Standardized procedures for the MSLT were published in 1986 2 and officially endorsed by the American Sleep Disorders Association in 1992. 3 Procedures involve the measurement of sleep latency (time to sleep onset) during four to five nap opportunities dispersed at two hour intervals throughout the day beginning 1.5-3 hours after awakening. 2 Scores on the MSLT are thought to represent a measure of physiological sleep tendency. 4 The MSLT has become the most widely used measure of physiological sleep propensity, 5 and it has been demonstrated to be sensitive to the effects of partial and total sleep deprivation 6 and pharmacological manipulations of alertness. 7 Finally, the MSLT has been found to have high (r =.97, p<.001) test-retest reliability at three to six month intervals in healthy individuals. 8 To date, only one study has been published examining the scoring reliability of the MSLT in a clinical population. 9 The findings of that study suggested the MSLT has excellent interrater reliability (r=.85-.88). However, those results were based on a very small sample (n=21) and used only three raters for reliability estimation. In addition, no assessment of intrarater reliability, a limiting factor for interrater reliability, was undertaken. The present study was undertaken to address the interrater and intrarater scoring reliability of the MSLT in a large clinical population across a range of sleep disorders and levels of sleepiness.
METHODS
PatientsTwo hundred consecutive patients presenting to the sleep center that were administered the MSLT were used for the present study. Data were collected during the first six months of 1995. Patients included 145 men and 54 women, aged 14-75 years (mean=47; s.d.=11.5). Coefficients for the mean number of REM onsets during the MSLT were .81 for intrarater and .88 for interrater reliability. Intrarater and interrater agreement (kappa coefficients) for the presence of at least one REM onset during the MSLT was .78 and .86, respectively. For the presence of greater than one REM onset, a kappa of .78 was obtained for intrarater agreement and .91 for interrater agreement.
Conclusions:The clinical MSLT displays excellent interrater and intrarater reliability estimates for both sleep latency and REM onset scores in a sleep-disordered population.