The Implementation of NICE Severity Modifier in the Single Technology Appraisal of Mirvetuximab Soravtansine:...

NICE's severity modifier uses absolute and proportional QALY shortfalls to determine whether higher cost-effectiveness thresholds should apply. We use the appraisal of mirvetuximab soravtansine in platinum-resistant ovarian cancer to show how modelling choices can affect the severity weight and to draw out practical recommendations for future appraisals.

We reviewed the company and EAG approaches to estimating QALY shortfalls, focusing on utility values, starting age and standard-care overall survival extrapolation. We considered how these assumptions changed expected QALYs under standard care, the resulting QALY shortfalls and the severity weight applied.

The severity weight was sensitive to all three assumptions. The EAG's preferred utility values increased expected QALYs under standard care and reduced the severity weight from 1.7 to 1.2 in the EAG base case. Starting age was also influential, with older ages making the higher severity threshold less likely to be met. Survival extrapolation affected expected QALYs under standard care; longer-tailed curves reduced absolute and proportional QALY shortfalls, whereas shorter-tailed curves made the higher severity weight more likely.

Severity-modifier decisions can change under plausible modelling assumptions, particularly when estimates are close to the threshold between weights. When severity weights are used, appraisals should include a standard sensitivity analysis reporting the absolute QALY shortfall, proportional QALY shortfall and resulting severity weight under alternative assumptions. Clearer reporting of these calculations would make the basis for severity-weight decisions easier to scrutinise across appraisals.