Critical care medicine--a review of the outcome prediction in critical care.

Predicting the outcome of critically ill patients has undergone considerable evolution over the last two decades. Economic constraints require the channelling of resources and efforts to patients with reasonable chances of survival. This paper reviews the development of the Therapeutic Intervention Score (TISS), Simplified Acute Physiological Score (SAPS), Acute Physiological And Chronic Health Evaluation (APACHE) Score and Mortality Prediction Model (MPM) systems which have been extensively validated for large groups of critical care patients with mixed diagnoses and found to correlate well with observed outcome in general. They allow evaluation of new therapies, monitoring of resource utilisation and quality assessment of intensive care units. The issues surrounding the use of the APACHE score from the literature are discussed. Estimation of specific individual mortality risks have not been as successful as that for large groups of critical care patients in general using the available prognostic scores and until better systems are available, clinical decisions on any individual outcome should not be based on any prognostic score alone.