Measuring the local burden of disease. A study of years of life lost in sub-Saharan Africa.

BACKGROUND An effective health policy necessitates a reliable characterization of the burden of disease (BOD) by cause. The Global Burden of Disease Study (GBDS) aims to deliver this information. For sub-Saharan Africa (SSA) in particular, the GBDS relies on extrapolations and expert guesses. Its results lack validation by locally measured epidemiological data. METHODS This study presents locally measured BOD data for a health district in Burkina Faso and compares them to the results of the GBDS for SSA. As BOD indicator, standard years of life lost (age-weighted YLL, discounted with a discount rate of 3%) are used as proposed by the GBDS. To investigate the influence of different age and time preference weights on our results, the BOD pattern is again estimated using, first, YLL with no discounting and no age-weighting, and, second, mortality figures. RESULTS Our data exhibit the same qualitative BOD pattern as the GBDS results regarding age and gender. We estimated that 53.9% of the BOD is carried by men, whereas the GBDS reported this share to be 53.2%. The ranking of diseases by BOD share, though, differs substantially. Malaria, diarrhoeal diseases and lower respiratory infections occupy the first three ranks in our study and in the GBDS, only differing in their respective order. Protein-energy malnutrition, bacterial meningitis and intestinal nematode infections occupy ranks 5, 6 and 7 in Nouna but ranks 15, 27 and 38 in the GBDS. The results are not sensitive to the different age and time preference weights used. Specifically, the choice of parameters matters less than the choice of indicator. CONCLUSIONS Local health policy should rather be based on local BOD measurement instead of relying on extrapolations that might not represent the true BOD structure by cause.