Related Experiment Videos
Predicting the risk of death in patients in intensive care unit
Asadolah Saadat-Niaki1, Dariush Abtahi
1Department of Anesthesiology, Shaheed Beheshti University of Medical Sciences, Tehran 16179, Iran. saadatniaki@yahoo.com
Background:
The ability to identify critically ill patients who will not survive until hospital discharge may yield substantial cost savings. The aim of this study was to validate the mortality prediction model II (MPM II) in in-hospital mortality of critically ill patients for quality management and risk-adjusted monitoring.
Methods:
The data were collected prospectively from consecutive admissions to the Intensive Care Unit of Imam Hossein Medical Center in Tehran. A total of 274 admissions were analyzed using tests of discrimination and calibration of the logistic regression equation for mortality prediction model II at admission (MPM0 II) and at 24th hour (MPM24 II).
Results:
The mortality prediction model II exhibited excellent discrimination (receiver operating characteristic curve area). Calibration curves and Hosmer-Lemeshow statistics demonstrated good calibration of both models on outcome.
Conclusion:
We recommend using mortality prediction model II in Iranian ICUs for routine audit requirements. Mortality prediction model II is not affected by the standards of treatment after admission to ICU. The information needed to calculate mortality prediction model II is easy to collect, and the model is applicable to all ICU admitted patients.