RESUMO
OBJECTIVES: The availability of cause-specific mortality data for children in South Africa is limited. Hospital-based data have the potential to contribute to understanding of the causation of childhood death in South Africa. The objectives of the study were to gain insights into the causes of death in a South African children's hospital. DESIGN: Prospective, descriptive study of death notification forms. SETTING: Red Cross War Memorial Children's Hospital, Cape Town. METHODS: Data from 1999 to 2003 were analysed by direct and underlying causes of death (using a modified Global Burden of Diseases (GBD) classification) and demographic variables. Death rates per 1 000 hospital admissions were calculated for certain common causes of death. Seasonal correlates of mortality were examined. RESULTS: There were 1 978 deaths. The number of deaths per year increased by 11.4% over the period. The death rate rose from 15.9 to 18.4 per 1 000 admissions from 1999 to 2002, declining to 17.4/1 000 in 2003. The death rate was higher for females than for males (18.4/1 000 versus 17.6/1 000, p = 0.007). Sixty per cent of deaths occurred in children less than 1 year old. GBD group I diseases (infectious, nutritional, perinatal) accounted for the greatest proportion of deaths (58.6%), followed by noncommunicable diseases (29.1%), and injuries (7.9%). HIV/AIDS accounted for 60% of infectious deaths (31.6% of all deaths). Diarrhoea-related mortality was 3 times higher in summer than in winter. Congenital conditions dominated GBD group II (57.5%). CONCLUSION: The analysis shows the value of routinely recording data on childhood hospital deaths. The results mirror those of the South African Medical Research Council's Burden of Disease studies but also reflect the hospital's tertiary functions. Female children were at higher risk of death. Childhood HIV-related deaths are a major challenge to the health system.