Clinical utility of the Epworth sleepiness scale

Abstract

Purpose: The Epworth sleepiness scale (ESS) is a widely used tool which has been validated as a measure of sleepiness. However, the scores within individual patients referred for clinical sleep services vary considerably which may limit the clinical use of the ESS. We sought to determine the test-retest reliability of the ESS if scores were classified as either normal or sleepy.

Methods: We measured the ESS in patients presenting to our sleep center at a clinical visit and again when a sleep study was done. Demographic and clinical information was extracted from the electronic medical record.

Results: Average ESS scores were similar on 2 administrations, mean (SD) of 9.8 (5.4) and 10.2 (6.2). Bland-Altman analysis showed upper and lower limits of agreement of 7.5 and - 6.7, respectively. No demographic or clinical variables were identified which contributed to the intra-individual variability. Of the patients who presented with an initial ESS < 11, 80% had a second ESS < 11. Of the patients who presented with an initial ESS ≥ 11, 89% had a second ESS ≥ 11. Cohen's kappa for the two administrations of the ESS was 0.67 (95% CI of 0.51-0.83). Using previously published reports, we calculated Cohen's kappa for polysomnographic determination of the apnea-hypopnea index (AHI) with values ranging from 0.26 to 0.69.

Conclusions: Individual ESS scores varied considerably within individual patients, but with classification into either normal or sleepy, the test-retest reliability was substantial and in line with other clinical measures including polysomnographic determination of the AHI.

Keywords: Epworth sleepiness scale; Kappa; Polysomnogram; Test-retest reliability; Variability.