Abstract
BACKGROUND: Falls among hospitalized older adults represent a critical patient safety concern, yet comprehensive assessments of fall prevention knowledge, attitudes, behaviors (KAB), education quality, and health literacy in Chinese hospital settings remain scarce. This study examined these domains and their relationships with fall outcomes. METHODS: This cross-sectional study enrolled patient-caregiver dyads at a tertiary hospital in China between February 2023 and October 2025. Participants completed validated assessments measuring fall prevention knowledge (18-item scale), attitudes (9-item scale), behaviors (14-item scale), education quality metrics (cascade framework with teach-back assessment), health literacy (composite score), and communication barriers. Fall/near-fall events during hospitalization were ascertained through structured interviews. Progressive multivariable logistic regression models examined associations between KAB domains, education quality, health literacy, and fall outcomes. RESULTS: Among 3,223 participants, 950 (29.5%) reported experiencing at least one fall or near-fall event. Fall prevention knowledge was modest (mean accuracy 58-59%), with critical gaps in recognizing the risks of prior falls (29.6%) and the value of environmental modifications (29.3%). Although attitudes were positive and comparable between groups, caregivers reported significantly higher behavioral engagement than patients (48.80 vs. 45.10; p < 0.001); however, adherence to healthcare provider recommendations was notably low across both groups (mean 2.12 ± 1.14/5). The education cascade revealed substantial system attrition: while 69.8% of respondents were informed of fall risks, only 46.6% demonstrated verified comprehension via teach-back. Health literacy was frequently inadequate (41.5% scored ≤2/5), and KAB domains showed negligible intercorrelations (r ≤ 0.02). In adjusted multivariable models, KAB scores did not predict fall outcomes. Instead, higher health literacy (aOR = 1.11, 95% CI 1.03-1.19; p = 0.006) and greater comfort asking staff questions (aOR = 1.17, 95% CI 1.07-1.27; p < 0.001) were independently associated with increased probability of reporting events, despite low overall model discrimination (AUC = 0.577). CONCLUSION: Hospital fall prevention requires system-level interventions beyond patient education, including standardized comprehension verification, literacy-sensitive communication, and integration of clinical risk assessment with environmental safety protocols to address multifactorial determinants inadequately captured by KAB-centered frameworks.