chronic diseases and no primary caregiver exhibited distinct psychosocial configurations characterized primarily by different combinations of emotional and sleep symptoms. Multidimensional assessment may be more informative than relying on routine demographic or clinical characteristics alone. The smaller depressive symptoms-sleep disturbance profile requires external validation before use for clinical stratification.