Abstract
The purpose of this study was to clarify the efficiency of using a sleep log in the sleep/awake assessment of frail elderly people. In study 1, 1-h-duration sleep logs recorded by nurses or caregivers and actigrams from wrist-activity devices were simultaneously monitored in four frail elderly participants for 21 days (total, 504 h). In study 2, 1-min-duration measurements of both kinds of data in another subject were taken for 3 days (4,320 min). In study 1, the actigraph data agreed with the sleep log data between 57% and 78% of the time for "sleep", whereas for "awake" they agreed between 91% and 97% of the time. In study 2, the concordance rates were 86% for "sleep" and 97% for "awake". These results suggested that the sleep log was as accurate as the actigraph data in sleep/awake assessment of the frail elderly with an adequate judging standard and prior observation /record training. Moreover, the wakeful state was sometimes misjudged as "sleep" by the actigram in slowmoving subjects. Thus the sleep log might assess wakefulness status more accurately than the actigram when the movements of the participant are slow on awakening.
