Age and gender effects in self-reported urinary schistosomiasis in Tanzania

OBJECTIVE To characterize age-gender prevalence profiles of urinary schistosomiasis according to the questionnaire responses, compare the profiles to field survey data from selected regions, and determine if the profiles varied spatially throughout Tanzania. METHODS In 2004, a national school-based questionnaire survey for self-reported schistosomiasis and blood in urine (BIU) was conducted in all regions of mainland Tanzania, to assist targeted mass distribution of praziquantel. Field survey data were collected in six north-western and five coastal regions using microscopic examination of urine samples for the presence of Schistosoma haematobium eggs and assessment of micro-haematuria with chemical reagent strips. Bayesian logistic regression models were created to calculate age-gender profiles adjusted for demographic and ecological covariates and spatial correlation in the questionnaire data. Separate odds ratios (OR) for age-gender effects were calculated in each administrative area. RESULTS Data were obtained from > 2.5 million schoolchildren. Boys had higher prevalence of self-reported schistosomiasis and BIU than girls. In boys, prevalence according to the questionnaire and field surveys followed similar age profiles. However, in girls, prevalence according to the field surveys increased in older age groups, but flattened out or decreased according to the questionnaire, indicating the latter underestimated prevalence in older girls. In the models, little spatial correlation was evident in the OR for the age-gender effects, suggesting that these did not vary spatially. CONCLUSION Age-gender patterns of urinary schistosomiasis were consistent in different geographical areas of Tanzania. Because the questionnaire underestimated prevalence in older girls, we propose that upward calibration of observed prevalence is done for older females only.