Objective: To compare registrar sleep and supervision hours before and after a change in roster to accommodate more senior registrar (SR) positions, and to identify risky patterns of sleep on night shifts.
Design, setting and participants: Prospective study of 21 registrars on two different roster templates from September 2010 to May 2011 in the intensive care unit of Wellington Regional Hospital, Wellington, New Zealand.
Intervention: Roster change from 13 registrars and one SR to 10 registrars and four SRs.
Main outcome measures: Mean sleep and supervision hours by shift; episodes of sleep ≤ 5 hours, wakefulness ≥17 hours, sleep during shift, waking before 16:00 before night shifts.
Results: 990 sleep surveys were analysed. There was no significant difference between groups in mean sleep or supervision hours for any shift. Two hundred and thirty-six night shifts were analysed. Registrars slept ≤5 hours before 19/236 (8.1%) night shifts; had ≥17 hours wakefulness before 79/236 night shifts (33.5%); woke by 16:00 107/ 236 (45.3%) times; and slept during 86/236 (36.4%) night shifts. Registrars arrived at work having either woken before 16:00 or had ≤5 hours of sleep on 114/236 (48.3%) night shifts.
Conclusions: Changing the registrar roster to meet the training demands of our senior trainees did not adversely affect registrar sleep or supervision. Registrars may be taking on unnecessary risk due to poor sleep hygiene around night shifts. We suggest sleep education and scheduled sleep time during night shifts.