ED syphilis and gonorrhea/chlamydia cotesting practices before and after the implementation of an electronic health record-based alert.

BACKGROUND The prevalence of syphilis is increasing in many countries, including the USA. The ED is often used by underserved populations, making it an important setting to test and treat patients who are not evaluated in outpatient clinical settings. We aimed to assess the utility of an ED-based syphilis and gonorrhoea/chlamydia cotesting protocol by comparing testing practices before and after its implementation. METHODS We implemented an electronic health record (EHR) alert that prompted clinicians to order syphilis testing in patients undergoing gonorrhoea/chlamydia testing. We performed a retrospective cohort analysis that compared outcomes between the preimplementation period (January-November 2018) and the postimplementation period (January-November 2019). Patients were tested for Treponema pallidum antibody (TPA) using a multiplex flow immunoassay (MFI), and positive results were confirmed by rapid plasma reagin (RPR). The primary implementation outcome was the number of syphilis tests/month, and the primary clinical outcome was the number of syphilis diagnoses/month (defined as positive TPA MFI and RPR). We performed an interrupted time-series analysis to evaluate the effect of implementing the alert over time. RESULTS Four-hundred and ninety-four and 1106 unique patients were tested for syphilis in the preimplementation and postimplementation periods, respectively. Syphilis testing increased by 55.6 tests/month (95% CI 45.9 to 65.3, p<0.001) following alert implementation. Patients tested in the postimplementation period who were tested using the alert were much younger (difference: 14 years (95% CI 12 to 15)) and were more likely to be female (difference: 15% (95% CI 8 to 21)) and African-American (difference: 11% (95% CI 5 to 17)) than patients tested by clinician-initiated testing. Presumptive syphilis diagnoses increased from 3.4 diagnoses/month to 7.9 diagnoses/month (difference, 4.5 (95% CI 2.2 to 6.9), p<0.001). CONCLUSIONS Our study demonstrates that use of a targeted EHR alert testing protocol can increase syphilis testing and diagnosis and may reduce clinician bias in testing.