Deep Learning Model For Personalized Prediction Of Positive Mrsa Culture Using Time-series Electronic Health Records

This article discusses a deep learning-based predictive model for MRSA-positive cultures in hospitalized patients. The model showed better performance compared to traditional machine learning models and successfully predicted positive cultures not only on the index day but also over two weeks from the index day. The model also “learned” patient-specific features and was able to transfer to different datasets and tolerate imbalanced outcomes.