ICU severity of illness scores: APACHE, SAPS and MPM

This review aims to evaluate the latest versions of the Acute Physiology and Chronic Health Evaluation, Simplified Acute Physiology Score and Mortality Probability Model scores, make comparisons and describe their strengths and limitations. Additionally, we provide critical analysis and recommendations for the use of these scoring systems in different scenarios.

The last generation of ICU scoring systems (Acute Physiology and Chronic Health Evaluation IV, Mortality Probability Model 0-III (MPM0-III) and Simplified Acute Physiology Score 3) was widely validated in different regions of the world and in distinct settings comprising general ICU patients as well as specific subgroups such as critically ill cancer patients, cardiovascular, surgical, acute kidney injury requiring renal replacement therapy and those in need of extra-corporeal membrane oxygen. Conflicting results are reported, and in general the scores presented a good discrimination despite a worse calibration as compared with the ones described in the original studies that generated them. Nonetheless, such calibration is often improved when customizations are performed both at ICU and region or country level.

ICU scoring systems provide a valuable framework to characterize patients' severity of illness for the evaluation of ICU performance, for quality improvement initiatives and for benchmarking purposes. However, to ensure the best accuracy, constant updates as well as regional customizations are required.