Characterising perinatal mortality in South Africa: A cross-sectional analysis using Demographic Health Survey data

Abstract

Background: Perinatal mortality defined as the death of a fetus at or after 22 weeks of gestation and the death of liveborn in the first 7 days of life is a significant public health concern in Sub- Saharan Africa including countries like South Africa. While South Africa may be on track in meeting the Sustainable Development Goal (SDG) targets of reducing perinatal mortality, there are significant disparities within the country, with some provinces showing considerably higher rates of perinatal mortality. This underscores the need to investigate individual and community level factors associated with perinatal mortality to facilitate more effective targeted interventions. Aim: This study aims to investigate individual (maternal and neonatal) and community-level characteristics associated with perinatal mortality in South Africa using the Demographic and Health Survey (DHS) data in 2016. Methods: A cross-sectional design was used in analysing secondary data from 8,514 women of reproductive age. Perinatal mortality was the primary outcome and was calculated using total counts of perinatal mortality divided by the number of live births and multiplied by 1,000 at national and provincial level. Multivariable logistic regression models were used to identify the factors associated with perinatal mortality adjusting for survey design. Furthermore, random effects were incorporated at cluster level in the multilevel logistic regression to estimate community level effects. Results: The national perinatal mortality rate was 31 per 1,000 live births, with significant regional disparities. Women with a history of terminated pregnancies had four times the odds of perinatal mortality compared to those without (AOR 4.24; 95% CI: 2.25–6.48). Women aged 25–34 years (AOR 2.51; 95% CI: 1.46–4.31) and those aged 34+ (AOR 3.62; 95% CI: 2.14– 6.14) were significantly likely to experience perinatal mortality. Women who tested positive iii for HIV during pregnancy were significantly less likely to have perinatal mortality (AOR 0.60; 95% CI: 0.41–0.86). Perinatal mortality significantly varied by province, higher odds of perinatal mortality were observed in the Northern Cape (AOR 2.16; 95% CI: 1.12–12.33) and North West (AOR 3.31; 95% CI: 1.16–9.42) compared to the Western Cape. Household wealth was protective, with women from richer households less likely to experience perinatal mortality than those from poorer households (AOR 0.58; 95% CI: 0.36–0.96). About 20% of the variance was attributed to cluster/community effects. Conclusions: The study showed high levels of perinatal mortality at national and provincial level. Individual level factors like terminated pregnancy, maternal age, parity and HIV transmission, as well as community level factors like household wealth and province were associated perinatal mortality. Policymakers should consider tailored interventions at multiple levels including improving access and quality of maternal healthcare particularly among women of low socioeconomic status and enhancing uptake family planning services among women of reproductive age. Additionally, interventions tailored for provinces with high levels of perinatal mortality should be considered.