Sociodemographic factors shaping knowledge-attitude-behaviour pathways in postpartum fall prevention

Postpartum falls threaten maternal and infant safety, yet evidence on postpartum women's fall-prevention knowledge, attitude, and behaviour (KAB) and their sociodemographic correlates remains limited. This study examined postpartum fall-prevention KAB and associations among KAB dimensions and in-hospital fall occurrence.

In this cross-sectional study, 1,220 postpartum women from a tertiary hospital in Chengdu, China, completed a general questionnaire and a self-developed Postpartum Fall Prevention KAB Scale (Cronbach's α = 0.926). Multivariable linear regression examined associations of maternal education, companion education, and residence with KAB scores. Structural equation modelling examined associations among knowledge, attitude, behaviour, and in-hospital fall occurrence.

Higher maternal education was associated with higher knowledge (B = 0.22, 95% CI 0.06 to 0.38; p = 0.009), attitude (B = 0.23, 95% CI 0.02 to 0.43; p = 0.030), and behaviour scores (B = 0.57, 95% CI 0.02 to 1.11; p = 0.044). Residence was associated with attitude score (B = 0.77, 95% CI 0.13 to 1.41; p = 0.020), but not with knowledge or behaviour. Companion education was associated with behaviour score (B = -0.67, 95% CI -1.15 to -0.20; p = 0.005), but not with knowledge or attitude. In the structural model, knowledge was associated with attitude (estimate = 0.289, 95% CI 0.206 to 0.379; p < 0.001) and behaviour (estimate = 0.663, 95% CI 0.455 to 0.908; p < 0.001), and attitude was associated with behaviour (estimate = 0.471, 95% CI 0.304 to 0.634; p < 0.001). behaviour was not associated with in-hospital fall occurrence (estimate = 0.001, 95% CI -0.001 to 0.002; p = 0.506).

Maternal education, residence, and companion education were associated with postpartum fall-prevention KAB scores. The findings support associations broadly consistent with the KAB framework, but causality cannot be inferred. Tailored fall-prevention education should be evaluated in longitudinal, multi-centre studies.