Application of cox proportional hazard model to study the influence of determinants on under five mortality in Odisha, India

This study investigates under-five mortality (U5M) dynamics in Odisha using National Family ‎Health Survey (NFHS) data through five models incorporating socio-economic, environmental, ‎demographic, nutritional, and media variables. Employing the Cox Proportional Hazard model ‎facilitates time-to-event analysis, accommodating censored observations and handling multiple ‎covariates. The model's assumption of proportional hazard ratios over time suits the study of ‎U5M, a time-dependent event. Table 1 presents hazard ratios from NFHS-4 to NFHS-5, ‎revealing evolving trends in U5M risk factors. Significant contributors include maternal ‎education, with hazard ratios decreasing across education levels, indicating improved ‎associations with U5M in NFHS-5. Rural areas exhibit consistently higher hazard ratios, while ‎wealthier households show potential rises in U5M risk. Varied impacts on U5M risk emerge for ‎birth-related factors. These findings underscore the changing landscape of U5M risk factors in ‎Odisha, offering a foundation for evidence-based policymaking to address evolving socio-‎demographic dynamics impacting child survival in the region‎.