Estimating the distribution of causes of death among children age 1-59 months in highmortality countries with incomplete death certification

Background: Our objective was to develop a methodology to estimate causes of death among children age 1-59 months in high child mortality countries without adequate vital registration (VR) systems. Methods: We systematically reviewed community-based studies reporting at least two causes of death among children 1-59 months of age identified from published and unpublished sources. We included (i) studies conducted after 1979, (ii) for duration of 12 months or an exact multiple, (iii) with ≥25 deaths in children <5 years, (iv) each death represented once and (v) <25% of deaths due to unknown causes. A study-based multinomial logistic regression model was applied to country-level data to estimate causes of child death. Results: Of the 216 studies reviewed, 81 were included in the analysis comprising 79 067 under-5 deaths from 25 countries. After adjusting for risk factors and intervention coverage, the estimated distribution of causes of deaths in children 1-59 months of age in sub-Saharan Africa and Southeast Asia was: pneumonia (21 and 31%), diarrhoea (25 and 31%), malaria (26 and 2%), injury (3 and 4%), meningitis (3 and 4%), measles (3 and 2%) and other causes (20 and 27%), respectively. Conclusion: From studies reporting as few as two different causes of death, statistical modelling can be used to estimate the causes of child mortality for settings with incomplete VR. Pneumonia and diarrhoea remain the leading causes of death among children 1-59 months of age in sub-Saharan Africa and Southeast Asia.