In this proposal, investigators seek to address conclusively two knowledge gaps: 1) the lack of data on the impact of a schedule including a day off clinic before night shift on performances of residents in anesthesia or critical care medicine and 2) the lack of data on the relationship between resident sleep deprivation and their non technical skills impairment. Investigators designed a prospective, randomized, blinded in cross over evaluation of medical pratices. Investigators will compare resident's performance during a high fidelity simulation session, on crisis managment in intensive care unit, after a night shift in intensive care unit, of residents who work on an intervention schedule with no clinical activity before night shift, with resident's performance of residents who had a traditionnal schedule including an usual clinical day before night shift. Investigators also assess cognitive performances, sleepiness and self esteem before and after night shift. Investigators specific aims will be:

To test the hypothesis that residents working on an intervention schedule will have non technical skills, assessed by Ottawa GRS global score, significantly less impaired after a night shift, than those on traditionnal schedule

To test the hypothesis that residents working on an interveniton schedule will have cognitive performances, sleepiness and self esteem less impaired after night shift than those on a traditionnal schedule

Main objective of this study is to compare the non-technical skills after night shift, of anaesthesia and critical care residents who did not have clinical activity before the night shift with those who had at least 8 hours of clinical activity before. The workforce is constrained by the number of residents in training at Clermont-Ferrand and Lyon University Public Hospital. Thus, it is possible to consider the inclusion of 60 residents for 55 evaluable. With such numbers, the clinically relevant difference of 6 points for the Ottawa GRS score (for a standard deviation of 11) can be demonstrated with an alpha risk of bilateral error of 5% and a power of 80%. Two interim analyzes will be carried out including the first 20 interns and then 20 additional interns (therefore for 40 interns) with the aim of stopping the study for "effectiveness", more precisely stopping the study for evidence of a statistically significant difference for the primary outcome measure, with risks of first error observed respectively at 0.0005 and 0.014 comparisons (correction due to multiples). A simulation of the statistical power will also be proposed during the performance of these analyzes; if necessary, a new estimation of the workforce will be proposed to the competent authorities. Investigators do not plan loss of follow up.

Inclusion Criteria:

• residents in Anesthesia or Critical care medicine
• Who effect night shift in intensive care unit
• Able to take a day off clinic before night shift
• Able to give consent

Exclusion Criteria:

• Refusal of participation

The investigators will include volunteer residents in anesthesia of critical care medicine of Gabriel Montpied university hospital of Clermont Ferrand.