Natural Language Processing in Medicine: A Review

Abstract Natural language processing (NLP) is a form of machine learning which enables the processing and analysis of free text. When used with medical notes, it can aid in the prediction of patient outcomes, augment hospital triage systems, and generate diagnostic models that detect early-stage chronic disease. These applications may be particularly useful in critical care where there is more patient data to analyse and prediction of patient mortality is routine. In addition to its natural language understanding (NLU) ability, NLP can also accomplish natural language generation (NLG), providing an interface for patients to ask questions and access relevant information in the form of chatbots. There are challenges to the use of NLP in medicine. Unbiased training data is an essential requirement if the conclusions reached by NLP algorithms are to be trusted. Clinicians will need training to understand how NLP can be safely used as part of routine practice. In the future, NLP applications are likely to be integrated into the clinical environment, working with clinicians to suggest problem lists, as patient facing applications to streamline triage systems, and as a tool to interrogate vast amounts of free text data, which could contribute to personalised, up to the minute evidence based medicine.