Evaluation of the psychometric properties of the EORTC chemotherapy-induced peripheral neuropathy questionnaire (QLQ-CIPN20).

Abstract

To investigate the scale structure and psychometrics of the EORTC chemotherapy-induced peripheral neuropathy module (QLQ-CIPN20). Using confirmatory factor analyses (CFA), we tested two hypothesized scale structure models of the QLQ-CIPN20 in 473 patients with non-small cell lung cancer, 281 patients with heterogeneous cancer diagnoses, and 500 patients with colorectal cancer. We also modeled the two hypothesized models as bi-factor models. These included a general factor, in addition to the specific domain factors. Additional models were investigated with exploratory factor analysis (EFA). Known groups validity was evaluated where justified. CFA could not confirm the two hypothesized models (Model 1: CFI<0.926; TLI<0.914; RMSEA>0.077 and Model 2: CFI<0.906; TLI<0.887; RMSEA>0.105) in any of the three samples. Including a general factor to these two hypothesized models to produce a bi-factor model also did not yield satisfactory results. Using EFA, we identified four different factor structures in the three samples that were unstable due to cross loadings of the items. When scoring the QLQ-CIPN20 as a simple, additive checklist evidence was found for known groups validity in the first two samples based on Common Toxicity Criteria (CTC-AE), and in the third sample based on exposure to CIPN-inducing chemotherapy. Neither CFA nor EFA yielded support for a stable subscale structure for the QLQ-CIPN20. Scoring the questionnaire as a simple additive checklist results in acceptable validity.