In Search of a Gold Standard Patient-Reported Outcome Measure for Use in Chemotherapy- Induced Peripheral Neuropathy Clinical Trials

Purpose: To test a reduced version—CIPN15—of the European Organisation for Research and Treatment of Cancer (EORTC) Quality of Life Questionnaire Chemotherapy-Induced Peripheral Neuropathy scale (QLQ-CIPN20) to establish a possible gold-standard patient-reported outcome measure for chemotherapy-induced peripheral neuropathy (CIPN). Methods: Using a prospective, longitudinal, case–control design, patients (n = 121) receiving neurotoxic chemotherapy completed the CIPN15 at baseline and 12 weeks and underwent objective neurological assessment using the 5-item Total Neuropathy Score-Clinical (TNSc). Healthy controls (n = 30) completed the CIPN15 once. Structural validity was evaluated using factor analysis. Because a stable factor structure was not found, a sum score was used to evaluate measures of the CIPN15’s psychometric properties—reliability, validity, sensitivity, and responsiveness—as follows: internal consistency via Cronbach’s α and item–item correlations; test–retest reliability via correlation between 2 CIPN15 scores from each patient; concurrent validity via correlation between CIPN15 and 5-item TNSc scores; contrasting group validity via comparison of CIPN15 scores from patients and healthy controls; sensitivity via descriptive statistics (means, standard deviation, ranges); and responsiveness via Cohen’s d effect size. Results: Most patients received single agent oxaliplatin (33.7%), paclitaxel (21.2%), or more than 1 neurotoxic drug concurrently (29.8%). Factor analysis revealed no stable factor structure. Cronbach’s α for the CIPN15 sum score was 0.91 (confidence interval [CI] = 0.89-0.93). Test–retest reliability was demonstrated based on strong correlations between the 2 scores obtained at the 12-week time point ( r = 0.86; CI = 0.80-0.90). The CIPN15 and 5-item TNSc items reflecting symptoms (not signs) were moderately correlated ( r range 0.57-0.72): concurrent validity. Statistically significant differences were found between patient and healthy control CIPN15 mean scores ( P < .0001): contrasting group validity. All items encompassed the full score range but the CIPN15 linearly converted sum score did not: sensitivity. The CIPN15 was responsive based on a Cohen’s d of 0.52 (CI = 0.25-0.79). Conclusion: The sum-scored CIPN15 is reliable, valid, sensitive, and responsive when used to assess taxane- and platinum-induced CIPN.