DocumentCode :
591371
Title :
ICU outcome predictions using physiologic trends in the first two days
Author :
Kayaalp, Mehmet
Author_Institution :
Lister Hill Nat. Center for Biomed. Commun., Nat. Inst. of Health, Bethesda, MD, USA
fYear :
2012
fDate :
9-12 Sept. 2012
Firstpage :
977
Lastpage :
980
Abstract :
Aims: This study aims to accurately predict patient mortality in the ICU. Given all physiologic measurements in the first 48 hours of the ICU stay, the Bayesian model of the study predicts outcome with a posterior probability. Methods: This study modeled the outcome as a binary random variable dependent on trends of daily physiologic measures of the patient, where trends were conditionally independent given the outcome. A two-day trend is a sequence of two discrete values, one for each day. Each value (low, medium, high or unmeasured) is a function of the arithmetic mean of that measure on the corresponding day. Results: The prediction performance of the model was measured as the minimum of sensitivity and positive predictive values. The model yielded a score of 0.39 along with a Hosmer-Lemeshow H statistic of 36, which measures calibration. The perfect scores would be 1.0 and 0, respectively. Conclusion: The prediction performance of the study was an improvement over the established ICU scoring metric SAPS-I, whose score was 0.32. Calibration of the model outputs was comparable to that of SAPS-I.
Keywords :
Bayes methods; biomedical measurement; calibration; patient treatment; physiological models; random processes; statistical analysis; Bayesian model; Hosmer-Lemeshow H statistics; ICU outcome predictions; binary random variable dependent; calibration; patient mortality prediction; physiologic measurements; physiologic trends; prediction performance; time 48 hr; Bayesian methods; Biomedical measurements; Calibration; Heart rate; Market research; Predictive models;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Computing in Cardiology (CinC), 2012
Conference_Location :
Krakow
ISSN :
2325-8861
Print_ISBN :
978-1-4673-2076-4
Type :
conf
Filename :
6420559
Link To Document :
بازگشت