PCN204 Evaluation of the Item-level Psychometric Properties of EORTC QLQ-C30 and QLQ-LC13 for Patients with Non-small Cell Lung Cancer (NSCLC)

Collecting patient-reported outcomes in clinical trials requires balancing patient burden with concept relevance through psychometrically sound measures. Evaluating psychometric properties of items from completed studies may identify that can be removed future (e.g., are not statistically acceptable regarding measurement quality). This study evaluated this idea using item-level analyses the EORTC QLQ-C30 and QLQ-LC13 patients NSCLC. research used data a Cancer Immunotherapy Data Mart, comprising eight NSCLC (six phase-III, two phase-II). There were 5,274 at baseline, between 18-90 years old (median = 64) 66% males. The measure cancer patients’ functioning, quality life, disease/treatment-related symptoms. Psychometric conducted for each instrument classical test theory (CTT) item response (IRT) Rasch Partial Credit model (PCM). Internal reliability was excellent good QLQ-LC13. Each showed construct validity demonstrated unidimensionality; all individual had significant positive factor loadings. fit PCM well. All provided promising displayed adequate discrimination. Item difficulties ranged -0.43 to 1.86 logits -0.97 1.09 Three (Q5, Q15, & Q17) one (Q2) flagged as potentially less productive measurement, given floor effects (i.e., majority experiencing item's description), along disordered thresholds and/or small Two pairs considered redundant (Q6 Q7 Q29 Q30). remaining or performed well based on multiple indications. Overall, reliable, valid, sound. Few identified potential candidates removal assessments patients.