Project title: Development of a clinical question answering system to aid in decision making
Project description: The use of clinical knowledge systems such as UpToDate that provide reliable information at the point of care has been shown to improve patient safety and decision-making. With similar content to UpToDate, Mayo Clinic's AskMayoExpert (AME) is an online knowledge system that primarily contains over 5000 (and increasing) specialist-vetted answers to FAQs for point of care use. However, because of the overabundance of clinical resources and guidelines, adding new answers manually to AME and ensuring that it is consistent with evidence is time consuming. This problem is also faced with other systems such as UpToDate. This project, funded by Dr. Jonnalagadda’s career grant, proposes to investigate the feasibility of using a novel text mining based informatics approach to semi-automate the management of a clinical knowledge system, using AME as the test bed. Although the methods will be applicable to any clinical knowledge system and any topic, they will be evaluated using two important test topics from cardiology (which has the biggest focus in AME) - atrial fibrillation (a topic exhaustively covered in AME) and congestive heart failure (a topic less covered, but is an increasingly complex vast field with knowledge from huge literature). While the existing content of AME is private, the methods and the code we develop to assist in generating the content will be released open-source as part of the Open Health Natural Language Processing (OHNLP) consortium in UIMA framework.