Child Deaths in Ghana Cluster by Region and Family, 2022 Health Survey Finds

Under-five mortality in Ghana remains a stubborn public health challenge, and a new analysis of the country's most recent Demographic and Health Survey offers a detailed statistical portrait of where child deaths concentrate and which maternal and household characteristics are associated with them. The study, a population-based cross-sectional secondary analysis of the 2022 Ghana Demographic and Health Survey published in BMC Pediatrics, examined 8,580 weighted children and identified 266 deaths among children under five, corresponding to a mortality rate of 31 deaths per 1,000 births. The headline figure of 3.1 percent, with a 95 percent confidence interval of 2.8 to 3.5 percent, masks considerable geographic variation.

The regional distribution of under-five mortality was highest in the Oti, Western North, Volta, Ahafo, and Greater Accra regions, and the researchers report significant variations across multiple administrative areas. That spread matters because Ghana has invested in national initiatives, including the integrated Reproductive, Maternal, Newborn, Child, and Adolescent Health and Nutrition policy framework and the National Health Insurance Scheme, yet the burden clearly does not fall evenly across the country. Methodologically, the study leaned on multilevel logistic regression, an approach suited to data in which children are nested within households and communities.

Rather than treating every child as an independent observation, multilevel models partition variation in mortality into components attributable to individual and maternal factors and components attributable to higher-level contextual differences, such as region of residence. The authors fitted a series of models and judged them using the receiver operating characteristic curve and the area under the curve, alongside the Akaike and Bayesian Information Criteria, which reward models that fit well without unnecessary complexity. Statistical significance was set at p below 0.05.

The results complicate some conventional assumptions about child survival. Children of mothers with lower levels of education showed varying associations with under-five mortality, consistent with the idea that maternal schooling shapes health knowledge, care-seeking, and household resources. More striking, however, was the finding on antenatal care: children whose mothers reported higher antenatal visit frequencies, specifically eight or more visits, showed significantly higher odds of under-five mortality in model 2 and in the full combined model, with an adjusted odds ratio of 4.38 and a 95 percent confidence interval of 1.69 to 11.37.

That counterintuitive association deserves careful interpretation. The authors nonetheless flag antenatal care utilization as a key factor associated with child mortality in Ghana, and their conclusion emphasizes improving the quality, affordability, and accessibility of these services alongside education. Parity, the number of children a mother has borne, showed a strong positive association with under-five mortality rather than a protective one, with an adjusted odds ratio of 8.05 and a 95 percent confidence interval of 4.87 to 13.32.

In practical terms, children born to mothers with higher parity faced markedly higher odds of dying before their fifth birthday in this dataset. One of the study's most technically interesting results is the residual intraclass correlation coefficient, reported as 0.178. This means that roughly 18 percent of the residual variation in under-five mortality was attributable to differences between higher-level units, such as communities or regions, after accounting for the measured individual and maternal factors.

In other words, where a child lives continues to matter even once family-level characteristics are taken into consideration, underscoring the regional disparities the descriptive analysis revealed. The authors frame their findings against the Sustainable Development Goals, and their recommendation is directed at governments: improve the quality, affordability, and accessibility of education and antenatal care services to reduce child mortality and advance progress toward those goals. Given that the survey data show deaths concentrated in particular regions and linked to maternal education, parity, and care patterns, the analysis suggests that national averages conceal the heterogeneity that any effective policy response would need to address.

As with all cross-sectional analyses, the associations reported here describe patterns in the 2022 survey data rather than demonstrated causal effects, and the study is based on weighted secondary data from that single survey round.

Subject of Research: Risk factors and regional distribution of under-five mortality in Ghana based on the 2022 Demographic and Health Survey

Article Title: Multi-level modeling of risk factors of under-five mortality in Ghana: evidence from the 2022 demographic and health survey

Article References: Mensah, M., Gyapong-Osei, N., & Opoku, S. (2026). Multi-level modeling of risk factors of under-five mortality in Ghana: evidence from the 2022 demographic and health survey. BMC Pediatrics. https://doi.org/10.1186/s12887-026-07946-5

Image Credits: AI Generated

DOI: 10.1186/s12887-026-07946-5

Keywords: under-five mortality, Ghana, Demographic and Health Survey, multilevel modeling, maternal education, antenatal care, parity, regional disparities, child health, BMC Pediatrics, Sustainable Development Goals, Multi-level