Hepatitis C Elimination as a Test Case for AI in Care Delivery

Hepatitis C (HCV) is routinely curable, yet roughly two-thirds of Americans living with HCV remain untreated. The central obstacle is not therapeutic efficacy but a fragmented care delivery cascade with failures at each step along a continuum spanning diagnosis, linkage to care, starting treatment, and confirming cure. In this article, we map practical opportunities for artificial intelligence at each transition and make the case that HCV offers an unusually tractable test of whether AI can improve the delivery of treatments already known to work. The strongest current evidence supports AI-assisted case finding and retrieval of patients lost to follow-up, while proposed roles in treatment and closed-loop follow-through require careful prospective evaluation. The approach should build on universal screening, reflex and point-of-care testing, low-barrier care, and trusted human navigation, and its value should be judged by whether it shifts human effort from administrative work toward patient relationships. We propose multi-state demonstrations with a staged AI layer, prespecified success and failure thresholds, subgroup equity measures, and monitoring of administrative versus patient-facing time. Findings and lessons from an HCV test case could serve as a template for other conditions in which effective treatment exists but fragmented delivery impedes population health impact at scale.