We studied the sensitivity, face validity, and con- struct validity of the rule of using two or more SIRS criteria for the diagnosis of severe sepsis in the first 24 hours after ICU admission. We found that the SIRS-criteria rule missed one patient in eight with severe sepsis