Estimates of Under-Five Mortality from a Mobile Phone Survey during the COVID-19 Outbreak: an Observational Study Comparing Three Instruments in Malawi

Julio E. Romero-Prieto, London School of Hygiene and Tropical Medicine (LSHTM)
Boniface Dulani, Institute of Public Opinion and Research
Bruno Masquelier, Louvain University (UCL)
Malebogo Tlhajoane, London School of Hygiene and Tropical Medicine
Stéphane Helleringer, Columbia University
Jethro Banda, Malawi Epidemiology Intervention Research Unit
Georges Reniers, London School of Hygiene and Tropical Medicine (LSHTM)

Under-five mortality (U5M) estimates for low- and middle-income countries are primarily derived from detailed birth/pregnancy histories collected through in-person household surveys. These are, however, resource intensive and vulnerable to interruption during epidemics and crises. Remotely deployed Mobile Phone Surveys (MPS) can circumvent these disadvantages, but their suitability for monitoring population-level mortality has not been demonstrated. We examine Malawian MPS data from Summary Birth Histories, Truncated Pregnancy Histories (TPH), and Full Pregnancy Histories (FPH) for estimating U5M. Considering the limited penetration and the unequal distribution of mobiles in Malawi, sampling strategies were implemented to increase the participation of less represented populations and post-stratification methods were applied to attenuate potential selection bias. Resulting probabilities of dying q(x)–before 28 days, 12 months, and 60 months of life–are compared against external estimates from the DHS and the UN Inter-agency Group of Child Mortality Estimation (UN-IGME). Estimates from the TPH instrument are affected by downward bias. Estimates of q(28d), q(12m), and q(60m) from the FPH instrument are in line with those published by the UN-IGME, but they are also suggestive of a mortality excess during the COVID-19 outbreak in 2020-22. MPS are a promising method for collecting U5M, and particularly so via the FPH instrument.

Keywords: Mortality and Longevity, Data and Methods, Randomized controlled experiments

See paper.