Longitudinal profiles of sleep during residential PTSD treatment.

We investigated longitudinal profiles of objectively measured sleep periods (SP) over the course of residential treatment for posttraumatic stress disorder (PTSD) in a cohort of U.S. male veterans. Participants (N = 190) slept on mattress-actigraphy systems in a Veterans Affairs (VA) residential PTSD program. The final sample included 4078 sleep periods. Latent class mixed model analyses were used to identify between-subject profiles in sleep period durations, controlling for daily medication intake, over the first fifty days of residency. Logistic regression was used to determine the association of pre-treatment characteristics with identified profiles. Three longitudinal profiles of sleep period characterized most of the sample: 'stable' (56%), 'decreasing' (35%), and 'increasing' (8%). Less severe pre-treatment PTSD avoidance symptoms predicted membership in the 'decreasing' class and increased depression symptoms were predictive of membership in the 'increasing' class. Nearly half of the sample reported a summary change of 1 h or more over the course of the initial 50-nights in the program. Future work is needed to identify how these profiles might drive inpatient treatment decisions regarding the provision of adjunctive sleep-focused treatment such as cognitive-behavioral treatment for insomnia or hypnotic medications.