Assessing sleep in psychiatric inpatients: nurse and patient reports versus wrist actigraphy

This study was designed to evaluate the conventional techniques of assessing sleep, nursing and patient report, of inpatients on a clinical psychiatric unit. Nurses assessed sleep/wake status at hourly checks and patients completed a sleep diary. For three nights patients wore a wrist actigraph, a portable instrument which provides objective data about sleep/wake activity. The nursing and patient data obtained were compared with actigraphy data. Nursing staff evaluated sleep with satisfactory agreement (76.5% night 1 and 81.6% night 3) that improved over the first three nights of hospitalization (p < 0.03). When the nurses' report did not agree with the actigraph, they tended to overestimate sleep. Patients tended to underestimate their total sleep time and total time awake after sleep onset. Time in bed and initial sleep latency were overestimated. There was great intersubject variability, making determination of agreement impossible. This data suggest that treatment teams on psychiatric units should in general consider nursing reports of sleep more accurate than patient self-report. However, since nursing staff and patients observe different aspects of sleep, both sources of data are important to inpatient treatment teams on clinical units.