ICU Outcome Predictions using Physiologic Trends in the First Two Days.

Kayaalp M

Computing in Cardiology (39)977–980.


Aims—This study aims to accurately predict patient mortality in the ICU. Given all physiologicmeasurements in the first 48 hours of the ICU stay, the Bayesian model of the study predicts outcomewith a posterior probability.Methods—This study modeled the outcome as a binary random variable dependent on trends ofdaily physiologic measures of the patient, where trends were conditionally independent given theoutcome. 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 thecorresponding day.Results—The prediction performance of the model was measured as the minimum of sensitivityand positive predictive values. The model yielded a score of 0.39 along with a Hosmer-LemeshowH 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 establishedICU scoring metric SAPS-I, whose score was 0.32. Calibration of the model outputs was comparableto that of SAPS-I.

Kayaalp M ICU Outcome Predictions using Physiologic Trends in the First Two Days. 
Computing in Cardiology (39)977–980.