A Patient-Centered Utility Index for Non–Small Cell Lung Cancer in the United States

Abstract

Background. A preference-based quality-of-life index for non–small cell lung cancer was developed with a subset of Functional Assessment of Cancer Therapy (FACT)–General (G) and FACT–Lung (L) items, based on clinician input and the literature. Design. A total of 236 non–small cell lung carcinoma patients contributed their preferences, randomly allocated among three survey groups to decrease burden. The FACT-L Utility Index (FACT-LUI) was constructed with two methods: 1) multiattribute utility theory (MAUT), where a visual analog scale (VAS)–based index was transformed to standard gamble (SG); and 2) an unweighted index, where items were summed, normalized to a 0 to 1.0 scale, and the result transformed to a scale length equivalent to the VAS or SG MAUT-based model on a Dead to Full Health scale. Agreement between patients’ direct utility and the indexes for current health was assessed. Results. The agreement of the unweighted index with direct SG was superior to the MAUT-based index (intraclass correlation for absolute agreement: 0.60 v. 0.35; mean difference: 0.03 v. 0.19; and mean absolute difference 0.09 v. 0.21, respectively). Mountain plots showed substantial differences, with the unweighted index demon-strating a median bias of 0.02 versus the MAUT model at 0.2. There was a significant difference (P = 0.0002) between early (I-II) and late stage (III-IV) patients, the mean difference for both indexes being greater than distribution-based estimates of minimal important difference. Limitations. The population was limited to non–small cell lung cancer patients. However, most quality-of-life literature consulted and the FACT instruments do not differ-entiate between lung cancer cell types. Minorities were also limited in this sample. Conclusions. The FACT-LUI shows early evidence of validity for informing economic analysis of lung cancer treatments.