Meaningful change threshold estimation for the non-small cell lung cancer symptom assessment questionnaire (NSCLC-SAQ): psychometric analysis from a phase 3 trial (LIBRETTO-431).

Abstract

Purpose: Meaningful change thresholds are important to help interpret patient-reported outcome scores. To date, meaningful within-patient change (MWPC) thresholds have only been proposed for NSCLC-SAQ total score. This study proposed clinically MWPC thresholds, and group-level minimal important change/difference (MIC/MID) thresholds for both improvement and worsening for the Non-Small Cell Lung Cancer- Symptom Assessment Questionnaire (NSCLC-SAQ) total and symptom scores.

Methods: Blinded patient data (N = 246) from the Phase 3 LIBRETTO-431 clinical trial were used in a pre-specified meaningful change analysis. A combination of anchor- and supportive distribution-based methods were used to estimate the MWPC, MIC, and MID thresholds. Triangulation across anchor estimates was then performed using a correlation-weighted average to provide a single MWPC, MIC, and MID estimate for improvement and worsening.

Results: NSCLC-SAQ total score and symptom scores showed moderate to high correlations with various anchors (ranging from 0.306 to 0.890), with threshold estimates being provided from multiple anchors (except for cough). Triangulation suggested MWPC, MIC, and MID thresholds for improved total score were - 2.5, -3.5, and - 2.0, respectively. For worsening, the proposed thresholds were 2.0, 0.5, and 2.0, respectively. The MWPC, MIC, and MID thresholds for improved symptom scores ranged from - 0.5 to -1.5, and the worsening thresholds for symptom scores ranged from 0.5 to 1.0.

Conclusion: This study provides the first worsening and improvement estimates of MWPC, MIC, and MID for NSCLC-SAQ total and symptom scores. The thresholds proposed in this study can be used to inform interpretation of NSCLC-SAQ scores in clinical trials.