Minimal clinically important difference for the EORTC QLQ-C30 and QLQ-BN20 questionnaires in glioblastoma patients using several distribution and anchor-based methods
Résumé
Many researches have been conducted in the last decades on the determination of the minimal clinically important difference
(MCID) for health-related quality of life (HRQOL) scores. Several methods have been proposed, generally classified in distribution or
anchor-based methods. A joint use of both distribution and anchorbased approaches, using several distributions and anchors, should be recommended to study the consistency and reliability of the results. Our objective was to explore and to compare seven
approaches from the distribution and anchor based methods of the MCID in EORTC questionnaires using a small sample size, to
highlight the advantages and disadvantages of each of them and the consistency of the results. Methods: Data from a randomized phase II clinical trial in unresectable glioblastoma patients were used. Patients completed the EORTC QLQ-C30 questionnaire and the QLQ-BN20 brain cancer module at baseline and during the treatment. Four distribution-based approaches were applied using
baseline scores: 0.2 standard deviation (SD), 0.3 SD, 0.5 SD and the standard error of measurement. Three anchors were used: the patient rating of change (PRC), global HRQOL change (GHC) using item 30 of the QLQ-C30 and the Karnofsky performance status (PS). Correlation between HRQOL and the anchors were checked. Results: Among the 134 patients involved in this study,
102 patients (76 %) had at least one HRQOL score available at baseline and were included in the analysis based on the distribution.
For the distributions, the minimal and maximal MCID were (4–18) and (2–16) for the QLQ-C30 and QLQ-BN20 questionnaires
respectively. Regarding the anchor, the role functioning dimension of the QLQ-C30 questionnaire was correlated with both PS and
GHC anchors (r[0.30). The MCID calculated for this dimension was 5 points for deterioration and 2 points for improvement.
Conclusions: The strength of our study is the diversity of anchors explored in order to estimate the MCID (evaluated by the patient
himself and one evaluated by the clinicians), but the sample size is challenging in this context of poor prognostic disease using anchor based approaches. These results will be compared to other published results for brain cancer tumor with QLQ-BN20.