Effect of centre-, patient- and procedure-related factors on intensive care resource utilisation after cardiac surgery.

To determine associations between intensive care resource utilisation and centre-, patient- and procedure-related factors. Prospective multicentre cohort study. Twenty-one European intensive care units. Four thousand four hundred adult patients who had undergone cardiac or thoracic aortic surgery in 21 centres. None (observational study). Primary outcomes were duration of artificial ventilation and intensive care unit (ICU) length of stay. Exposures were centres and patient- and procedure-related factors. Both outcomes varied fourfold among centres. Median time to extubation varied from 5 to 19 h and ICU length of stay varied from 22 to 91 h. Cox regression analysis was performed to adjust risks of prolonged ventilation and ICU length of stay for patient-, procedure- and centre-related factors. Patient- and procedure-related factors were the main risk factors among individual patients, accounting for nearly two thirds of the risk of prolonged ventilation and ICU length of stay. Centre-related variation accounted for the remaining risk. In European ICUs resource utilisation is highly variable after cardiac surgery. Up to two thirds more patients could be treated with current ICU resources if the most efficient strategies and structures were applied across all centres