S2512 Clinical Utility of Algorithmic Detection of Diagnosed but Untreated HCV

Introduction: Hepatitis C represents a significant cause of morbidity and mortality with 0.9% of US adults estimated to have a positive HCV viral load. Around 85% of those exposed to HCV develop a chronic infection, which can lead to complications such as cirrhosis and liver cancer. There are currently several once-daily oral treatment options, and nearly 100% of patients attain sustained viral response (SVR) with proper treatment. The CDC reports that up to one-third of patients with a positive HCV antibody screen lack a follow-up HCV viral load PCR, and only around one-third of patients receive an appropriate treatment within 1 year. To identify patients diagnosed with chronic HCV without confirmed treatment, researchers at Mount Sinai developed an algorithm that utilizes Java and structured query language to screen for adults in EPIC with positive HCV-RNA testing without subsequent prescription of a directly acting agent (DAA). This study aims to evaluate the clinical utility of this algorithm. Methods: We conducted a retrospective cohort study by screening 3500 subjects with the above algorithm for untreated HCV infection between January 2008 and December 2020. Patients aged 18 years or older with PCR-confirmed HCV who did not receive any form of HCV treatment were included in this study. Patients identified by the algorithm underwent manual chart review, and once confirmed to have diagnosed but untreated HCV, they were contacted for further evaluation and treatment. Results: Of the 3,500 patients screened, 155 patients were initially identified by the algorithm, with 83 cases verified as truly diagnosed but untreated chronic HCV on manual review. The 72 patients excluded had documentation of treatment from outside providers. Over 1 year, 14 patients identified are now cured and 2 have been evaluated pending initiation of treatment. Conclusion: Given that 53% of the patients flagged by the algorithm were truly untreated, it is recognized that the algorithm requires some degree of manual supervision. The primary issue identified is that the algorithm does not recognize documentation of treatment from outside providers. Regardless, this algorithm can be an effective tool for flagging cases to review and presents an opportunity to make a significant clinical impact by identifying patients with diagnosed but untreated chronic HCV. Further development of similar clinical aides and integration into EMRs can provide a powerful tool for panel management and improved clinical outcomes.