Key points are not available for this paper at this time.
Background Dementia poses significant challenges for family caregivers, yet little is known about the developmental trajectories of family resilience and their predictors in this population. Understanding these patterns is essential for developing targeted interventions to support caregivers and enhance care quality. This study aims to identify and predict trajectories of family resilience among dementia caregivers using complementary logistic regression and decision tree approaches, and to explore key factors influencing these trajectories. Methods A three-wave longitudinal study was conducted among 239 dementia caregivers from 19 communities in Huzhou, China, between June 2023 and July 2024. Growth Mixture Modeling ( GMM ) was applied to trajectories. Binary logistic regression and decision tree models were employed to analyze predictors of these trajectories, and Receiver Operating Characteristic ( ROC ) curves were used to evaluate model performance. Results Two distinct family resilience trajectories were identified: a “low resilience-rapidly declining” group and a “high resilience-slowly rising” group. Key predictors included caregiver empowerment, self-efficacy, self-rated health, dementia knowledge, social support, and relationship with the patient. The logistic regression model demonstrated higher sensitivity, whereas the decision tree model showed higher specificity. The complementary use of both models enhanced predictive accuracy and interpretability. Conclusion Family resilience among Chinese dementia caregivers follows heterogeneous trajectories primarily predicted by caregiver empowerment and self-efficacy. The identified thresholds offer candidate criteria for early identification of high-risk families, pending external validation. These findings support the development of tiered, empowerment-focused interventions within community-based dementia care systems.
Lv et al. (Wed,) studied this question.