• Patricia L Turner, Aaron G Ilano, Yue Zhu, Steven B Johnson, and Nader Hanna.
• University of Maryland Medical Center, University of Maryland School of Medicine, Baltimore, MD, USA.
• J. Am. Coll. Surg. 2011 Mar 1;212(3):287-94.

BackgroundSimilarity between the ICU patient data acquired by the APACHE system and the patient outcomes data acquired by the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) suggests that variables in NSQIP data could be effective mortality predictors. We theorized that identifying ACS-NSQIP preoperative data points predictive of patient outcomes would assist in identifying potential complications earlier.Study DesignBetween 2006 and 2008 at the University of Maryland Medical Center, we identified 340 surgical ICU patients included in both databases as our study cohort. Median APACHE score calculation permitted division of study subjects into quartiles from which to examine length of stay, mortality variables, and ACS-NSQIP-recorded postoperative occurrences and preoperative risk factors. Outcomes were compared using each database's percentage of patients who had died.ResultsMortality was positively correlated with APACHE score increases. Initially, the average ICU length of stay increased, then declined. Most common postoperative occurrences were ventilator dependence >48 hours, pneumonia, unplanned intubation, sepsis, and septic shock. It was also noted that preoperative variables, such as American Society of Anesthesiologists classification and serum albumin levels, had an association with a poorer prognosis.ConclusionsAPACHE score predictions are consistent with ACS-NSQIP-recorded postoperative outcomes. Higher APACHE scores correlated with increased incidence of postoperative outcomes and were associated with earlier mortality in the most extremely ill. Poorer states of health before surgery also correlated with poor outcomes postoperatively.Copyright © 2011 American College of Surgeons. Published by Elsevier Inc. All rights reserved.

14 keywords...

hide…