Minimal clinically important differences in the EORTC QLQ-BM22 and EORTC QLQ-C15-PAL modules in patients with bone metastases undergoing palliative radiotherapy

Srinivas Raman, Keyue Ding, Edward Chow, Ralph M. Meyer, Abdenour Nabid, Pierre Chabot, Genevieve Coulombe, Shahida Ahmed, Joda Kuk, A. Rashid Dar, Aamer Mahmud, Alysa Fairchild, Carolyn F. Wilson, Jackson S. Y. Wu, Kristopher Dennis, Carlo DeAngelis, Rebecca K. S. Wong, Liting Zhu, Michael Brundage

Validated tools for evaluating quality of life (QOL) in patients with bone metastases include the EORTC QLQ-BM22 and QLQ-C15-PAL modules. A statistically significant difference in metric scores may not be clinically significant. To aid in their interpretation, we performed analyses to determine the minimal clinically important differences (MCID) for these QOL instruments. Both anchor-based and distribution-based methods were used to determine the MCID among patients with bone metastases enrolled in a randomized phase III trial. For the anchor-based approach, overall QOL as measured by the QLQ-C15-PAL module was used as the anchor and only the subscales with moderate or better correlation were used for subsequent MCID analysis. In the anchor-based approach, patients were classified as improved, stable or deteriorated by the change in the overall QOL score from baseline to follow-up after 42 days. The MCID and confidence interval was then calculated for all subscales. In the distribution-based approach, the MCID was expressed as a proportion of the standard deviation and standard error measurement from the subscale score distribution. A total of 204 patients completed the questionnaires at baseline and follow-up. Only the dyspnea and insomnia subscales did not have at least moderate correlation with the overall QOL anchor. Using the anchor-based approach, 10/11 subscales had an MCID score significantly different than 0 for improvement and 3/11 subscales had a significant MCID score for deterioration. The magnitude of MCID scores was higher for improvement in comparison with deterioration. For improvement, the anchor-based approach showed good agreement with the distribution-based approach when using 0.5 SD as the MCID. However, there was greater lack of agreement between these approaches for deterioration. We present the MCID scores for the EORTC QLQ-BM22 and QLQ-C15-PAL QOL instruments. The results of this study can guide clinicians in the interpretation of these instruments. NCT01248585.