Responsiveness to change in health status of the EQ-5D in patients treated for depression and anxiety

Kenneth Sandin, Gemma Shields, Ragne G.H. Gjengedal, Kåre Osnes, Marianne T. Bjørndal, Silje E. Reme, Odin Hjemdal

The EQ-5D is a commonly used generic measure of health but evidence on its responsiveness to change in mental health is limited. This study aimed to explore the responsiveness of the five-level version of the instrument, the EQ-5D-5 L, in patients receiving treatment for depression and anxiety. Patient data (N = 416) were collected at baseline and at end of treatment in an observational study in a Norwegian outpatient clinic. Patients were adults of working age (18-69 years) and received protocol-based metacognitive or cognitive therapy for depression or anxiety according to diagnosis. Responsiveness in the EQ-5D was compared to change in the Beck Depression Inventory-II (BDI-II) and the Beck Anxiety Inventory (BAI). Effect sizes (Cohen's d), Standardised response mean (SRM), and Pearson's correlation were calculated. Patients were classified as "Recovered", "Improved", or "Unchanged" during treatment using the BDI-II and the BAI. ROC analyses determined whether the EQ-5D could correctly classify patient outcomes. Effect sizes were large for the BAI, the BDI-II, the EQ-5D value and the EQ VAS, ranging from d = 1.07 to d = 1.84. SRM were also large (0.93-1.67). Pearson's correlation showed strong agreement between change scores of the EQ-5D value and the BDI-II (rs -0.54) and moderate between the EQ-5D value and the BAI (rs -0.43). The EQ-5D consistently identified "Recovered" patients versus "Improved" or "Unchanged" in the ROC analyses with AUROC ranging from 0.72 to 0.84. The EQ-5D showed good agreement with self-reported symptom change in depression and anxiety, and correctly identified recovered patients. These findings indicate that the EQ-5D may be appropriately responsive to change in patients with depression and anxiety disorders, although replication in other clinical samples is needed.