OBJECTIVES: Estimate EQ-5D-5L utility values for 1^st line ES-SCLC patients in the IMpower133 trial.

METHODS: Patients completed the EQ-5D-5L questionnaire electronically every 3 weeks during study treatment in the IMpower133 trial. Repeated measure linear models were fitted, first evaluating the effect of treatment (Tecentriq + Carboplatin + Etoposide vs. Carboplatin + Etoposide) on EQ-5D scores, and second, the effect on the proximity to death for the pooled population, classified in 4 groups: a) 5 weeks before death, b) > 5 to ≤ 10 weeks before death, c) > 10 to ≤ 30 weeks before death, d) > 30 weeks before death. For the model fitting, proc mixed in SAS/STAT 9.2 was used. The most recent published EQ-5D-5L tariffs were used for certain key countries.

RESULTS: 96.8% of the 403 intent-to-treat patients, completed the EQ-5D-5L at least once. In total 3199 utility index scores were computed. Utility estimates for patients while on treatment for Tecentriq + Carboplatin + Etoposide were higher to those for Carboplatin + Etoposide alone but this difference was not statistically significant. Those estimates were 0.736 vs. 0.723, 0.770 vs. 0.757, 0.730 vs. 0.725, 0.755 vs 0.750, 0.748 vs. 0.739 and 0.762 vs. 0.756 for Australia, Canada, France, Netherlands, UK and USA respectively. The proximity to death estimates were consistent with similar lung cancer studies where utilities were analyzed as a function of time. Using UK as base-case, utility estimates (95% CI) were 0.508 (0.458 – 0.558), 0.691 (0.654 – 0.729), 0.740 (0.716 – 0.765) and 0.765 (0.744 – 0.786) for time groups a, b, c, and d, respectively.

CONCLUSIONS: There is paucity of robust utility estimates in the literature for ES-SCLC, suitable to inform economic evaluation for Health Technology Assessment. This study provides robust estimates of country specific EQ-5D utilities for this disease area, assessing both treatment effect and time dependence.