The clinical importance of changes in the 0 to 10 numeric rating scale for worst, least, and average pain intensity: analyses of data from clinical trials of duloxetine in pain disorders

Abstract

Data on 1,700 patients pooled from 5 randomized, placebo-controlled duloxetine studies (3 in diabetic peripheral neuropathic pain and 2 in fibromyalgia) were analyzed to determine clinically important differences (CIDs) in the 0 to 10 Numeric Rating Scale-Pain Intensity (NRS-PI) for patient-reported "worst" and "least" pain intensity while validating the previously published level for "average" pain. The correspondence between the baseline-to-endpoint raw and percentage change in the NRS-PI for the worst, least, and average pain were compared to patients' perceived improvements at endpoint as measured by the 7-point Patient Global Impression of Improvement (PGI-I) scales. Stratification by baseline pain separated the raw but not the percent change scores. The PGI-I category of "much better" or above was our a priori definition of a CID. Cutoff points for the NRS-PI change scores were determined using a receiver operator curve analysis. A consistent relationship between the worst and average NRS-PI percent change and the PGI-I was demonstrated regardless of the study, pain type, age, sex, or treatment group with a reduction of approximately 34%. The least pain item CID was slightly higher at 41%. Raw change CID cutoff points were approximately -2, -2.5 and -3 for least, average, and worst pain respectively.

Perspective: We determined an anchor-based value for the change in the worst, least, and average pain intensity items of the Brief Pain Inventory that best represents a clinically important difference. Our findings support a standard definition of a clinically important difference in clinical trials of chronic-pain therapies.