Reliable and timely data on causes of death are needed to guide effective policy and programmes in countries to improve child survival.
Definition:
Death rates (per 1,000 live births) in children under 5 years, by cause, age group with cause defined by ICD-10 codes for common conditions.
Disaggregation:
Age, Cause
Method of measurement
Data sources: Data from civil registration with complete coverage (80% or over) and medical certification of cause of death, or nationally representative epidemiological studies of causes of child death (special studies analysing causes of death based on verbal autopsy studies or other sources for countries without civil registration data).
Method of estimation:
Estimates of child causes of death were prepared by WHO and the Child and Adolescent Cause of Death Estimation group (CA CODE). The number of live births are obtained from the World Population Prospects produced by the United Nations Population Division from the latest revision that were available at the time of estimation. WHO regularly receives mortality-by-cause data from Member States, as recorded in national civil registration systems. These statistics are evaluated for their completeness and quality. Complete and nationally-representative data were then grouped by ICD codes into the cause categories, and the proportions of these causes with regard to the total number of deaths of children under 5 years were then computed. For low mortality countries without adequate vital registration data, the cause distribution was estimated using a Bayesian multinomial regression model applied to death registration data. For high mortality countries without adequate vital registration data, the cause distribution was estimated using a Bayesian multinomial regression model applied to (largely) verbal autopsy (VA) data from research studies. Cause-specific under-five mortality estimates from the WHO technical programmes and UNAIDS are taken into account in assigning the distribution of deaths to specific causes.
For complete methods please see: Perin J, Prieto-Merino D, Wahi A, Pramanik S, Eilerts-Spinelli H, Villavicencio F, McCormick T, You D, Abhirup D, Strong K, Black R, Liu L. Systematic estimates of global causes of neonatal and under 5 mortality in 2000-24: secondary data analysis using Bayesian multinomial logistic regression BMJ 2026; 393 :e088686 doi:10.1136/bmj-2025-088686.
Predominant type of statistics: predicted and adjusted.
Method of estimation of global and regional aggregates:
Estimates for UN Member States using Bayesian multinomial logistic regression
Other possible data sources:
Special studies
Preferred data sources:
Civil registration with complete coverage and medical certification of cause of death
Unit of Measure:
Rate per 1,000 live births
Expected frequency of data dissemination:
Biennial
Expected frequency of data collection:
Biennial
Comments:
Three of the cause groups used for this indicator require special definition:
“Other Group 1 (postneonatal)” includes other infections and conditions that occurred in the newborn period but resulted in a postneonatal death, i.e., tetanus, neonatal sepsis, prematurity, and birth asphyxia/trauma.
“Other noncommunicable diseases (postneonatal only)” includes other NCDs, i.e. cancers.
“Other Group 1 and Other noncommunicable (neonatal and under-5 only)” includes other infectious causes of the newborn, i.e. HIV/AIDS, measles, pertussis.
Estimates are useful for country comparisons and trends over time but not for sub-national level unless produced for subnational populations, or for monitoring programme specific outputs and outcomes.
If you have any feedback, you are welcome to write it here.
If you need to access the old Global Health Observatory data, you can do it here. But before you leave, please provide us your feedback about our new data portal.