Determinants of neonatal, infant and under-five mortalities: evidence from a developing country, Bangladesh

**Background:** Despite global progress, 5.3 million children under five died from preventable causes in 2018, nearly half in the first month of life. Bangladesh has outperformed India and Pakistan in reducing child mortality and achieved SDG 3 and MDG 4 targets ahead of schedule. This study aims to identify the macroeconomic determinants driving these reductions from 1991 to 2018 and discuss implications for other developing countries.

**Methods:** Annual time-series data were extracted from the World Bank Open Data source covering 1991–2018. Variables included neonatal, infant, and under-five mortality rates (dependent variables) and seven independent macroeconomic variables: births attended by skilled healthcare staff, pregnant women receiving prenatal care, newborns protected against tetanus, healthcare expenditure per capita, GDP growth, female labor force participation, and population growth. Unit root testing confirmed first-order integration for all variables, leading to the selection of Fully Modified Ordinary Least Squares (FMOLS) and Dynamic Ordinary Least Squares (DOLS) regression as appropriate analytical techniques for estimating long-run relationships.

**Key Results:** Both FMOLS and DOLS estimations consistently showed that 'births attended by skilled healthcare staff' had a negative and highly significant (p ≤ 0.01) coefficient for child mortality and infant mortality. 'Pregnant women receiving prenatal care' was negatively and highly significant (p ≤ 0.01) for both child and infant mortality in FMOLS estimation. 'Newborns protected against tetanus' was highly significant (p ≤ 0.01) with a negative coefficient across both estimation methods for child and infant mortality. 'Healthcare expenditure per capita' showed a negative and highly significant (p ≤ 0.01) effect on both child and infant mortality in both DOLS and FMOLS estimations. 'GDP growth' significantly (p ≤ 0.05) reduced both child and infant mortality. Conversely, 'female labor force participation' was positively and highly significant (p ≤ 0.01) for both infant and child deaths. 'Population growth' positively influenced both child and infant mortality and was highly significant (p ≤ 0.01) in FMOLS estimation. When disaggregated by mortality type, the same five variables (skilled birth attendance, prenatal care, tetanus protection, healthcare expenditure, GDP growth) negatively and significantly influenced infant deaths, under-five deaths, and neonatal deaths, while female labor force participation and population growth positively and significantly influenced all three mortality outcomes. The consistency of coefficient signs and significance levels across both DOLS and FMOLS estimations was considered robust.

**Clinical Implications:** The findings highlight that low-cost, scalable interventions—particularly tetanus vaccination, skilled birth attendance, and increased healthcare spending—are effective in reducing child mortality in resource-limited settings. Bangladesh's success since establishing the Expanded Program on Immunization (EPI) in 1979 demonstrates the impact of sustained vaccination coverage, though mortality remained at 37 deaths per 1000 live births in 2014 despite over 90% vaccine coverage. The study underscores that child mortality is not solely a health issue but a social problem with health consequences, requiring integrated social and health policies. Other developing countries with similar macroeconomic profiles can emulate Bangladesh's strategies to accelerate progress toward SDG 3 and MDG 4. Limitations include the single-country focus and the potential influence of political, economic, social, technological, environmental, and legal factors not captured in the model.