Accounting for stillbirths in maternal health metrics: a cross-country analysis.

Background: Live birth is only one of four potential pregnancy outcomes, alongside stillbirth, miscarriage, and induced abortion. While morbidity and mortality associated with all pregnancy outcomes are included in the numerator of many maternal metrics, often only live births are included in the denominator. This inconsistency makes interpreting trends challenging and may exacerbate the deprioritisation of monitoring other pregnancy outcomes. We assess the effect of using (1) total births (live births and stillbirths) or (2) total pregnancies (total births plus miscarriages/induced abortions) as the denominator on estimates of maternal and pregnancy-related mortality ratios (MMR and PRMR). Methods: Using data from Demographic and Health Surveys (DHS) conducted from 1996 to 2023, we estimated the proportion of pregnancies reported to end in live birth, stillbirth, or miscarriage/induced abortion in full pregnancy histories (DHS-VIII) or reproductive calendars (DHS-VII and earlier) for 46 countries in Africa, Asia, Latin American and the Caribbean, and Oceania. We calculated MMR and PRMR from the DHS sibling survival histories, adjusting the denominator by the reported distribution of pregnancy outcomes to account for either total births or total pregnancies. Findings: There was substantial cross-country heterogeneity in the proportion of pregnancies reported as ending in a live birth, ranging from 70% (Cambodia 2021) to 96% (Papua New Guinea 2017). Pregnancies reported as ending in stillbirth ranged from 0.3% (Timor-Leste 2016) to 4.1% (Lesotho 2014). Variability across countries might reflect differences in the distribution of pregnancy outcomes, temporal trends, and reporting practices. These differences result in non-uniform biases from using live births as the denominator. Using total births reduced the MMR and PRMR by up to 2.8% (Cote-d'ivoire 2021). Using total pregnancies reduced the MMR and PRMR by up to 23% (Cambodia 2021). Interpretation: Pregnancy-related morbidity and mortality can occur with any pregnancy outcome, not only live births. Progress in the availability of global stillbirth estimates means using total births as the denominator in maternal metrics is increasingly feasible in some countries and, in turn, could further strengthen momentum to institutionalise stillbirth reporting in civil registration systems. As the end of the SDG era approaches, the use of a more conceptually accurate maternal denominator based on total births should be explored in parallel with existing measures. However, better estimates of miscarriage and induced abortion are needed before total pregnancies can be used in global maternal metrics. Funding: UG and JW were supported by the UK Research and Innovation (EP/Y031172/1) and the Leverhulme Trust (RC-2018-003) for the Leverhulme Centre for Demographic Science. HES and LL were supported by the Eunice Kennedy Shriver National Institute of Child Health and Human Development (R01HD107015) and the Gates Foundation Child and Adolescent Causes of Death Estimation (CA-CODE, INV-038624).