Minimal clinically important differences in the brief pain inventory in patients with bone metastases

Purpose The brief pain inventory (BPI) is often used to assess pain and functional interference as a result of pain in cancer patients. Minor improvements or deteriorations in BPI may be statistically significant due to large sample sizes but may not necessarily be clinically relevant. The purpose of this study was to determine the minimal clinically important differences (MCID) in the functional BPI in patients with pain due to bone metastases. Methods BPI scores were collected from patients with painful bone metastases who visited the Rapid Response Radiotherapy Program for palliative radiotherapy. Pain and functional interferences scores were also collected monthly for three months. Patients were categorized into “complete or partial response,” “pain progression,” and “indeterminate response” based on their pain scores as recommended by the latest consensus definitions. Anchor-based determination of MCIDs of functional interference scores was calculated by determining the difference between the mean follow-up scores and the mean baseline scores for patients from each of the three response groups. Distribution-based estimates were obtained utilizing 0.2, 0.3, and 0.5 standard deviation (SD) effect sizes and the standard error of measurement. The anchor-based method results were compared with the distribution-based method results. Results Statistically significant MCIDs were determined for all of the functional interference items of BPI for patients with “complete or partial response”; whereas, no statistically significant MCIDs in BPI scores could be determined for patients with “pain progression.” Some of the functional interference items of BPI had statistically significant MCIDs for patients with “indeterminate response,” although these were generally smaller than patients with complete or partial response. Using the distribution-based approach, an effect size of 0.5 SD was the closest estimate for determining the MCID for both patients with complete or partial response and those with indeterminate response. Conclusions The MCIDs determined for pain improvement were rather large, where as statistically significant MCIDs could not be detected for pain deterioration. Knowledge of MCIDs utilizing the BPI will allow physicians to evaluate the impact of treatment (or no treatment) on a patient’s functional abilities. Knowledge of MCIDs may allow for sample size determination in future clinical trials.