OBJECTIVES: The Institute for Clinical and Economic Review (ICER) conducted 3 economic evaluations over the past decade for treatments for nonalcoholic steatohepatitis (NASH) in the United States. Our objective was to review the cost-effectiveness (CE) modeling approaches, data sources, and assumptions across ICER evaluations and identify implications and recommendations for future CE analyses in NASH.
METHODS: We reviewed the CE analyses from the final evidence reports for the 2016, 2020, and 2023 ICER evaluations for NASH treatments and extracted details on model structures, natural history sources, efficacy endpoints, treatment discontinuation and stopping rules, cost and utility sources, and CE results.
RESULTS: The CE models used similar Markov-based structures with health states defined by fibrosis stages and cirrhosis complications. NASH resolution was included in 2016 only, and cardiovascular events were included in 2020 and 2023. Natural history sources were similar across evaluations. Treatment efficacy endpoints shifted from NASH resolution without worsening of fibrosis (2016) to improvement and worsening of fibrosis (2020 and 2023). Treatment discontinuation and stopping rules varied significantly (2016: no discontinuation, treat until F4; 2020: no discontinuation, treat until decompensated cirrhosis or hepatocellular carcinoma; 2023: 17% discontinuation per year, treat until F4). Cost and utility sources evolved over time as NASH-specific estimates became available. When comparing obeticholic acid with the standard of care (in common across analyses), incremental CE ratios decreased (from $2,748,300 to $568,000 per quality-adjusted life-year gained) and value-based pricing estimates increased (from $5,100 to $38,200 per year) with each successive evaluation.
CONCLUSIONS: Among the changes in modeling approaches, data sources, and assumptions across ICER evaluations in NASH, differences in treatment discontinuation and stopping rule assumptions appeared to have the greatest impact on economic value. Future CE analyses of treatments for NASH should use clinically realistic discontinuation rates and stopping rules and include sensitivity analyses on these influential parameters.
