P055 MARKOV CHAIN ECONOMIC IMPACT MODEL FOR ANTI-TNF NON-RESPONSE STRATIFICATION OF PATIENTS WITH MODERATE TO SEVERE ULCERATIVE COLITIS

Abstract

Abstract Objective The annual cost of ulcerative colitis (UC) in the US is estimated to be as high as $14.9 billion, of which therapeutics account for up to 45%. The mean annual drug cost per patient with moderate to severe UC is $28,586. However, only 49% of patients achieve endoscopic healing on their first anti-tumor necrosis factor (anti-TNF) therapy and up to half of UC patients on biologics discontinue therapy within 12 months of initiation which leads to high rates of hospitalization and surgery. We performed a health care payer budget impact analysis to assess the influence of a precision medicine RNA-based classifier (PrismUC) that predicts, at baseline, anti-TNF non-response in patients with moderate to severe UC. Methods We developed a 12-month budget impact model for a 1-million-member health plan by analyzing peer-reviewed literature of treatment patterns in commercial and Medicare populations moderate to severe UC patients. Inputs included drug and medical costs, UC prevalence, biologic assignment patterns, therapy response rates, therapy persistence, therapy cycling, endoscopic healing rates, and risk of surgery. Drug and medical costs were obtained from published literature. A Markov model was created to estimate indirect effects of treatment efficacy, timing and cost of disease progression in subjects identified as anti-TNF non-responders (patients who fail to achieve endoscopic healing). Non-responders were assigned either to an alternative first-line therapy (vedolizumab) or to the standard of care (SOC) as determined by clinician judgments and payer formularies. The SOC included cycling patients through various biologics such as anti-TNF, vedolizumab, tofacitinib and ustekinumab. The cost of SOC was compared that of implementing a classifier to stratify patients by primary non-response to anti-TNF. Final cost savings included the cost of drugs, hospitalizations, and surgeries. Results Introducing a classifier that identifies anti-TNF non-responder UC patients to a 1-million-member health plan resulted in annual cost savings ranging from $6.8 to $9.6 million total (Table 1) within a 12-month period with and $9.0 to $12.7 thousand per patient (Table 2), depending on classifier performance metrics. Conclusion The PrismUC precision medicine classifier test that identifies non-responders to anti-TNF therapy will provide significant economic savings to payers by decreasing expenditures on unnecessary drugs and shortening times to endoscopic healing. These savings provide an economic incentive for payers to proactively engage with diagnostics manufacturers and encourage development of such tools to improve care management. A subsequent classifier that predicts response to therapies with alternative mechanisms of action is in development and expected to yield even higher cost savings.