Low test-retest reliability of the Epworth Sleepiness Scale within a substantial short time frame.

ABSTRACT

The Epworth Sleepiness Scale (ESS) is a widely used, validated questionnaire for effectively examining patients' sleepiness in a range of different situations. Test-retest reliability is an important aspect of a questionnaire, which, according to only a few studies, was found to be low in the case of the ESS. All these studies applied long intervals between the tests, thereby increasing the possibility of fundamental change in circumstances, which in turn affect the reliability of the test. The aim of the present study was to investigate the test-retest reliability of the ESS in a short time frame to provide stability of the test circumstances. We also compared the originally used and current accepted statistical methods of test-retest evaluation. We examined 100 unselected patients consecutively referred to the sleep laboratory with the ESS questionnaire, using a test-retest paradigm with an interval of 1 h between two ESS tests. The Lin's concordance coefficient was found to be low, whereas the Pearson's correlation revealed good reliability. Our result provides evidence on the poor test-retest reliability of the ESS, despite the examination protocol excluding changes in test circumstances.