Postoperative patients in the intensive care unit: Identifying those who do not really need it.

Purpose: We evaluated severity-scoring systems as predictors of intensive care unit (ICU) need and created a new model for identifying postoperative patients who do not really need ICU.

Materials And Methods: The American Society of Anesthesiologists (ASA), the Physiological and Operative Severity Score for the enUmeration of Mortality and Morbidity (POSSUM), and Simplified Acute Physiology Score III scores of 100 postoperative patients were calculated, and cutoff values for necessary ICU admission were obtained. Criteria for a "necessary admission" were death, length of stay more than 48 hours, need for vasoactive agents, or mechanical ventilation for more than 24 hours. Patients whose scores were greater than the cutoff values for the 2 most discriminative variables were defined as high-risk patients; and the rest, as low-risk patients (LRPs). Relative risk, positive predictive value, and negative predictive value were calculated.

Results: The POSSUM-total (P-total) and ASA were the 2 most discriminative scores. High-risk patients (patients with ASA scores≥3 and P-total≥35) needed ICU 4.83-fold more than LRPs. The new model had the highest relative risk and negative predictive value (0.85) among all variables and the second highest positive predictive value (0.73) after P-total.

Conclusions: The new model can predict LRPs more accurately than each scoring system alone. The care of LRPs in intermediate care units can prevent overuse of ICUs. But the lack of outcome comparison for predicted LRPs in ICUs vs intermediate care units is the most important limitation of our study.