Background

With the expansion of Prevention of Mother to Child Transmission (PMTCT) services in Senegal, there is growing interest in using PMTCT program data in lieu of conducting unlinked anonymous testing (UAT)-based ANC Sentinel Surveillance. For this reason, an evaluation was conducted in 2011–2012 to identify the gaps that need to be addressed while transitioning to using PMTCT program data for surveillance.

Methods

We conducted analyses to assess HIV prevalence rates and agreements between Sentinel Surveillance and PMTCT HIV test results. Also, a data quality assessment of the PMTCT program registers and data was conducted during the Sentinel Surveillance period (December 2011 to March 2012) and 3 months prior. Finally, we also assessed selection bias, which was the percentage difference from the HIV prevalence among all women enrolled in the antenatal clinic and the HIV prevalence among women who accepted PMTCT HIV testing.

Results

The median site HIV prevalence using routine PMTCT HIV testing data was 1.1% (IQR: 1.0) while the median site prevalence from the UAT HIV Sentinel Surveillance data was at 1.0% (IQR: 1.6). The Positive per cent agreement (PPA) of the PMTCT HIV test results compared to those of the Sentinel Surveillance was 85.1% (95% CI 77.2–90.7%), and the percent-negative agreement (PNA) was 99.9% (95% CI 99.8–99.9%). The overall HIV prevalence according to UAT was the same as that found for women accepting a PMTCT HIV test and those who refused, with percent bias at 0.00%. For several key PMTCT variables, including “HIV test offered” (85.2%), “HIV test acceptance” (78.0%), or “HIV test done” (58.8%), the proportion of records in registers with combined complete and valid data was below the WHO benchmark of 90%.

Conclusions

The PPA of 85.1 was below the WHO benchmarks of 96.6%, while the combined data validity and completeness rates was below the WHO benchmark of 90% for many key PMTCT variables. These results suggested that Senegal will need to reinforce the quality of onsite HIV testing and improve program data collection practices in preparation for using PMTCT data for surveillance purposes.