Assessing the Relationship Between EQ-5D-3L Index Scores and Disease-Specific Patient-Reported Outcome Measures in Eosinophilic Oesophagitis

OBJECTIVES: To evaluate in patients with Eosinophilic Oesophagitis (EoE), the correlation between 3-level EQ-5D (EQ-5D-3L) utility values and disease-specific patient-reported outcome measures (PROMs) - the Dysphagia Symptom Questionnaire (DSQ) and the EoE Impact Questionnaire (EoE-IQ) - and determine the suitability of EQ-5D-3L for assessing EoE-specific quality of life (QoL).

METHODS: In this post-hoc analysis of the R668-EE-1774 (NCT03633617) trial - part B, baseline EQ-5D-3L utilities (UK tariffs) were collected following a protocol amendment for 71 patients (aged ≥12 years). DSQ (range: 0–84; lower scores indicate less severe dysphagia) and EoE-IQ scores (range: 1 [no impact] to 5 [extremely impacted]) at baseline were summarised according to the EQ-5D-3L utility level (<0.4, 0.4–0.9, and >0.9). Correlation coefficients were calculated to determine the level of association between utility levels and EoE-IQ and DSQ scores. Three-dimensional scatterplots were developed to visually explore the relationships between the PROMs.

RESULTS: The mean EQ-5D-3L utility at baseline was 0.809 (standard deviation: 0.247), with 46.5% of patients having a utility of 1.00. When stratified by EQ-5D-3L utility scores (<0.4, N=8; 0.4–0.9, N=30; >0.9, N=33), differences were observed in baseline characteristics: mean age, 18.9/25.1/32.2 years; mean EoE duration, 3.8/5.6/6.6 years; and mean peak eosinophil count, 76.1/104.1/101.4 per high-power field, respectively. Patients in the lower utility subgroups (<0.4/0.4–0.9/>0.9) reported numerically more severe DSQ (means: 41.2/40.8/33.6, respectively) and EoE-IQ scores (means: 2.7/2.4/2.1, respectively). There were weak correlations between EQ-5D-3L utilities and EoE-IQ (r=−0.33) and DSQ scores (r=−0.30), while the DSQ and EoE-IQ scores were moderately correlated (r=0.43). Scatterplots revealed no clear visual relationship between EQ-5D-3L utilities and EoE-IQ and DSQ scores.

CONCLUSIONS: EQ-5D-3L utility measures showed a weak correlation with disease-specific PROMs in patients with EoE, suggesting that it may not be suitable for assessing health-related QoL. Clinicians should consider using EoE-specific QoL measures to accurately demonstrate the impact of EoE on patients’ QoL.