Worst itch numerical rating scale for prurigo nodularis: a psychometric evaluation

Background Study TR03 evaluated the safety and efficacy of nalbuphine ER for prurigo nodularis (PN; NCT02174419). Objective We conducted supplementary analyses to assess the psychometric properties of the Worst Itch Numeric Rating Scale (WI-NRS), the TR03 primary endpoint. Methods Study TR03 was a double-blind, placebo-controlled, phase 2 trial in PN patients with documented scores >= 5 on the WI-NRS (0 [no itch]-10 [worst itch imaginable]) on >= 5 of 7 days before baseline. Using TR03 data, the WI-NRS's psychometric properties, including reliability, validity and ability to detect change, were evaluated. A responder threshold was estimated to facilitate interpretation of WI-NRS score changes. Results Amongst 62 treated patients, improvements in mean [SD] (median) WI-NRS scores were observed between baseline (8.2 [1.21] (8.1)) and week 10 (5.8 [2.43] (6.0)). The WI-NRS had an intraclass correlation coefficient of 0.96 (95% confidence interval, 0.93-0.98) in 42 patients who had stable Itch verbal rating scale (VRS) scores from week 9-10, supporting strong test-retest reliability. Construct validity was supported, with strong correlations at week 10 with Average Itch NRS (r = 0.87) and Itch VRS single-day/weekly mean scores (r = 0.81/0.89) and moderate correlations with ItchyQoL (TM) total/domain scores (r = 0.41-0.43). The WI-NRS discriminated between predefined severity subgroups based on the Itch VRS and detected changes in itching severity (effect-size estimate: -2.05; standardized response mean: -1.21). An anchor-based threshold based on a two-category improvement in the single-day Itch VRS suggests a responder threshold of >= 3.8 points (similar to 40% improvement). Conclusions The WI-NRS demonstrates good measurement properties, supporting its use in evaluating treatment change in PN.