Abstract P165: Geographic Co-morbidity of Cardiovascular Diseases in South Africa

Background: The growing burden of non-communicable diseases in Africa is well recognized. However, the interplay between chronic conditions and environmental and lifestyle factors is not well documented in this setting. We used statistical co-morbidity models to jointly estimate risks of four cardiovascular conditions: hypertension, coronary heart disease (CHD), stroke and hypercholesterolaemia that are known to share unhealthy diet and lifestyle as common risk factors. Methods: Data are from 13,827 adults (mean ±SD age 39 ±18, 58.4% women) interviewed in the 1998 South African Health and Demographic Survey aggregated to 52 health districts. We estimated shared-component spatial prevalence models to known individual risk factors, in addition to allowing for disease-specific spatially structured district random effects. Results: There appeared to be an associations between the district prevalence of hypertension and CHD (ρ =0.33), and stroke and hypercholesterolaemia (ρ =0.31). However, on the unadjusted disease-specific prevalence maps, it was not possible to prove geographical trends in prevalence. On covariate-adjusted smoothed maps, the disease-specific log-odds showed clearer spatial patterns. Hypertension and stroke were concentrated highly in south-western parts of the country; whereas CHD and hypercholesterolaemia were highly concentration in central north-eastern and the top north-eastern corridors, respectively. The shared component, which we took to represent nutrition and other lifestyle factors not accounted for in the model, had a larger effect on vascular disease prevalence in south-western areas of the country. The shared component appeared to have greater effect on hypertension and CHD, though the relative weights were not statistically significant. Conclusion: This study provides novel evidence on disease-specific risk and geographic distribution of cardiovascular risk in South Africa. Understanding the underlying risk interactions between diseases might be useful for public health resource allocation in this setting.