Validation of smear-positive tuberculosis diagnoses in Bangladesh.

Abstract

Questions have been raised as to the accuracy of smear-positive tuberculosis (TB) diagnoses at DOTS centres in Bangladesh. To assess the validity of diagnoses by smear microscopy at 62 DOTS centres in various health facilities. Using a cross-sectional study design, smear-positive TB cases were included from high, medium and low-performing DOTS centres from July to December 2010. The centres were categorised according to their case detection rate: centres with case detection rates of >70%, 50-70% and <50% were classed as high, medium and low-performing, respectively. Microscopy was performed on spot and early morning sputum samples for all cases. A convenience sample of 333 recently diagnosed smear-positive TB cases was included. The average number of patients enrolled per centre was 5.4; the highest number of cases was diagnosed from low-performing DOTS centres. Of the 333 cases, 302 (90.7%) were confirmed, while 31 (9.3%) were smear-negative. Differences in microscopy results could be due to the poor quality of the sputum sample, of smearing, staining or reading, or erroneous reporting by the DOTS centres. Preset programme targets and the nationwide drug stockout may have led to overreporting by the DOTS centres.