TITLE:
Identifying Indicators of Infant Mortality Using Survival Analysis
AUTHORS:
Nasrin Khatun
KEYWORDS:
Censoring Data, Survival Analysis, Cox Proportional Hazards Model
JOURNAL NAME:
Open Journal of Statistics,
Vol.15 No.4,
August
4,
2025
ABSTRACT: In practice, when dealing with censored data, survival analysis must be employed. In this case, parametric and non-parametric models are appropriate to analyze survival data to obtain optimal estimates of the parameters of interest. To identify significant determinants of infant mortality in rural Bangladesh, survival data have been extracted from the Bangladesh Demographic and Health Survey (BDHS), 2017-2018. In this study, an event involving an infant death within the past 12 months; otherwise, 0 will be used for censoring purposes. The main aim of this study is to find out the relationship between infant death and demographic factors. Used the Cox proportional hazard model (COXPH) to determine the responsible factors and found that age group, religion, and residence area significantly affected mortality. This study found that urban areas had a higher survival rate than rural areas. On the other hand, the age group 20 - 34 has a higher survival probability than other groups. And also, the “Others” have a higher mortality rate than the Muslim religion. Notably, background factors are effective on health facilities which help to increase the survival rate.