Exploring the Feasibility and Validity of Proxy Pregnancy Histories Generated via a Mobile Phone Survey in Malawi.

Jethro Banda, Malawi Epidemiology Intervention Research Unit
Albert Dube, Malawi Epidemiology and Intervention Research Unit
Lena Kan, Johns Hopkins University
Mayank Date, Johns Hopkins University
Malebogo Tlhajoane, London School of Hygiene and Tropical Medicine
Diwakar Mohan, Johns Hopkins University
Amelia Crampin, Karonga Prevention Study
Georges Reniers, London School of Hygiene and Tropical Medicine (LSHTM)
Li Liu, Johns Hopkins University

Background: Proxy Pregnancy History (PPH) is a tool to collect pregnancy histories through proxy reporting. PPH can increase the effective sample size for estimating rates of child mortality and reproductive outcomes and allows data collection irrespective of the mother’s survival status. However, its validity is poorly understood. Methods: We conducted phone interviews with 500 index women and a nominated sister as proxies about the index’s pregnancy histories. We compare individual reports and aggregated mortality rates. Results: Proxies were older, more likely to own phones, and more educated than indexes. 80.4% and 86.5% reported the same number of child births and deaths respectively as index. Among matched pregnancies (n=1160), over 90% had the same pregnancy outcome, and sex and vital status of children. Most mortality probabilities were marginally higher for index. Conclusions: We observe high correspondence in pregnancy history reporting between index and proxies, suggesting PPH may be a promising methodology.