查看文章

Background

Syphilis is becoming more prevalent and is diagnosed primarily through serologic tests. Confirmatory treponemal tests (TTs), such as Treponema pallidum particle agglutination (TPPA) and fluorescent treponema antibody absorption (FTA-Abs), are typically used. Although only one TT is needed for syphilis diagnosis, multiple TTs are still common in the disease course. Discrepant TT results can cause confusion and delay treatment. We aimed to explore the clinical characteristics of patients with discrepant TT results and develop a machine learning based tool to evaluate the risk of TT discrepancies.

Methods

In the retrospective cohort study, we linked electronic health records from a medical centerand national claims records in Taiwan between January 2001 and September 2018. Medical histories and demographic characteristics were considered as variables of interest to identify risk factors and …