Test-Retest Reliability of an Interactive Voice Response Version of the EQ-5D in a Sample of Cancer Survivors

Abstract

Electronic data capture technologies, such as interactive voice response (IVR) systems, are emerging as important alternatives for collecting patient-reported outcome data. The objective of this study was to assess the test-retest reliability of an IVR version of the EQ-5D. Outpatient cancer clinic patients (n = 127) were asked to complete the IVR-based EQ-5D twice, 2 days apart. The analyses tested for mean differences (paired t-test) and test-retest reliability (intraclass correlation coefficient [ICC]) to assess measurement stability over time. Equivalence of the means was established if the 95% confidence interval (CI) was within the minimally important difference (MID) interval; namely -0.035 to 0.035 for the EQ-5D index and -3.0 to 3.0 for the visual analog scale (i.e. EQ VAS). Adequacy of the ICC was established by testing whether it differed from a value of 0.70. Both administrations were completed per protocol by 114 subjects (EQ-5D index) and 110 subjects (EQ VAS). For the EQ-5D index, the means (SD) of the first and second administrations were 0.871 (0.14) and 0.871 (0.15), respectively. The 95% CI of the mean difference was -0.013, 0.013, within the equivalence interval. The ICC was 0.876 (95% lower bound of 0.826) and was significantly different from 0.70. The EQ VAS means (SD) were 81.3 (17.5) and 80.8 (17.5), respectively. The 95% CI of the mean difference was -0.598, 1.617, within the equivalence interval. The EQ VAS ICC was 0.944 (95% lower bound of 0.919) and was significantly greater than 0.70. This analysis provides substantial evidence that the scores obtained from the IVR version of the EQ-5D are reliable upon repeated administrations.