Routine data from prevention of mother-to-child transmission (PMTCT) HIV testing not yet ready for HIV surveillance in Mozambique: a retrospective analysis of matched test results

BACKGROUND: Opt-out HIV testing is offered at 70% of antenatal care (ANC) clinics in Mozambique through the prevention of mother-to-child transmission (PMTCT) program. If routine data from this program were of sufficient quality, their heightened coverage and continuous availability could complement or even replace biannual sentinel serosurveys that currently serve as the primary HIV surveillance system in Mozambique. METHODS: We assessed the efficacy of routine HIV testing data from prevention of mother-to-child transmission programs for estimating the prevalence of HIV infection among pregnant women. The PMTCT program uses sequential point-of-care rapid tests conducted on site while ANC surveillance surveys use dried blood spots tested sequentially for HIV-1/2 antibodies at a central laboratory. We compared matched routine PMTCT and ANC surveillance test results collected during 2007 and 2009 ANC surveillance surveys from 36 sentinel sites. RESULTS: After excluding 659 women without PMTCT data, including 83 who refused rapid testing, test results from a total of 20,563 women were available. Pooling the data from both years indicated HIV prevalence from routine PMTCT testing was 14.4% versus 15.2% from surveillance testing (relative difference -5.1%; absolute difference -0.78%). Positive percent agreement (PPA) of PMTCT versus surveillance tests was 88.5% (95% Confidence Interval CI: 85.7-91.3%), with 19 sites having PPA below 90%; Negative percent agreement (NPA) was 98.9% (CI: 98.5-99.2%). No significant difference was found among three regions (North, Center and South), however both PPA and NPA were significantly higher in 2009 than 2007 (p < 0.05). CONCLUSIONS: We found low PPA of PMTCT test results compared to surveillance data which is indicative either of testing errors or data reporting problems. Nonetheless, PPA improved significantly from 2007 to 2009, a possible positive trend that should be investigated further. Although use of PMTCT test results would not dramatically change HIV prevalence estimates among pregnant women, the impact of site-level differences on surveillance models should be evaluated before these data are used to replace or complement ANC surveillance surveys.