Psychometric and structural properties of the Karolinska Exhaustion Disorder Scale: a 1,072-patient study

Abstract

Abstract Objective Exhaustion disorder is a stress-related diagnosis that was introduced in 2005 to the Swedish version of the International Statistical Classification of Diseases and Related Health Problems, 10th edition (ICD-10). The Karolinska Exhaustion Disorder Scale (KEDS) was developed to assess exhaustion disorder symptomatology. While the KEDS is intended to reflect a single construct and be used based on its total score, the instrument's characteristics have received limited attention. This study investigated the KEDS’s psychometric and structural properties in a large clinical sample. Methods The study relied on data from 1,072 patients diagnosed with exhaustion disorder that were included in two clinical trials in Sweden. We investigated the dimensionality, homogeneity, and reliability of the KEDS using advanced statistical techniques, including exploratory structural equation modeling (ESEM) bifactor analysis. Results A one-factor confirmatory analytic model exhibited a poor fit, suggesting at least a degree of multidimensionality. The ESEM bifactor analysis found the general factor to explain about 72% of the common variance extracted, with an omega hierarchical coefficient of 0.680. Thus, the ESEM bifactor analysis did not clearly support the scale’s essential unidimensionality. A homogeneity analysis revealed a scale-level H of only 0.296, suggesting that KEDS’s total scores do not accurately rank individuals on the latent continuum assumed to underlie the measure. The KEDS’s reliability was modest, signaling considerable measurement error. Conclusion Findings reveal important limitations to the KEDS with possible implications for the status of exhaustion disorder as a nosological category. Trial registration This study was pre-registered on Open Science Framework (osf.io) on April 24, 2022 (https://osf.io/p34sq/).