Background The positive predictive value (PPV) of a screening test (ST) is a function of prevalence and ST specificity and is expected to decrease with decreasing prevalence unless ST specificity approaches 100%. Consequently, an additional test (AT) following positive STs may be indicated if prevalence is low. Our objective was to determine the impact of CT and GC prevalence on per cent AT positive following positive STs by Gen-Probe Combo 2 CT and GC using data from public clinics in the state of Mississippi.

Methods Based-on CDC's electronic prevalence monitoring databases from 2005 to 2007, we stratified 126 clinics (with >400 females tested) served by Mississippi State Public Health Laboratory (MSPHL) based on ST positivity. We calculated the per cent AT positive among 6553 CT ST positive and 1841 GC ST positive specimens. We further examined the impact of the quantitative Combo 2 GC results (relative light units (RLU)) for a sample of 508 specimens from clinics with low (<2.0%, family planning) and high (>6.0%, STD) ST positivity by abstracting the RLU values from hard copy records.

Results Per cent CT AT positive declined significantly (p<0.0001) from 96.3% for specimens from clinics with >10.0% ST positivity to 90.9% for specimens from a single clinic with <6% ST positivity (see Abstract P3-S1.11 table 1). GC ST positivity was <6% for 109 (87%) of the clinics. In spite of the lower GC ST positivity, the per cent GC AT positive was also >90%, ranging from 95.4% for GC ST positivity <2% to 97.7% for GC ST positivity 3.0%–4.0%. However, the per cent GC AT positive was not associated with GC ST positivity (p=0.17). Discordant GC AT results were confined to GC ST positive specimens with RLU <1 million (results not shown). The per cent of ST specimens with RLU <1million and the per cent AT negative among these lower RLU positives were also not associated with clinic ST positivity (p=0.14 and p=0.78, respectively).

Conclusions Performing APTIMA CT or GC ATs added little to Combo 2 ST PPV, although the decrease in per cent AT positive with decreasing ST positivity observed in this study raises concern about Combo 2 PPV at CT prevalence levels lower than 6%. The lack of impact of GC prevalence on GC ST RLU or AT results is unexpected and might indicate that the Combo 2 ST PPV is very high even at the lower GC prevalence. In other words, most negative GC AT results are false rather than true negatives and the patients should be treated.