Predictability of sleep in insomnia: sleep patterns of patients from a sleep psychology clinic.

The present study aims at identifying sleep patterns in insomnia in a clinical sample using three strategies to define poor nights. Sleep diaries and self-reported questionnaires were collected from 77 clinical patients with insomnia. The conditional probabilities of observing a poor night after 1, 2, or 3 consecutive poor nights were computed according to three strategies with same criteria for sleep onset latency, wake after sleep onset, and sleep efficiency, but varying criterion for total sleep time. Latent profile analyses were conducted to derive sleep patterns. Uni- and multivariate analyses were conducted to characterise the sleep patterns identified. A total of 1586 nights were analysed. The strategy used significantly influenced the average percentage of reported poor nights. Two to three sleep patterns were derived per strategy. Within each strategy, sleep patterns differed from each other on sleep variables and night-to-night variability. Results suggest the existence of sleep patterns in insomnia among individuals consulting in psychological clinics. Adding a total sleep time of 6-h cut-off as a criterion to define poor nights increases the accuracy of the strategy to define poor night and allows to identify sleep patterns of poor nights in insomnia.