Objectives: To develop an influenza pandemic ICU triage (iPIT) protocol that excludes patients with the highest and lowest predicted mortality rates, and to determine the increase in ICU bed availability that would result. Design and setting: Post-hoc analysis of a study evaluating two triage protocols, designed to determine which patients should be excluded from access to ICU resources during an influenza pandemic. ICU mortality rates were determined for the individual triage criteria in the protocols and included criteria based on the Sequential Organ Failure Assessment (SOFA) score. Criteria resulting in mortality rates outside the 25th and 75th percentiles were used as exclusion criteria in a new iPIT-1 protocol. The SOFA threshold component was modified further and reported as iPIT-2 and iPIT-3. Main outcome measure: Increase in ICU bed availability. Results: The 25th and 75th percentiles for ICU mortality were 8.3% and 35.2%, respectively. Applying the iPIT-1 protocol resulted in an increase in ICU bed availability at admission of 71.7% +0.6%. Decreasing the lower SOFA score exclusion criteria to =s 6 (iPIT-2) and =s 4 (iPIT-3) resulted in an increase in ICU bed availability at admission of 66.9% + 0.6% and 59.4 + 0.7%, respectively (P