Introduction:

scoring systems for prediction have been developed to measure the severity of the disease and the prognosis of patients in the intensive care unit. They are useful for clinical decision-making, standardizing research and comparing the quality of care.

Materials and Methods:

an observational cohort analytical study in which the medical records of 283 oncological patients admitted to the intensive care unit (ICU) of the CIOSAD Oncology Research Center from January 2014 to January 2016 were reviewed. The probability of mortality with the APACHE IV and MPM II prognostic scores was estimated, binary logistic regression was performed with the variables of the models in their original studies, calibration, discrimination was determined and Akaike AIC and Bayesian BIC information criteria were calculated.

Results:

in the evaluation of the performance of the prediction scores, APACHE IV showed greater predictive capacity (AUC = 0.95) versus MPM II (AUC = 0.78); the two models showed adequate calibration with Hosmer and Lemeshow statistics for APACHE IV (p = 0.39) and for MPM II (p = 0.99). The BIC delta is 2.9 showing positive evidence against APACHE IV. The AIC statistic is lower for APACHE IV indicating that it is the score with the best fit to the data.

Conclusions:

APACHE IV has a good performance in the prediction of mortality of critically ill oncologic patients. It is a useful tool for the clinician in his daily work, allowing him to distinguish patients with a high probability of mortality. (Acta Med Colomb 2018; 43: 81-89).