Background Routine surveillance of Military Health System (MHS) data provides actionable information on STI rates. National increases in syphilis rates led to the publication of a 2015 MHS syphilis analysis demonstrating a significant rise in military syphilis cases between 2010 and 2015. The current study had two objectives 1) validate the current syphilis estimates under the DoD surveillance case definition and 2) evaluate the validity of clinical staging of syphilis cases within the surveillance period.

Methods Of the initial 2976 incident cases identified in the 2015 surveillance review, we sampled 500 cases. We developed and applied a standardized review algorithm for case determination and expert chart review to provide evidence of clinical stage of disease at the time of surveillance case capture.

Results Out of 500 total cases evaluated, 181 (36%) were determined to not be cases of syphilis. Surveillance cases identified through Reportable Medical Events (RME) had a positive predictive value (PPV) of 0.82 compared to those cases identified through administratively available (ICD9-coded) data with a PPV of 0.42. Similarly, surveillance classification of clinical staging of infection was grossly inaccurate with respect to Latent, Primary or Secondary (P&S), or Late infection with accuracy dependent on use of RME (PPV 0.49) vs ICD-9 codes (PPV 0.30) for case identification.

Conclusion A full one third of DoD surveillance case identified cases of syphilis in the Military Health system are not true cases of syphilis. The predominate cause of this misclassification was the reliance on appropriate use of ICDs by providers. The use of administrative data (ICD codes) for incidence and disease stage surveillance should be done with caution due to inappropriate use of coding, misinterpretation of labs, and overly inclusive case definitions. RMEs provide better accuracy (PPV) for correctly identifying incident cases but are still inaccurate with respect to clinical stage.

Disclosure No significant relationships.