Minimal clinically important differences in the EORTC QLQ-C30 and brief pain inventory in patients undergoing re-irradiation for painful bone metastases

PurposeThe EORTC QLQ-C30 and the Brief Pain Inventory (BPI) are validated tools for measuring quality of life (QOL) and the impact of pain in patients with advanced cancer. Interpretation of these instrument scores can be challenging and it is difficult to know what numerical changes translate to clinically significant impact in patients’ lives. To address this issue, our study sought to establish the minimal clinically important differences (MCID) for these two instruments in a prospective cohort of patients with advanced cancer and painful bone metastases.MethodsBoth anchor-based and distribution-based methods were used to estimate the MCID scores from patients enrolled in a randomized phase III trial evaluating two different re-irradiation treatment schedules. For the anchor-based method, the global QOL item from the QLQ-C30 was chosen as the anchor. Spearman correlation coefficients were calculated for all items and only those items with moderate or better correlation (|r| ≥ 0.30) with the anchor were used for subsequent analysis. A 10-point difference in the global QOL score was used to classify improvement and deterioration, and the MCID scores were calculated for each of these categories. These results were compared with scores obtained by the distribution-method, which estimates the MCID purely from the statistical characteristics of the sample population.ResultsA total of 375 patients were included in this study with documented pain responses and completed QOL questionnaires at 2 months. 9/14 items in the QLQ-C30 and 6/10 items in the BPI were found to have moderate or better correlation with the anchor. For deterioration, statistically significant MCID scores were found in all items of the QLQ-C30 and BPI. For improvement, statistically significant MCID scores were found in 7/9 items of the QLQ-C30 and 2/6 items of the BPI. The MCID scores for deterioration were uniformly higher than the MCIDs for improvement. Using the distribution-based method, there was good agreement between the 0.5 standard deviation (SD) values and anchor-based scores for deterioration. For improvement, there was less agreement and the anchor-based scores were lower than the 0.5 SD values obtained from the distribution-based method.ConclusionWe present MCID scores for the QLQ-C30 and BPI instruments obtained from a large cohort of patients with advanced cancer undergoing re-irradiation for painful bone metastases. The results from this study were compared to other similar studies which showed larger MCID scores for improvement compared to deterioration. We hypothesize that disease trajectory and patient expectations are important factors in understanding the contrasting results. The results of this study can guide clinicians and researchers in the interpretation of these instruments.