Data on Causes of

Understanding the Data

Mortality Coding Background

The World Health Organization (WHO) has defined a set of procedures or coding rules to be followed for selecting causes of death and for recording them on death certificates, with the aim of international comparability of mortality data.

Members of Determination of Cause of Death, or DeCoDe, panels at each CHAMPS site review all data available for each death and follow WHO guidance to attribute causes of death. These procedures include using ICD-10 (10th International Statistical Classification of Diseases and Related Health Problems) to classify conditions and recording causes of death in a format matching WHO’s death certificate. In some situations, a death will have an immediate cause (lower respiratory infection, for example) that occurred in an otherwise healthy child. In these situations, the immediate and underlying cause of death will be the same. In other situations, a death may have an immediate cause (lower respiratory infection, for example) that occurred as a consequence of a child having other diseases (HIV infection, for example). In this case, the condition that originally existed that gave rise to one or more complications that led to death is known as the underlying cause of death. Comorbid conditions are those that may have contributed to the cause of death, but not directly. CHAMPS records immediate and underlying causes of death and comorbid conditions. In rare cases, the data available to DeCoDe panels are not adequate to determine a cause of death with any certainty. In these situations, the cause of death is recorded as undetermined.

For stillbirths and newborns, causes of death are often related to conditions affecting the mother or complications of childbirth or preterm birth. For these situations, CHAMPS uses a WHO application of ICD-10 for the perinatal period, ICD-PM. Because deaths in young infants may be attributed to both perinatal causes and conditions that arise after birth, CHAMPS is working to standardize procedures for deaths occurring in this age group. In situations where ICD-10 and ICD-PM codes are inconsistent, ICD-PM principles will be applied. Note that results may change as data are updated and coding procedures become more refined.

Data Collection Methods

CHAMPS surveillance sites utilize notification systems which report all under-5 deaths and stillbirths to the local team within 24 hours of the child’s death. Site teams review each notification and contact family members of select cases to determine whether the case meets CHAMPS eligibility criteria (death at age <60 months or stillbirth; and residence in the defined geographic area).

If a case meets CHAMPS eligibility criteria, the team seeks informed consent from the parents or guardians to further investigate the cause of death through the postmortem minimally invasive tissue sampling (MITS) procedure and laboratory investigations. The laboratory investigations include microbiology, HIV, tuberculosis and malaria testing. Additional advanced diagnostics include multiplex molecular testing for many specific viral, bacterial, fungal and parasitic pathogens as well as tissue histopathology. The site teams also request consent to perform caregiver interviews, known as verbal autopsies, to explore the symptoms and conditions that may have led to the child’s death, and to collect any clinical records. Families are free to decide whether to take part in the procedures.

For those cases in which families grant consent and the MITS procedure has been performed, completing the tests and assembling the information requires 4 months. Then, an expert panel reviews all laboratory, clinical and verbal autopsy information on each case and assigns a final cause of death. This final cause of death is then provided to the family members.

Things to Note

CHAMPS sites are continuously entering and transmitting data, therefore more recent data may be missing or may have not yet completed data quality review.

Only a fraction of child deaths occurring in CHAMPS sites are reflected; the numbers are not adjusted for deaths in children not enrolled or for population size. No conclusions should be made about the overall contribution of each cause to total deaths for a CHAMPS site or other population.

Data that contribute to CHAMPS cause of death determinations are complex, and we're learning how expert panels should interpret it to assign causes of death. How we report the chain of causation will likely change over time.

Results presented here will change as the CHAMPS team refines methods and cleans data. If you find information that appears incorrect or confusing, contact us.