Comparative performance of the EQ-5D-5L and SF-6D index scores in adults with type 2 diabetes

To explore the comparative performance including discriminative and longitudinal validity of EQ-5D-5L and SF-6D index scores in adults with type 2 diabetes.

Data from an on-going cohort study of adults with type 2 diabetes in Alberta, Canada, were used. Known-groups approach was used to examine discriminative validity. Correlation and agreement indices and scatter and Bland-Altman plots were used to examine the relationship between the two measures. Longitudinal validity was explored using Wilcoxon signed-rank test, effect size, and standardized response mean.

In 1832 participants at baseline (age 64.3, standard deviation 10.6 years; 45% female), median EQ-5D-5L score was 0.85 [interquartile range (IQR) 0.17], and floor and ceiling effects of 0.1 and 16.1%, respectively; median SF-6D score was 0.72 (IQR 0.24), and floor and ceiling effects of 0.1 and 3.2%, respectively. EQ-5D-5L and SF-6D index scores were significantly correlated with an overall Spearman correlation coefficient of 0.73, and an ICC of 0.62 (95% CI 0.42-0.74). Both EQ-5D-5L and SF-6D scores demonstrated statistically significant differences in self-reported chronic conditions, depressive symptoms, and diabetes-related distress, and were able to detect changes in depressive symptoms and diabetes distress across all change groups.

Both EQ-5D-5L and SF-6D index scores provide valid measurement in this patient population. Considerable overlap between the measures means it is not necessary to include both in surveys, however, the advantages and disadvantages of each should be considered.