The Science Journal of the American Association for Respiratory Care

2009 OPEN FORUM Abstracts

EFFICACY AND TEST-RETEST RELIABILITY IN THE EPWORTH SLEEPINESS SCALE, A TOOL FOR ASSESSING DAYTIME SLEEPINESS

Martine J. Eon, John Dziodzio; Pulmonary Medicine/ Critical Care Medicine, Maine Medical Center/ Maine Sleep Institute, Portland, ME

Background An important part of assessing patients for sleep disorders is an estimation of Excessive Daytime Sleepiness (EDS). Techniques used to assess EDS are cumbersome and expensive, and other subjective screening tools correlate poorly with objective measurements. The Epworth Sleepiness scale (ESS) was developed to overcome issues in prior techniques, and was validated with objective measures like sleep latency. In our practice, we have noted patients with abnormal ESS scores and normal sleep studies, as well as patients with normal ESS with sleep studies categorized as severe by elevated Apnea Hypopnea Index (AHI). The hypothesis is that the ESS is a reliable tool to predict EDS and in turn Sleep Disordered Breathing (SDB). Methods To estimate efficacy of ESS in identifying SDB, we studied relationships between ESS and AHI, mean sleep latency (MSL), and age. To validate ESS as a repeatable measure, we administered it to a random sample of 101 adult patients twice, with a few weeks between, during which no treatment was given. Correlations between the ESS and AHI, and between ESS and sleep latency were determined by linear regression and test-retest precision assessed by paired t-test. Fourteen subjects with potentially confounded data were excluded. These patients had a change in prescribed medications or the addition of Continuous Positive Airway Pressure (Cpap) therapy prior to the second ESS. Results Test-retest precision was examined via the differences in paired data. The mean difference (“bias”) was 0.609, showing very little bias between pre and post data. The standard deviation (SD) of differences (“precision”) was 3.46, 95% confidence interval (CI) 6.92, overall range 13.8, over half of the maximum ESS score of 24. The correlation coefficient (R2) from regression of pre and post scores was 0.552, implying the regression explains only 55% of the variation, with 45% unexplained. Regression of mean ESS on AHI provided R2= 0.0041, with only 0.41% of the variance due to the regression and 99.6% unexplained. Regression of mean ESS on MSL provided R2= 0.012, explaining only 1.2% of the variance with 98.8% unexplained. Conclusions ESS failed to show any clinically useful relationship with either MSL or AHI, two basic measures of sleep performance; and the test-retest precision failed to demonstrate clinical significance; therefore we reject the hypothesis of clinical usefulness. Sponsored Research - None

You are here: RCJournal.com » Past OPEN FORUM Abstracts » 2009 Abstracts » EFFICACY AND TEST-RETEST RELIABILITY IN THE EPWORTH SLEEPINESS SCALE, A TOOL FOR ASSESSING DAYTIME SLEEPINESS