A comparison of visual analogue and numerical rating scale formats for the Lung Cancer Symptom Scale (LCSS): Does format affect patient ratings of symptoms and quality of life?

Problem and Purpose: The Lung Cancer Symptom Scale (LCSS), a site-specific health-related quality of life measure for patients with lung cancer, was originally developed using a Visual Analogue Scale (VAS) format. However, the VAS format is not readily compatible with data management and software programs using scanning. The primary aim of this study was to evaluate the convergence of ratings obtained with a Numerical Rating Scale (NRS), with an 11-pt response category format, to those obtained with a VAS format. The intent was to determine the degree of agreement between two formats to generalize the existing psychometric properties for the original measure to the new presentation. Design/setting: This methodological study evaluated the feasibility, reliability, and validity of a NRS format for the LCSS. The study was conducted at two cancer centers in New York City. Patients/procedures: Sixty-eight patients with non-small cell lung cancer (NSCLC) completed both versions of the LCSS along with demographic and feasibility questions on a single occasion. The VAS form was administered first, followed by the NRS form to prevent bias. The intraclass correlation coefficient (ICC), Lin's concordance correlation coefficient (CCC), and Bland-Altman plots were used to evaluate agreement and to characterize bias. Results: Cronbach's alpha for the NRS format total score was 0.89 for the 68 patients with NSCLC. Agreement was excellent, with both the ICC and CCC 0.90 for the two summary scores (total score and average symptom burden index) for the LCSS. Only five of the nine individual items showed this level of strict agreement. An agreement criterion of 0.80 (representing excellent) was observed for seven of the nine individual items (all but appetite loss and hemoptysis). Mean differences tended to be slightly lower for the VAS format compared to the NRS format (more so for the appetite and hemoptysis items), with evidence of scale shift for the same two items. The summary measures showed good concordance as measured by the ICC and CCC, but did display mean differences (VAS - NRS) of -2.7 and -3.1, respectively. Conclusions: Overall, the NRS format for the LCSS suitable for scanning has good feasibility, reliability (internal consistency), and convergent validity. The complete set of concordance evaluation measures supports the reproducibility of VAS scores by NRS scores, particularly for the two summary scores.