Establishment of cutpoints to categorize the severity of chronic pain using composite ratings with Rasch analysis

Abstract

Background: Establishment of cutpoints for classifying mild, moderate and severe pain is commonly based on single rating of worst or average pain. However, single pain measure may serve as a brief and partial surrogate for composite pain ratings. This study aimed to base composite pain ratings to establish optimal cutpoint that maximized the difference of pain interference on daily function and compare its utility with those based on single worst and average pain. Methods: Data were from a cohort study of 322 patients with chronic pain. Brief pain inventory (including four items measuring the least, worst, average and current pain) was administered. Rasch analysis and Serlin et al.' s (Pain, 61, 1995, 277) method were used to derive optimal cutpoint. Results: Rasch analysis calibrated the least, worst, average and current pain items into a unidimensional hierarchy and produced composite pain measurement. The optimal cutpoint for composite pain (mild, <= 4; moderate, > 4 - 6; severe, > 6 - 10 on the 0 - 10 numeric rating scale) differed from those cutpoints for worst (<= 6; > 6-8; >8-10) and average pain (<= 5; > 5- 7; > 7-10). The optimal cutpoint for composite pain was better able than those for worst and average pain to distinguish among groups on patient-rated pain quality and quality of life. The optimal cutpoint for average pain had better discriminant ability than that for worst pain. Conclusion: The results suggest that using optimal cutpoint for composite pain may be useful to classify clinically important groups in patients with chronic pain and that average pain may be an alternative choice if a single item is used.