RESUMO
AIM: Field-testing a paediatric mortality audit system at eight sites in South Africa to assess its feasibility, to collect local data on common causes of death in children under 5 y, and to determine health system failure, missed opportunities of intervention and substandard care (modifiable factors). METHODS: The study tested the Under-5 Healthcare Problem Identification Programme (U5PIP). The sites represent different levels of paediatric healthcare. STUDY PERIOD: 1 September 2003-31 August 2004. STUDY POPULATION: Under-5 patients admitted to study hospitals. CASES: Under-5 patients who died in study hospitals. RESULTS: In total, 1532 under-5 deaths occurred, representing a case-fatality rate of 7.8%. Main causes of death were lower respiratory tract infections (33%), gastroenteritis (15%) and septicaemia (12%). Sixty per cent of the deaths were HIV/AIDS related. Sixty-nine per cent of children who died were underweight. Administrative modifiable factors were present in 31% of deaths. Clinical personnel-related modifiable factors were detected in 26% at the clinic level and in 33-37% at the hospital level. CONCLUSION: The U5PIP is feasible for ongoing mortality reviews by paediatric teams as part of routine work. Information on common causes of death and modifiable factors in this study focus on the impact of HIV/AIDS, malnutrition and resource allocation, and can be used for interventions to improve paediatric healthcare.