Methodological Concerns in the Medico-Economic Modeling of Adjuvant Management

OBJECTIVES:

In France, 92% of health economic (HE) appraisals in oncology therapeutic area concern the advanced stage. Specifics characteristics of adjuvant stage management raise new challenge in modelling. The objective of this study was to review the HE appraisals in France published by the CEESP and provide a summary of the methodological concerns (MCs) regarding modelling approach reported in adjuvant indication.

METHODS:

All available appraisals for adjuvant indications published until June 2022 were reviewed. A cross-analysis of identified appraisals was done to identify the modelling choices, highlighted the major issues and challenges regarding modelling of adjuvant treatment.

RESULTS:

Since June 2022, 5 HE appraisals were published. Three of them concerned treatments for melanoma, one for breast cancer and one for gastroesophageal cancer. Four of the 5 models assessed used semi-markovian approach which include type of relapse. Although these models seem more appropriate for the assessment, minor reservation was made only for 1 model on the choice of number of heath state regarding disease history and data availabilities. The last model assed used AUC approach, which lead to major reservation link to immature data and inappropriate relapse modeling.

Among the five opinions, three used external data to model survival post-recurrence, including 2 with real-world data. According to the assessment, this external data creates uncertainty about the transposability of populations. The modeling of survival post-recurrence received an important reservation in 4 appraisals. These reservations are notably due to a lack of modelling assumption in the health state "metastatic recurrence".

CONCLUSIONS:

The cross-analysis of 5 HE appraisals allowed identifying challenges to be anticipated in an adjuvant indication. Especially, the economic evaluation of adjuvant treatments raises questions and concerns relating to immature data and modelling of post-recurrence states and others more specifics to oncology, such as the sequences of treatments with greater acuity.