Comparative responsiveness of verbal and numerical rating scales to measure pain intensity in patients with chronic pain.

UNLABELLED: Verbal rating scale (VRS) and numerical rating scale (NRS) are regularly used to assess and monitor pain in chronic pain patients. Although the NRS has been generally preferred, limited comparative responsiveness evidence was reported. This study compared the responsiveness of VRS and NRS measuring current pain and investigated the influence of different references (ie, worst, least, average, and current pain or their composite) on the NRSs' responsiveness. Two hundred fifty-four chronic pain patients attended a 10-day pain self-management program and were assessed with two 6-point VRSs (assessing current pain) and four 11-point NRSs (assessing worst, least, average, and current pain) at pre- and posttreatment. A patient-reported rating of pain improvement was used as the criterion for standardized response mean and receiver operating characteristic curve analyses. Results showed that the VRSs and NRSs exhibited small responsiveness in all patients, but the magnitude of responsiveness became moderate to large in patients with improved pain. However, in patients with pain improvements, the NRS current pain item and composite score (made up of the 4 pain items) were found to have significantly larger responsiveness and greater discriminatory ability to detect the presence of improvement than other current pain VRSs and the NRSs assessing worst, least, and average pain. Potential implications for clinical practice are discussed. PERSPECTIVE: This study shows that the current pain and composite NRSs were more responsive than the current pain VRSs and the NRSs measuring other individual pain references in patients with improved pain, undertaking a short-term, self-management program. The results help inform the selection of pain intensity measures in studies using similar types of intervention.