The SCORPIO-PEP randomized trial demonstrated that ensitrelvir reduced the incidence of symptomatic COVID-19 after household exposure. The economic value of using ensitrelvir as post-exposure prophylaxis (PEP) for high-risk household contacts in Japan had not been assessed. The study aimed to estimate the cost-effectiveness of ensitrelvir PEP from the Japanese public healthcare payer perspective using a decision-analytic cost-utility model.
The authors developed a decision-analytic cost-utility model that linked trial-observed risks of symptomatic COVID-19 through day 10 to Japanese data on hospitalization, mortality, healthcare costs, quality-of-life weights, and life expectancy. Because the SCORPIO-PEP trial observed no COVID-19–related hospitalizations or deaths, the model did not include trial-observed severe outcomes; instead, any downstream reductions in hospitalization or mortality were modeled only as potential consequences of preventing symptomatic disease.
Inputs and assumptions were drawn from the trial and published Japanese sources. The analysis took the perspective of the public healthcare payer in Japan. Multiple analytic approaches were applied: deterministic sensitivity analyses, probabilistic sensitivity analysis (Monte Carlo simulation), structural scenario analyses, subgroup analyses, and threshold analyses to evaluate robustness and identify key drivers of cost-effectiveness.
In the primary analysis, treating a single exposed high-risk household contact with ensitrelvir increased expected costs by JPY 54,945 and yielded an incremental gain of 0.003267 QALYs. These results produced an incremental cost-effectiveness ratio (ICER) of JPY 16.82 million per QALY gained.
The base-case conclusion therefore was that ensitrelvir PEP, at current pricing and under the model’s assumptions, would not be considered cost-effective using a commonly cited Japanese threshold of JPY 5 million per QALY.
A probabilistic sensitivity analysis comprising 10,000 simulations was performed to account for parameter uncertainty. None of these 10,000 simulations met the JPY 5 million/QALY threshold, indicating low probability of cost-effectiveness at that willingness-to-pay level given the model inputs.
Deterministic and structural sensitivity analyses explored alternative assumptions about post-acute utility loss and attribution of downstream severe outcomes. These scenario analyses assessed how much results would change under more conservative or more optimistic parameter choices.
When the model applied the overall-trial relative risk to the high-risk baseline population, the estimated ICER was JPY 19.24 million per QALY. The analysis also examined the impact of assumptions about post-acute (longer-term) utility losses after symptomatic disease:
These scenario results show that the model’s estimates are sensitive to assumptions about the duration and magnitude of post-acute quality-of-life decrements, and that conservative assumptions substantially increase the ICER.
The analysis identified a drug acquisition-cost threshold at which ensitrelvir PEP would be cost-effective under the modeled assumptions: JPY 11,021 per course. At prices above this threshold, ensitrelvir would not meet the JPY 5 million/QALY benchmark given other input parameters in the model.
The authors conclude that, at current Japanese costs and under a range of conservative efficacy and post-acute utility assumptions, ensitrelvir used as post-exposure prophylaxis for high-risk household contacts is unlikely to be cost-effective at a willingness-to-pay threshold of JPY 5 million per QALY.
A key limitation is that the clinical trial observed no COVID-19–related hospitalizations or deaths, so any benefits on severe outcomes are model-based extrapolations rather than direct trial-observed effects. The unfavorable cost-effectiveness conclusion persisted across multiple conservative scenarios, including alternative assumptions about hospitalization attribution and mortality.
All data used in the analysis were aggregate, publicly available sources and published reimbursement information. The authors made all model inputs, analysis outputs, and the complete Python model publicly available on GitHub and in the Supplementary Materials. No individual-level patient data were used.
Under the inputs and assumptions reported in the source study, decision makers in Japan seeking interventions that meet a JPY 5 million/QALY threshold would find ensitrelvir PEP for high-risk household contacts unlikely to be cost-effective at current acquisition prices. The model highlights the importance of drug price, the magnitude and duration of post-acute quality-of-life effects, and direct evidence of severe-outcome reduction in determining value for money. Any reassessment of cost-effectiveness would require different prices, new evidence on severe outcomes, or alternative willingness-to-pay thresholds.