OBJECTIVES : In a placebo-controlled trial assessing the efficacy of regorafenib in patients with hepatocellular carcinoma (HCC) who progressed on sorafenib treatment, regorafenib showed improved survival but increased toxicity. We demonstrate the use of QTWIST analysis as an alternative to traditional economic modelling of the trial data for estimating QALYs.

METHODS : The regorafenib trial randomized 379 subjects to active treatment and 194 to matching placebo in 28-day cycles. Quality of life was assessed using the EuroQol-5 Dimension (EQ5D) instrument. Adverse events (AEs) were assessed using the National Cancer Institute (NCI) Common Terminology Criteria for Adverse Events (CTCAE). Only events of severity 3 (severe) or severity 4 (disabling/life-threatening) affecting ≥ 5% of subjects in each arm were considered in the analysis because of their greater expected impact on quality of life. Disutilities associated with disease progression and AEs were estimated using a mixed model. QTWIST was calculated as a weighted sum of the utility of each health state (pre-progression, toxicity, post progression) multiplied by the length of time spent in each health state. We considered AEs first as a composite variable and next as disaggregated individual AE variables.

RESULTS : Incremental improvement in QALYs was 0.428 months for the composite AE variable and 0.431 when individual events were considered. These results are comparable to a previous parametric survival analysis on the same trial data.

CONCLUSIONS : We demonstrate that when an appropriate cardinal utility measure is used to assess quality of life, QTWIST analysis produces QALYs comparable to economic models and may be preferred by agencies that are skeptical of decision models.