ASSESSING SLEEP IN ELDERLY INSOMNIACS INTRODUCTION SINCE THE INTRODUCTION OF POLYSOMNOGRAPHY (PSG) IN THE 1950S, THIS METHOD HAS BEEN REGARD- ED AS THE GOLD STANDARD FOR OBJECTIVE assessment

SINCE THE INTRODUCTION OF POLYSOMNOGRAPHY (PSG) IN THE 1950S, THIS METHOD HAS BEEN REGARDED AS THE GOLD STANDARD FOR OBJECTIVE assessment of sleep,1 providing measures of both sleep and wake time, in addition to classification of sleep stages. Even though PSG offers extensive information on sleep behavior and sleep physiology, it is very expensive and can sometimes be too invasive to be used in clinical studies in which the primary focus is quantification of sleep time, wake time, or both. Actigraphy has been suggested as an alternative assessment method to PSG. Actigraphy consists of an accelerometer and memory storage, both fitted into a watchlike device to be worn around the wrist. Based on differences in movements associated with wakefulness and sleep, actigraphy provides an estimate of sleep-wake schedules. Being independent of patients’ ratings and personal judgments of their sleep, actigraphy may offer a less time-consuming and less expensive alternative to PSG. In addition, actigraphy is regarded as a less invasive measure. Studies have shown actigraphy to correspond well with PSG in adults with impaired sleep (concordances ranging from 81%2 to 87%3). However, the level of agreement is often reported as crude percentage or correlation coefficients. This may overestimate the level of agreement because an estimate of agreement based on crude percentage fails to take into account the chance agreement due to the high base rate of sleep during the night.4 Furthermore, agreement has been reported as an average across participants, and this may mask individual differences in correspondence between actigraphy and PSG. This possibility of interindividual differences should be explored because it may have implications for the use of actigraphy as a method in detecting individual changes in sleep behavior. A more informative approach would be Tryon’s method of separately calculating and reporting sensitivity, specificity, and overall accuracy.5 In this view, sensitivity for sleep is the proportion of PSG-registered sleep epochs also identified as sleep by actigraphy, whereas specificity for sleep is the proportion of wake epochs correctly identified by actigraphy. By using this scheme, actigraphy has been shown to be more successful in detecting sleep (sensitivity), compared with detecting wakefulness (specificity).6, 7 Only a few studies have examined to what extent actigraphy is able to detect treatment changes, as compared with PSG.8-11 Verbeek et al11 compared actigraphy, PSG, and sleep diaries in a sample of 20 middle-aged insomniacs, concluding that actigraphy overestimated the amount of sleep compared to PSG. Also, Valliere and Morin10 demonstrated that actigraphy was sensitive enough to detect treatment changed following cognitive-behavior therapy in a similar-aged cohort. Although not all sleep changes A Comparison of Actigraphy and Polysomnography in Older Adults Treated for Chronic Primary Insomnia