Syphilis testing guidelines in China are usually based on symptomatic criteria, overlooking risk assessment and ultimately opportunities for disease detection and control. We used data from 10,695 sexually transmitted disease (STD) clinic patients in Guangxi, China, to assess the efficacy of a potential screening tool inquiring about behavioural and health risk factors in identifying the STD patients who should not be triaged for syphilis testing under current guidelines, but on the contrary receive such testing. Validity testing of the screening tool was performed and receiver-operating characteristic curves were plotted to determine an optimal total risk score cut-off for testing. About 40.9% of patients with positive toluidine red unheated serum test and Treponema pallidum particle agglutination test did not show hallmark signs of syphilis. The screening tool was more sensitive in detecting infection in non-triaged male versus female patients (highest sensitivity = 90% vs. 55%) and the cut-off score to warrant testing was lower in non-triaged female patients than in non-triaged male patients (cut-off = 1 vs. 2). Most of the cases were missed among female STD patients. In spite of selective testing based on behavioural and health indicators that improve case detection, cases were still missed. Our study supports universal testing for syphilis in the STD population.