Setting: The impact of Xpert(®) MTB/RIF as a follow-on diagnostic test after smear microscopy on tuberculosis (TB) notification has not yet been well defined.
Design: Quasi-experimental design with 86 evaluation and 49 control clinics in Kinshasa, Democratic Republic of Congo. Smear microscopy was supported at all 135 clinics, Xpert was placed in 15 evaluation clinics and a sputum transport system was implemented for 25 satellite clinics. The number of cases notified before and during the project (July 2012-June 2013) was obtained from the National TB Program.
Results: Of 27,147 presumptive TB cases presenting in clinics with access to Xpert, 5922 (21.8%) were smear-positive. Of 18,636 individuals with ⩾ 3 negative microscopy results, 6920 (37.1%) underwent Xpert testing, 991 (14.3%) of whom tested positive. The number of bacteriologically positive cases increased equally in evaluation clinics (15.1%, 95%CI -2.3 to 32.6) and control clinics (13.6%, 95%CI 2.6-29.3), for a difference in increase of 1.5% (95%CI -28.8 to 31.8). There was no difference in the change in smear-negative cases (-42.4%, 95%CI -111.5 to 26.6), nor in all types of TB notified (-6.1%, 95%CI -32.5 to 20.4) between the evaluation and control clinics.
Conclusion: In part due to a restrictive algorithm, Xpert as follow-on to smear microscopy did not increase the overall number of TB notifications, nor the number of bacteriologically positive cases.