Evaluation of Automated Processing of Electronically Reported Serological Tests for Syphilis Using Current and Historical Syphilis Results Compared with Traditional Reactor Grid Processing in Florida

Abstract

Abstract Background Syphilis in Florida increased 49% from 2016-2020. Moreover, many serological tests for syphilis (STS) do not indicate current infection. Traditionally, syphilis surveillance systems used reactor grids, a method for prioritizing STS for investigation based on age, non-treponemal titer, and/or sex. In 2022, Florida’s STD surveillance system implemented an automated method for processing electronically reported STS (eSTS), expanding upon the reactor grid, using an individual’s current STS (treponemal and non-treponemal), treatment history, and historical STS results aiming for more efficiently processing eSTS. We compared the new method of processing eSTS results against the reactor grid and determined potential value in time/cost savings of this change. Methods All eSTS (n = 4,144) from 1/2/2023–1/8/2023 were compared by how the logic-based method processed test results vs. how the reactor grid processed test results. Each method was compared using measurements of accuracy (e.g., sensitivity/specificity). Time and cost savings in eSTS processing were estimated. Results Using the surveillance case definition as reference, the accuracy of the logic-based method for processing eSTS was nearly double (82.3% vs. 43.6%), had greater specificity (79.0% vs. 33.0%), and increased positive predictive value (47.5% vs. 22.0%) when compared to the reactor grid method. Sensitivity (99.5% vs. 98.6%) and negative predictive value (99.9% vs. 99.2%) remained similar. The logic-based method is estimated to save 7,783 hours annually (~$185,000). Conclusions Processing eSTS based on current and historical STS results is significantly more accurate than using a reactor grid. Moreover, these improvements save time and resources that can be better allocated to other program prevention activities.