Subjective Ratings of Mental and Physical Health Correlate With EQ-5D-5L Index Values in Survivors of Critical Illness: A Construct Validity Study*

Abstract

Objectives: Survivors of critical illness commonly show impaired health-related quality of life (HrQoL). We investigated if HrQoL can be approximated by brief, easily applicable items to be used in primary care. Design: Secondary analysis of data from the multicenter, cluster-randomized controlled Enhanced Recovery after Intensive Care trial (ClinicalTrials.gov: NCT03671447) and construct validity study. Setting: Ten participating clusters of ICUs in the metropolitan area of Berlin, Germany. Patients: Eight hundred fifty ICU survivors enrolled in a mixed, medical or surgical ICU when they had an expected ICU length of stay of at least 24 hours, were at least 18 years old, and had statutory health insurance coverage. Interventions: None. Measurements and Main Results: Patients received follow-ups scheduled 3 and 6 months after ICU discharge. HrQoL was assessed with the EuroQol 5-Dimension 5-Level (EQ-5D-5L), and patients were asked to rate their current mental and physical health state from 0 (worst) to 10 (best). We fitted prediction models for the EQ-5D-5L index value using these two items and additional covariates, applying stepwise regression and adaptive lasso. Subjective mental health (Spearman: 0.59) and subjective physical health (Spearman: 0.68) correlated with EQ-5D-5L index values and were better predictors of EQ-5D-5L index values in the two-item regression (normalized root mean squared error [nRMSE] 0.164; normalized mean absolute error [nMAE] 0.118; R 2 adj 0.43) than the EQ-5D Visual Analog Scale (nRMSE 0.175; nMAE 0.124; R 2 adj 0.35). Stepwise regression with additional covariates further increased prediction performance (nRMSE 0.133; nMAE 0.1; R 2 adj 0.51). Conclusions: Asking patients to rate their subjective mental and physical health can be an easily applicable tool for a first impression of the HrQoL in primary care settings.