Several patients die in the intensive care unit (ICU) due to their acute disease and comorbid conditions. Moreover, after ICU discharge, some ICU survivors still die in the wards.

Previous studies have shown that the clinical condition of the patient at the time of ICU discharge may influence in-hospital prognosis. Non-modifiable factors, such as age and comorbid conditions certainly play a role. But inflammatory status (especially C-reactive protein - CRP), Sequential organ failure score (SOFA) score and Therapeutic intervention scoring system (TISS) 28 score have also been related with the risk of hospital death.

Admission to a high dependency unit may reduce the ICU length of stay (LOS). This strategy may also help to improve prognosis contributing to further stabilize the patient, facilitate his autonomy and the removal of invasive devices. However the benefit of a step-down strategy (from ICU to a high dependency unit) has never been evaluated.

The development of a score to evaluate the risk of patients discharged from the ICU may help to improve the allocation of resources, either to prolong the ICU stay, or admission to a high dependency units or the ward.

Methods Prospective, observational, international, multicentre study to be conducted in 2 countries (Portugal, Brazil). Inclusion criteria: Each centre may include a maximum of 110 consecutive patients discharged from the ICU with length of stay (LOS)>24h during a 6 month period.

Exclusion criteria: Limitation of care decision Primary objective: To determine the impact in hospital mortality and LOS (ICU and hospital) of a step-down strategy, from the ICU to a high dependency unit, before admission to the ward.

Secondary objectives: To develop and validate a score of the risk of death in the hospital after discharge from the ICU

Inclusion Criteria:

• Consecutive patients discharged from the ICU with LOS>24h

Exclusion Criteria:

• Presence of limitation of care decision

Patients discharge from the ICU with a ICU LOS>24h