The European journal of health economics : HEPAC : health economics in prevention and care · 2026
The application of excess hazard models for plausible long-term extrapolation in cost-effectiveness: a re-evaluation of immunotherapy in advanced non-small cell lung cancer using the KEYNOTE-024 trial.
СтатьяНазвание и аннотация: Перевод пока не подготовлен
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Kim Kun, Sweeting Michael, Wilking Nils, Jönsson Linus
Аннотация
BACKGROUND: Economic evaluations often rely on immature data to evaluate cost-effectiveness of new treatments. Pembrolizumab was approved for advanced non-small cell lung cancer with high PD-L1 expression based on interim data from the KEYNOTE-024 trial. Using 5-year follow-up data, we re-evaluated the cost-effectiveness analysis to assess optimal survival models and their impact on the incremental cost-effectiveness ratio (ICER).
METHODS: A partitioned survival model was developed from the national payer's perspective. Standard parametric distribution (SPD) models and excess hazard (EH) models were fitted to digitally reconstructed overall survival (OS) and progression-free survival (PFS) data from the 11-month interim and 5-year follow-up data from KEYNOTE-024. EH models incorporated general population mortality (GPM) rates to ensure stability and consistency in modelling OS and PFS. Model inputs were primarily derived from the prior assessment by the Swedish Health Technology Assessment agency. Uncertainties around survival models and inputs were examined in probabilistic sensitivity analyses (PSA).
RESULTS: We demonstrated that incorporating the relative survival framework, using EH models, provided a more suitable modelling approach than SPD models when applied to immature survival data from a clinical trial. The SPD exponential models produced overly pessimistic projections, while the SPD log-normal models produced overly optimistic projections. The EH log-normal models without a cure assumption aligned closely with the follow-up data, producing realistic and consistent projections across interim and follow-up data. Scatter plots from the EH models exhibited reduced variability than the SPD models in PSA.
CONCLUSIONS: Model selection should prioritize realistic survival projections with the use of EH models that incorporate GPM adjustment potentially preferred over SPD models. Models that accommodate waning disease-specific mortality with or without a cure assumption should be considered.
METHODS: A partitioned survival model was developed from the national payer's perspective. Standard parametric distribution (SPD) models and excess hazard (EH) models were fitted to digitally reconstructed overall survival (OS) and progression-free survival (PFS) data from the 11-month interim and 5-year follow-up data from KEYNOTE-024. EH models incorporated general population mortality (GPM) rates to ensure stability and consistency in modelling OS and PFS. Model inputs were primarily derived from the prior assessment by the Swedish Health Technology Assessment agency. Uncertainties around survival models and inputs were examined in probabilistic sensitivity analyses (PSA).
RESULTS: We demonstrated that incorporating the relative survival framework, using EH models, provided a more suitable modelling approach than SPD models when applied to immature survival data from a clinical trial. The SPD exponential models produced overly pessimistic projections, while the SPD log-normal models produced overly optimistic projections. The EH log-normal models without a cure assumption aligned closely with the follow-up data, producing realistic and consistent projections across interim and follow-up data. Scatter plots from the EH models exhibited reduced variability than the SPD models in PSA.
CONCLUSIONS: Model selection should prioritize realistic survival projections with the use of EH models that incorporate GPM adjustment potentially preferred over SPD models. Models that accommodate waning disease-specific mortality with or without a cure assumption should be considered.