P17.38.B THE USE OF EORTC QLQ-C30 SUMMARY SCORE IN CANCER CLINICAL TRIALS AND ITS PERFORMANCE AS COMPARED WITH THE GLOBAL HEALTH / QUALITY OF LIFE SCALE: A COMPARATIVE ANALYSIS OF EFFECT SIZES

Le résumé fourni par la source

Abstract BACKGROUND The EORTC QLQ-C30 is the most commonly used questionnaire for assessing general aspects of health-related quality of life (HRQoL) in patients with brain tumors. The summary score provides an overall measure of HRQoL based on 27 of the 30 QLQ-C30 items. We evaluated the use of the summary score in cancer clinical studies, including brain tumor studies, and analyzed its construct validity as compared to the 2-item global health status/quality of life (GH/QoL) scale of the QLQ-C30. MATERIAL AND METHODS We systematically searched 11 databases for original articles reporting on the validation or use of the QLQ-C30 summary score in adult cancer patients. Data were extracted and cross-checked by two reviewers using DistillerSR software. We collected information on study characteristics, use of the summary score, and use of other HRQoL domains. To assess construct validity, we compared the sensitivity of the summary score and the GH/QoL scale to detect cross-sectional and longitudinal differences, using a comparative analysis of their reported Cohen’s d effect sizes. We drafted a Bland-Altman plot of absolute Cohen’s d values to assess agreement and correlation between the summary score and the GH/QoL scale. Differences between both were assessed using a linear mixed-effects model with study as a random effect to account for within-study correlation. RESULTS We included 249 studies conducted from the first validation of the summary score in 2016 up to 2024, with 92 (37%) using the summary score as the main HRQoL domain of interest. Eighteen studies (29 comparisons) reported Cohen’s d values for both the summary score and the GH/QoL scale and were included in the comparative analysis. The Bland-Altman plot showed that the mean difference between the two constructs and its 95% confidence interval were above zero, indicating that the summary score had a structurally larger Cohen’s d value than the GH/QoL scale. This was confirmed by the paired t-test (mean difference 0.17; p = 0.018). CONCLUSION The QLQ-C30 summary score is increasingly used as the main HRQoL domain of interest in cancer clinical studies. Comparative analysis suggests that the summary score of 27/30 items from the QLQ-C30 questionnaire offers greater sensitivity than the 2-item GH/QoL scale in detecting cross-sectional group differences and changes in HRQoL over time. When evaluating overall HRQoL in brain tumor trials, the QLQ-C30 summary score should be considered the preferred endpoint.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.