Critical Care: Critical Care - ICU Management |

Bag Score: An ICU Survival Scoring System in a Veteran Population FREE TO VIEW

Simon Yau, MD; Siva Bhavani, MD; Charlie Lan, DO; Kiran Nair, DO; Gloria Li, MD; Sarah Perusich, BS
Author and Funding Information

Baylor College of Medicine, Houston, TX

Copyright 2016, American College of Chest Physicians. All Rights Reserved.

Chest. 2016;150(4_S):214A. doi:10.1016/j.chest.2016.08.227
Text Size: A A A
Published online

SESSION TITLE: Critical Care - ICU Management

SESSION TYPE: Original Investigation Poster

PRESENTED ON: Wednesday, October 26, 2016 at 01:30 PM - 02:30 PM

PURPOSE: APACHE (Acute Physiology and Chronic Health Evaluation) II scoring system is the most widely used severity-of-illness scoring system. It combines age, physiological measurements, co-morbidities, and indication for admission to estimate the likelihood of mortality in patients admitted to the medical intensive care unit (ICU). However, similar to other scoring systems, these calculations are time-intensive and difficult to perform efficiently in the ICU. Furthermore, the APACHE II score has not been validated in the veteran population. The aim of this study is to assess the validity of the APACHE II scoring system in predicting mortality in a VA medical ICU population and to develop a simplified survival scoring system. Our hypothesis is that the APACHE II score does not perform as well as it does in the general population.

METHODS: This is a retrospective chart review of 625 patients admitted to the Michael E. DeBakey VA medical ICU over one year. We excluded any ICU readmissions, cardiac ICU admissions, or transfers from other hospitals. We collected data necessary to calculate the APACHE II score and other variables that affect ICU patient outcomes (i.e. need for vasopressor, lactic acid, etc). Accuracy of the model was assessed by discrimination and calibration using area under the receiver-operating characteristic curve (AUC) and standardized mortality ratio (SMR), respectively. We performed logistic regression analysis to assess which variables were most predictive of hospital mortality. Finally, we developed a simplified scoring system based on the logistic regression model.

RESULTS: The mean age of our population was 67 years, and 96% of patients were male. Average APACHE II score was 25.9. Overall hospital mortality and 30-day mortality were 32% and 30%, respectively. When controlling for all other variables, logistic regression analysis revealed that higher Glasgow Coma Scale (GCS), higher serum albumin, and lower serum bilirubin during the first 24 hours of ICU admission were variables most predictive of survival. Using the logistic regression model, we developed a simplified BAG score ([4 x Serum Albumin] + GCS - Total Serum Bilirubin). Higher scores were associated with higher probability of survival. Patients with a BAG score > 15 had a 77% survival rate, while patients with a score ≤ 15 had a 21% survival rate. Furthermore, BAG score had good discriminative value with an AUC of 0.82. Although discrimination was acceptable with the APACHE II score (AUC = 0.81), calibration was poor (SMR 0.63).

CONCLUSIONS: APACHE II scoring system has acceptable discriminative function in our VA ICU population, but it did not perform as well as it did in the original cohort study. Calibration was poor. These discrepancies are likely due to our population’s older age, male predominance, and multiple comorbidities. Based on our population, we developed a simplified BAG score to predict survival.

CLINICAL IMPLICATIONS: BAG score could give clinicians a quick and accurate assessment of severity of illness. Although this scoring system is not intended to replace benchmark prognostication tools, a low score could alert clinicians of a high risk of mortality. This score needs to be further validated in other patient populations.

DISCLOSURE: The following authors have nothing to disclose: Simon Yau, Siva Bhavani, Charlie Lan, Kiran Nair, Gloria Li, Sarah Perusich

No Product/Research Disclosure Information




Citing articles are presented as examples only. In non-demo SCM6 implementation, integration with CrossRef’s "Cited By" API will populate this tab (http://www.crossref.org/citedby.html).

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging & repositioning the boxes below.

Find Similar Articles
CHEST Journal Articles
PubMed Articles
  • CHEST Journal
    Print ISSN: 0012-3692
    Online ISSN: 1931-3543