Automatic processing of Electronic Medical Records using Deep Learning

Background. Availability of large amount of clinical data is opening up new research avenues in a number of fields. An exciting field in this respect is healthcare, where secondary use of healthcare data is beginning to revolutionize healthcare. Except for availability of Big Data, both medical data from healthcare institutions (such as EMR data) and data generated from health and wellbeing devices (such as personal trackers), a significant contribution to this trend is also being made by recent advances on machine learning, specifically deep learning algorithms. Objectives. The objective of this work was to provide an overview of how automatic processing of Electronic Medical Records (EMR) data using Deep Learning techniques is contributing to understating of evolution of chronic diseases and prediction of risk of developing these diseases and associated complications. Methods. A review of the scientific literature was conducted using scientific databases Google Scholar, PubMed, IEEE, and ACM. Searches were focused on publications containing terms related to both Electronic Medical Records and Deep Learning and their synonyms. Results. The review has shown that a number of studies have reported results that provide unprecedented insights into chronic diseases through the use of deep learning methods to analyze EMR data. However, a major roadblock that may limit how effectively these paradigms can be utilized and adopted into clinical practice is in the interpretability of these models by medical professionals for whom many of them are designed. Conclusions. Despite the identified challenges automatic processing of EMR data with state-of-the-art machine learning approaches, such as deep learning, will push predictive power well beyond the current success rates. Hopefully, we will continue to see findings from these works to continue to transform clinical practices, leading to more cost effective and efficient hospital systems along with better patient outcomes and satisfaction.