OBJECTIVES: To estimate utility values based upon a proximity to death approach in patients with untreated non-squamous metastatic non-small-cell lung cancer (NSQ-mNSCLC), and to compare these values across the following country tariffs, United Kingdom, France, Sweden, Portugal, Canada, and Australia.

METHODS: EQ-5D-3L was collected in IMpower150, a phase III randomized controlled trial of 1,202 patients with NSQ-mNSCLC. The questionnaire was completed every 21 days during the treatment period and at 3 months and 6 months after disease progression. The index scores for each subject were calculated using tariffs from UK, France, Sweden, Portugal, Canada and Australia. A proximity to death approach was chosen since it reflects the decline in quality of life of that patient experience at the end of life. The time to death utilities were analysed separately for patients who were on treatment and patients who discontinued treatment. The reason of discontinuation was not accounted in this analysis. A repeated measurements model using unstructured correlation structure was assumed. Where convergence issues were observed, an exchangeable correlation structure was used.

RESULTS: A total of 15,573 observations were reported and 14,542 used for the analyses. For measurements taken during on treatment period, a decline in the mean utility score was observed approximately 6.6 months before death irrespective of the EQ-5D tariff used. For the measurements taken after patients discontinued treatment, their proximity to death curve showed a steeper decline in the mean utility score with the onset of decline being closer to death compared to the on-treatment utility.

CONCLUSIONS: The proximity to death methodology demonstrated the ability to capture the continuous decline of cancer patients’ HRQoL towards the end of life, which may be a more appropriate method for immunotherapies than standard methods. This conclusion held true independent of the country-specific tariff used.