Impact of Overall Survival Gains on ASMR for Oncological Drugs

OBJECTIVES: The French Haute Autorité de Santé (HAS) rates the added clinical value of drugs, so-called ASMR, on a scale ranging from “I – Major Improvement” to “V – No Improvement”. The relationship between gain in overall survival (OS) and ASMR was investigated based on reports of oncological drug appraisals.

METHODS: Data were obtained from the NaviHTA database, which contains detailed extractions of the Transparency Committee’s reports published since 2018. Assessments for oncological drugs with comparators considered as appropriate by the agency and a study population overlapping with the decision population were included. An ordered logit model was fitted to predict ASMR as a function of the gain in median OS or hazard ratio. Adjustments for orphan status and variables characterizing the quality of evidence were considered. The final selection of explanatory variables was guided by expert knowledge and model performance.

RESULTS: 63 decisions were included in the regression analysis, with ASMR ranging from “III – Moderate Improvement” to “V – No Improvement”. Mean OS gain in months was 7.02 for ASMR III, 6.45 for ASMR IV and 4.88 for ASMR V. The base case model used hazard ratio (HR) of OS categorised into: ≤0.7, 0.7 - 0.8, >0.8, as a predictor, as it performed better than median OS gain. Compared to a HR ≤0.7, a HR of 0.7 – 0.8 is associated with 6.8 times higher odds of higher ASMR (p-value < 0.001), and HR >0.8 with 19.5 times higher odds of higher ASMR (p-value < 0.001). 79% of cases could be accurately classified in 2 categories (ASMR I-III vs. ASMR IV-V) based on HR. Study design and orphan drug status had no significant influence.

CONCLUSIONS: The HAS rating of clinical added value for oncological drugs with an appropriate comparator is mostly driven by the OS HR.