Park JH, Shin SD, Song KJ, et al. - Researchers trained, validated and compared predictive models that use machine learning analysis for good neurological recovery in OHCA patients. Researchers analyzed 19,860 adult OHCA patients with presumed cardiac etiology and a sustained return of spontaneous circulation between 2013 and 2016. They analyzed 80% of the individuals for training and 20% for validation. Using six machine learning algorithms, they developed logistic regression (LR), extreme gradient boosting (XGB), support vector machine, random forest, elastic net (EN), and neural network. As per outcomes, the highest discrimination powers were seen with LR, XGB, and EN models, and all three were well calibrated. Patients were reclassified better according to their true risk using the XGB model vs the LR model, but the EN model vs the LR model resulted in a worse reclassification of patients. As per these findings, the XGB and LR algorithm are the best performing machine learning algorithm.