Objective: To assess the proportion of rifampicin-resistant tuberculosis (RR-TB) patients with potential earlier RR-TB diagnoses in Khayelitsha, South Africa.
Design: We conducted a retrospective analysis among RR-TB patients diagnosed from 2012 to 2014. Patients were considered to have missed opportunities for earlier diagnosis if 1) they were incorrectly screened according to the Western Cape diagnostic algorithm; 2) the first specimen was not tested using Xpert® MTB/RIF; 3) no specimen was ever tested; or 4) the initial Xpert test showed a negative result, but no subsequent specimen was sent for follow-up testing in human immunodeficiency virus-positive patients.
Results: Among 543 patients, 386 (71%) were diagnosed with Xpert and 112 (21%) had had at least one presentation at a health care facility within the 6 months before the presentation at which RR-TB was diagnosed. Overall, 95/543 (18%) patients were screened incorrectly at some point: 48 at diagnostic presentation only, 38 at previous presentation only, and 9 at both previous and diagnostic presentations.
Conclusions: These data show that a significant proportion of RR-TB patients might have been diagnosed earlier, and suggest that case detection could be improved if diagnostic algorithms were followed more closely. Further training and monitoring is required to ensure the greatest benefit from universal Xpert implementation.