OBJECTIVES: Mixture-cure models (MCMs) allow overcoming the reliability issues that traditional survival analysis methods face in the presence of statistical cure. In its recent decision, the Czech HTA body refused to accept this approach based on the assumption of different mortality rates between “cured” patients and normal population, which led to neglecting the presence of statistically cured patients and underestimating the value of evaluated medicine. Therefore, we aimed to analyze and discuss existing evidence related to the reliability and appropriateness of these models, conditions necessary for their use and their perspective within the HTA decision framework.

METHODS: In order to analyze and discuss the suitability of mixture-cure models, specifically in the context of diffuse-large B-cell lymphoma (DLBCL), we performed literature and HTA guidelines review, compared this approach with standard parametric models and put it into the perspective of current decision-making by local HTA authority.

RESULTS: Presence of “statistically cured” fraction of patients is accepted for 1L therapy of patients with DLBCL and becomes accepted for modelling of the therapeutic effects. Reliance on standard parametric models ignores the observed trends in survival of these patients, fails to provide reliable estimates for long-term extrapolation and underestimates the effects of the evaluated medicines.

CONCLUSIONS: Failed integration of novel concept of mixture-cure models into the HTA decision framework, which better reflect the natural history of the disease and improve the estimates for extrapolation of the tail of the KM curve represents an unnecessary barrier for access to novel medicines.