After-hours discharges from intensive care are associated with increased mortality.

Objective: To investigate the change in pattern of discharge of patients from an intensive care unit (ICU) to hospital wards and to determine the impact of discharge time on subsequent hospital mortality.

Design And Participants: A retrospective cohort study of 10 903 patients discharged alive from a single ICU between 1 January 1992 and 31 December 2002.

Main Outcome Measure: In-hospital mortality.

Results: Of the 10 903 patients discharged alive from the ICU, 486 (4.5%) died in hospital wards. When discharge times were categorised according to nursing shift (morning, 07:00-14:59; afternoon, 15:00-21:59; and night, 22:00-06:59), patients were more likely to be discharged on an afternoon shift (odds ratio, 3.63; 95% CI, 3.05-4.30) or night shift (4.52; 95% CI, 3.15-6.64) in 2000-2002 compared with 1992-1994. In a multiple logistic model, hospital mortality after discharge from the ICU was increased by higher APACHE II score (1.14; 95% CI, 1.12-1.16); admission to ICU from the operating room (1.47; 95% CI, 1.11-1.95) and from the general ward (1.75; 95% CI, 1.37-2.23); and discharge during the afternoon (1.36; 95% CI, 1.08-1.70) and night shifts (1.63; 95% CI, 1.03-2.57).

Conclusion: Over an 11-year period, more patients are being discharged from the ICU in the afternoon and night suggesting increasing pressure on ICU beds. Patients discharged on these shifts have an increased risk of death.