This document summarizes a study that analyzed mortality risk in patients by using lactate and lactate dehydrogenase (LDH) levels with support vector machines (SVM). The study used data from 686 patient records, including lactate levels and other variables. SVM was used to classify patients and predict survival based on three different variable sets. The best performing model used age, LDH, sex, diagnosis code, and lactate levels, and correctly classified over 98% of patients. The study concludes that analyzing medical data with machine learning can help with treatment planning and assessing patient risk, but requires multidisciplinary expertise and high quality data.