Measuring Childhood Mortality through Mobile Phone Interviews in Mozambique

Almamy Kante, Johns Hopkins Bloomberg School of Public Health
Cremildo Manhica, National Institute of Health (INS)
Akum Aveika, Johns Hopkins Bloomberg School of Public Health
Azarias Mulungo, National Institute of Health (INS)
Fred van Dyk, Johns Hopkins Bloomberg School of Public Health
Nordino Machava, National Institute of Health (INS)
Helen Kuo, Johns Hopkins Bloomberg School of Public Health
Charfuddin Saccor, Centro de Investigacão em Saude de Manhica (CISM)
Dustin Gibson, Johns Hopkins Bloomberg School of Public Health
Robert Black, Johns Hopkins Bloomberg School of Public Health
Ivalda Macicame, Instituto Nacional de Saúde - Moçambique
Agbessi Amouzou, Johns Hopkins Bloomberg School of Public Health

Introduction This study explored alternative methods for measuring childhood mortality in low- and middle-income countries using mobile phone surveys in Mozambique. Traditionally, face-to-face maternity history interviews are conducted through household surveys, but this research examined the mobile phone interviews with women of reproductive age. Method Two sampling strategies were used: one from the Countrywide Mortality Surveillance for Action (COMSA) database and the other using random digit dialing (RDD) with an Interactive Voice Response (IVR) system. The study reached 13,545 women through the COMSA database and 10,359 through RDD. Childhood mortality rates and trends were calculated for 2012-2021 and compared with United Nations estimates, the 2022 Demographic and Health Survey (DHS) and the main COMSA platform. Results Both methods yielded comparable results to DHS, with U5MR estimates from COMSA subsample and RDD/IVR sample being slightly lower than UN predictions and the main COMSA platform. NMR and IMR followed the same patterns as the U5MR. However, the RDD method consistently underestimated mortality compared to COMSA. Discussion Despite this, the study demonstrated the feasibility of mobile phone surveys for mortality measurement, but further research is required to refine methods, especially with RDD, and to address potential errors and biases in data collection.

Keywords: Data and Methods, Mortality and Longevity, Comparative methods , Population and Development

See extended abstract.