Three latent profiles of proactive health behaviors in older adult stroke patients and implications for stratified nursing: a cross-sectional study based on Pender’s health promotion model

Stroke has become the leading cause of death and disability among Chinese adults. The long-term rehabilitation and secondary prevention of elderly patients are highly dependent on proactive health behaviors, yet existing studies largely overlook group heterogeneity. Based on Pender’s Health Promotion Model, this study used latent profile analysis to explore the latent classes of proactive health behaviors among elderly stroke patients and their influencing factors, so as to provide evidence for stratified precision nursing.

This study adopted a cross-sectional design, from July to December 2025, 403 elderly stroke patients were selected by convenience sampling from Hefei First People’s Hospital. The survey instruments included a general information questionnaire, the Active Health Behavior Scale for Patients with Chronic Diseases (AHBS), the Self-Efficacy for Managing Chronic Disease 6-Item Scale (SES6C), the Social Support Rating Scale (SSRS), and the Brief Illness Perception Questionnaire (BIPQ). Latent profile analysis (LPA) was used to identify the latent classes of proactive health behaviors, with AIC, BIC, aBIC, entropy, LMR, and BLRT used to determine the optimal model. Univariate analysis and multinomial logistic regression (with the low proactive health behavior-passive compliance type as reference) were used to explore the influencing factors.

The proactive health behavior score of the 403 patients was 59.42±12.56. LPA identified three latent classes: low proactive health behavior-passive compliance (53.6%), moderate proactive health behavior-cognitive-behavioral imbalance (20.8%), and high proactive health behavior-active engagement (25.6%). The model demonstrated a good fit (AIC=9467.672, BIC=9555.648, aBIC=9485.840, entropy=0.953, LMR and BLRT: P<0.001). Multinomial logistic regression analysis revealed that, with the low proactive health behavior-passive compliance class as the reference, monthly income of 1000 to 2999 yuan (OR=0.139, 95%CI: 0.029 to 0.659), monthly income of 3000 to 5000 yuan (OR=0.119, 95%CI: 0.023 to 0.605), chronic disease management self-efficacy score (OR=1.238, 95%CI: 1.141 to 1.343), illness perception score (OR=1.246, 95%CI: 1.181 to 1.314), and social support score (OR=1.337, 95%CI: 1.250 to 1.431) were associated with classification into the moderate proactive health behavior-cognitive-behavioral imbalance class. Furthermore, primary school education or below (OR=0.030, 95%CI: 0.008 to 0.118), junior or senior high school education (OR=0.305, 95%CI: 0.098 to 0.939), monthly income of 1000 to 2999 yuan (OR=0.063, 95%CI: 0.015 to 0.263), monthly income of 3000 to 5000 yuan (OR=0.219, 95%CI: 0.052 to 0.923), living with spouse (OR=20.693, 95%CI: 2.620 to 163.263), living with children (OR=8.693, 95%CI: 1.523 to 49.625), family history of stroke (OR=5.688, 95%CI: 1.875 to 17.259), having 3 to 5 chronic diseases (OR=4.998, 95%CI: 1.409 to 17.656), chronic disease management self-efficacy score (OR=1.107, 95%CI: 1.046 to 1.173), illness perception score (OR=1.209, 95%CI: 1.125 to 1.301), and social support score (OR=1.383, 95%CI: 1.286 to 1.489) were associated with classification into the high proactive health behavior-active engagement class (all P<0.05).

Elderly stroke patients showed obvious group heterogeneity in proactive health behaviors. Multiple sociodemographic and psychological factors influenced their profile types, and clinical practice should implement stratified individualized interventions based on latent profile characteristics.