Measurement Issues and Implications for Improvements in Stillbirth Rates from Retrospective Surveys in Low-Income Settings

Stillbirth data are routinely collected through retrospective household surveys, e.g., the Demographic and Health Surveys, in low-income settings. However, such data are subject to a number of known biases, including those due to misclassification of neonatal deaths and retrospective self-reporting. Previous research has shown that the extent and direction of misclassification vary across contexts. Women of reproductive age, the respondents of retrospective surveys, do not always have accurate information on their babies' birth outcomes as they are not consistently informed by health care providers during and after delivery. Moreover, little is known about the under-reporting associated with retrospective surveys in stillbirths. 

This pre-formed panel offers new insights into these evidence gaps through three studies from Guinea Bissau and Tanzania using mixed methods. Results from these studies could not only inform the interpretation and adjustment of retrospective survey and model-based stillbirth estimates for better progress tracking, but also identify suboptimal quality of care issues in delivery and postpartum care to further improve maternal and newborn outcomes. The pre-formed panel is co-organized by Johns Hopkins University, the Bandim Health Project, and University of Southern Denmark.