Spatial epidemiology of hospital-diagnosed brucellosis in Kampala, Uganda

BACKGROUND A retrospective case-control study was undertaken to examine the spatial risk factors for human brucellosis in Kampala, Uganda. METHODS Information on age, sex and month of diagnosis was derived from records from plate agglutination tests undertaken at Mulago Hospital, Kampala. Information on Parishes (LC2s) where patients reside was sourced from the outpatient registration book. In-patient fracture cases were selected for use as controls using 1:1 matching based on the age, sex and month of diagnosis. The locations of cases and controls were obtained by calculating Cartesian coordinates of the centroids of Parish level (LC2) polygons and a spatial scan statistic was applied to test for disease clustering. Parishes were classified according to the level of urbanization as urban, peri-urban or rural. RESULTS Significantly more females than males were found to show sero-positivity for brucellosis when compared with the sex ratio of total outpatients, in addition female brucellosis patients were found to be significantly older than the male patients. Spatial clustering of brucellosis cases was observed including around Mulago Hospital (radius = 6.8 km, p = 0.001). The influence of proximity to the hospital that was observed for brucellosis cases was not significantly different from that observed in the controls. The disease cluster was confounded by the different catchment areas between cases and controls. The level of urbanization was not associated with the incidence of brucellosis but living in a slum area was a significant risk factor among urban dwellers (odds ratio 1.97, 95% CI: 1.10-3.61). CONCLUSIONS Being female was observed to be a risk factor for brucellosis sero-positvity and among urban dwellers, living in slum areas was also a risk factor although the overall risk was not different among urban, peri-urban and rural areas of the Kampala economic zone.