Construction of an Interpretive Structural Model-Analytic Network Process model for preschool children's drowning risk factors based on the Haddon matrix.

Objective: This study aims to systematically analyze the complex mechanisms of risk factors for drowning in preschool children, clarify their structural relationships and relative importance, and provide scientific evidence for constructing a precise and effective drowning prevention and control system for preschool children. Methods: Based on the theoretical framework of the Haddon matrix (which aims to systematically identify and categorize risk factors from the four dimensions of host (child), agent (water), physical environment, and social environment), the Interpretive Structural Model (ISM) and Analytic Network Process (ANP) were integrated. First, ISM was used to hierarchically deconstruct 18 risk factors, revealing their hierarchical transmission paths; then, ANP was utilized to quantify the global weights of each factor in the network and determine their relative impact levels. Results: (1) ISM analysis indicated that the risk factors presented a five-layer hierarchical structure: the surface layer included direct factors such as children's risky behaviors (F4) and lack of immediate protection (F5); the middle layer centered around caregiver behaviors and supervision (F13); the bottom layer consisted of fundamental factors such as community safety systems (F15) and family economic background (F16). (2) Three key transmission chains were identified: social ecological pathways, direct pathways of the physical environment, and emergency response pathways. (3) ANP weight analysis showed that risky behaviors (28.06%) and caregiver supervision (23.26%) were the two core factors, with their combined weight exceeding 50%, followed by immediate response and lack of protection (14.69%). Conclusion: This study demonstrates that the drowning risk for preschool children stems from a complex, multi-factorial system transmission. Based on the principle of "structural layering and priority weighting" derived from our ISM-ANP model, we propose a three-tiered prevention and control strategy framework. This framework prioritizes short-term interventions targeting children's risky behaviors and caregiver supervision, medium-term reinforcement of physical and environmental barriers, and long-term optimization of the socio-ecological foundation. The findings underscore that effective prevention requires a systematic governance approach. The proposed layered strategy provides a theoretical basis and decision-making reference for constructing a targeted and precise drowning prevention and control system for preschool children.