Data routinely collected by health facilities could become an additional source for measuring maternal mortality, but their quality needs to improve. Researchers from Karolinska Institutet and Tanzania propose four measures for assessing these data. The work is published in the Bulletin of the World Health Organization.
Maternal mortality refers to deaths during pregnancy, childbirth or the six weeks following pregnancy. Reliable information about these deaths is important for monitoring progress in maternal health. Yet accurate estimates remain difficult in many low- and middle-income countries, where systems for registering deaths are often incomplete or lacking.
Today, more than 80 percent of births take place in health facilities in about two thirds of low- and middle-income countries. This means that information routinely collected by hospitals and other health facilities could potentially complement existing sources of maternal mortality data. However, underreporting of maternal deaths is a concern.
“Health facility data could provide additional information for estimates of maternal mortality, but we first need to understand and improve the completeness and quality of the reporting,” says Claudia Hanson , Professor of Global Health Systems Research at the Department of Global Public Health .
Four measures to assess data quality
The researchers propose four measures to assess data quality. These examine whether reported maternal mortality levels and geographical patterns are plausible, the relationship between reported stillbirths and maternal deaths, the distribution of causes of death, and the age distribution among women who died.
Using data from Tanzania, the researchers found problems with quality and completeness across all four measures and the available data sources. For example, some districts reported implausibly low maternal mortality, while other patterns suggested that maternal deaths or stillbirths had not always been reported.
The researchers therefore stress that the four measures should be considered together and should not be regarded as a universal standard. Expected patterns may vary between countries and settings. They propose regular national assessments of data quality and independent checks of how completely health facilities report maternal deaths.
“If the quality of facility data improves, these data could eventually contribute to national and global estimates and give us additional information about maternal mortality,” says Claudia Hanson.
The work was carried out within the Countdown to 2030 initiative together with researchers and government representatives in Tanzania and other collaborating institutions. It was funded by the Bill & Melinda Gates Foundation, while Claudia Hanson received support from the Swedish Research Council. The authors report no competing interests.
Publication
Makuwani AM, Kagoye S, Masanja H, Ismail HR, Mpembeni R, Shabani J, Ng’weshemi SK, Masika G, Katalambula L, Boerma T, Hanson C. Measurement of maternal mortality, United Republic of Tanzania . Bull World Health Organ. 2026 Sep 1;104(9):631-639A. doi: 10.2471/BLT.25.295623
