Reliability of sleep log data versus actigraphy in veterans with sleep disturbance and PTSD

Abstract

The goal of the study was to assess inter-rater reliability of the daily sleep log (a self-rating) with actigraphy (an objective measure of sleep based on activity) in veterans with Posttraumatic Stress Disorder (PTSD). This analysis focused on time asleep and number of awakenings during bedtime. Study participants consisted of 21 veterans with a lifetime diagnosis of Posttraumatic Stress Disorder and current sleep disturbance symptoms. Data collection included study participants' daily charting of sleep logs and actigraphy (utilizing study participants' activity level). Data analysis included the following: (1) interrater reliability for the tabulation of self-reported sleep logs by two trained raters using 99 nights of sleep from 10 cases; (2) comparison of sleep log data versus actigraphic findings for sleep time during 241 bedtimes; (3) comparison of sleep log data versus actigraphic findings for awakenings during 241 bedtimes. Findings showed that the two raters had intraclass correlation scores of .801 for time spent asleep and .602 for time spent in bed-acceptable scores for tabulation of the sleep logs. Comparison of patients' sleep logs versus actigraphy for 241 nights showed that 10 out of 21 study participants had acceptable intraclass correlations of 0.4 or above for duration of sleep. However, sleep logs and actigraphic data on number of sleep awakenings showed poor intraclass correlation, with only 1 subject having an intraclass correlation greater than .30. In conclusion, these data strongly suggest that sleep logs do not reproduce actigraphic records in patients with PTSD even though the sleep logs were reliably quantified. Sleep logs especially under-count awakenings in PTSD patients with sleep complaints.