A SURVIVAL ANALYSIS OF INFANT MORTALITY RATE: EVIDENCE FROM BARAU DIKKO TEACHING HOSPITAL, KADUNA, NIGERIA

Abstract

Infant mortality remains a major public health concern, particularly in low- and middle-income countries, where neonatal and infant deaths contribute substantially to under-five mortality. This study investigated the factors associated with infant mortality among infants admitted to Barau Dikko Teaching Hospital, Kaduna, Nigeria, using survival analysis techniques. A retrospective cohort design was adopted, utilising secondary data extracted from hospital records covering the period from 2015 to 2025. The Cox proportional hazards model was employed to examine the effects of birth weight, gestational age, five-minute Apgar score, maternal age, maternal educational attainment, and the number of antenatal care visits on infant survival. A total of 100 infants were included in the analysis, of whom 20 experienced the event of interest (death) during the follow-up period, while the remaining observations were right-censored. The findings revealed that gestational age was the only statistically significant predictor of infant mortality (HR = 0.755, 95% CI: 0.572–0.995, p = 0.046). Specifically, each additional week of gestation was associated with an approximately 24.5% reduction in the hazard of infant death. Although birth weight exhibited a protective effect (HR = 0.999, p= 0.081), its association with infant mortality was not statistically significant at the 5% significance level. Likewise, five-minute Apgar score, maternal age, maternal educational attainment, and the number of antenatal care visits did not significantly influence infant survival. The fitted Cox model demonstrated good predictive performance, with a concordance statistic of 0.874. However, assessment of the proportional hazards assumption using Schoenfeld residuals indicated significant violations for gestational age and the number of antenatal care visits, suggesting that the effects of these covariates varied over time. These findings indicate that extended Cox models incorporating time-dependent covariates may provide a more appropriate framework for future analyses. The study concludes that gestational age is the most important determinant of infant survival among the factors examined. The findings underscore the importance of interventions aimed at preventing preterm births, improving the quality and utilisation of antenatal care services, and strengthening neonatal healthcare to reduce infant mortality. The study further contributes hospital-based empirical evidence from Northern Nigeria, thereby enriching the existing literature on infant survival and providing useful information for healthcare practitioners and policymakers.