Background
Provider-initiated HIV testing and counselling (PITC) is promoted as a means to increase HIV case finding. We assessed the effectiveness of PITC to increase HIV testing rate and HIV case finding among outpatients in Rwandan health facilities (HF).

Methods
PITC was introduced in six HFs in 2009-2010. HIV testing rate and case finding were compared between phase 1 (pre-PITC) and phase 3 (PITC period) for outpatient-department (OPD) attendees only, and for OPD and voluntary counseling & testing (VCT) departments combined.

Results
Out of 26,367 adult OPD attendees in phase 1, 4.7 % were tested and out of 29,864 attendees in phase 3, 17.0 % were tested (p < 0.001). The proportion of HIV cases diagnosed was 0.25 % (67/26,367) in phase 1 and 0.46 % (136/29864) in phase 3 (p < 0.001). In multivariable analysis, both testing rate and case finding were significantly higher in phase 3 for OPD attendees. In phase 1 most of the HIV testing was done in VCT departments rather than at the OPD (78.6 % vs 21.4 % respectively); in phase 3 this was reversed (40.0 % vs 60.0 %; p < 0.001). In a combined analysis of VCT and OPD attendees, testing rate increased from 18.7 % in phase 1 to 25.4 % in phase 3, but case finding did not increase. In multivariable analysis, testing rate was significantly higher in phase 3 (OR 1.67; 95 % CI 1.60-1.73), but case finding remained stable (OR 1.09; 95 % CI 0.93-1.27).

Conclusion
PITC led to a shift of HIV testing from VCT department to the OPD, a higher testing rate, but no additional HIV case finding.