Symptom assessment in small cell lung cancer (SCLC) in a randomized trial: A psychometric analysis of Patient Symptom Assessment in Lung Cancer (PSALC)

Abstract

18101 Background: Lung symptoms may be very burdensome for patients (pts) with SCLC. PSALC is a symptom scale that was developed for use in pts with SCLC for assessment of 9 symptoms (shortness of breath, cough, chest pain, hemoptysis, appetite loss, sleep interference, hoarseness, fatigue, interference with daily activities) scored from 1 (not at all) to 4 (very much), but has not been formally validated. The objective of this study is to evaluate the validity of the PSALC through a psychometric analysis of trial data. Methods: Data were analyzed from a randomized, open-label, multicenter trial with 141 pts with SCLC receiving either oral topotecan with best supportive care (BSC) or BSC alone. PSALC and EQ-5D were administered to pts at baseline and at 3-week intervals at subsequent visits. Endpoints for psychometric analysis included statistics for internal consistency, construct validity, reliability, and responsiveness to change in clinical status. Results: Factor analysis (n=131) indicated that there was only one factor, therefore PSALC total score (PSALC-TS) could be used for psychometric analysis. Weighted Cronbach’s alpha from all visits was 0.78. Construct validity was supported by association with established measures: higher baseline PSALC-TS (worse symptoms) was associated with worse ECOG (12.31, 16.37, 17.49 for ECOG score=0, 1, 2, respectively, p=0.0002); baseline PSALC-TS was significantly correlated with EQ-5D utility index (Pearson correlation coefficient [CC] =-0.493, p<0.0001, n=125) and EQ-5D VAS score (Pearson CC=-0.495, p<0.0001, n=122). Reliability was supported by intraclass CC of 0.68 for PSALC-TS evaluated before any change in clinical status, and concordance CC of 0.65, calculated using PSALC-TS at baseline and before first visit. PSALC-TS was responsive to change in clinical status from baseline to tumor response (responsiveness statistic (RS) =-0.99) and to tumor progression (RS=0.94). Conclusions: Retrospective analysis of trial data suggests that PSALC is a reliable, valid, and responsive scale for measuring SCLC symptoms. If feasible in SCLC population, a prospective validation study could be used to further evaluate the validity of this symptom scale. No significant financial relationships to disclose.