The Epworth Sleepiness Scale (ESS) has been shown to have only moderate reproducibility in repeat assessments of patients in a sleep clinic population.
It should probably not be used in clinical settings to monitor the impact of therapeutic interventions in individuals or for prioritising access to services such as sleep studies, according to the findings of a Sydney study.
The study, comprising 108 adult patients, compared the ESS in patients at a first specialist visit and again on the night of diagnostic polysomnography. The test-retest interval was a median of 64 days.
Most patients (90%) had sleep-disordered breathing (Respiratory Disturbance Index >5 events/hr) and 26% had severe disease ( ≥30 events/hour).
The study found the range of scores across the two measurements was 0-23. The mean score was 10.1 at the clinic visit and 9.0 at polysomnography.
“Although the mean difference (bias) was 1.1, the 95% limits of agreement were between −8.5 and +10.6. An absolute difference in scores of at least 3 was observed in 60 (56%) participants,” the study said.
There was no correlation between ESS and RDI and no evidence that factors such as age, gender, BMI, medication use or sleep study parameters were significant predictors of test–retest variability in the ESS score.
“Together these results raise concerns about the utility of the ESS as a reliable measure of EDS in a sleep clinic population,” the study said.
Senior investigator Dr Hima Vedam, from the respiratory and sleep medicine department at Liverpool Hospital, told the limbic their findings were also consistent with other studies.