*Krishnaj Gourab, MD
Oakland University William Beaumont School of Medicine
Demo-Interactive – Research Track
Saturday, Sept 29, 2012: 1:20 PM – 1:35 PM – Demo Pavilion
(Research in Progress)
Substantial progress has been made in making electronic resources for medical education available to trainee physicians. It is also recognized that for effective learning, these resources are best delivered and accessed at the point of patient care. However, during more involved patient care situations (either the requirement of being able to evaluate a patient within a limited time or during management of a fast-evolving critical situation) there are time constraints to accessing these resources. Paradoxically, these are also the situations when the potential for utilizing, retaining and learning from these resources are the maximum.
The purpose of this abstract is to present a proof of concept for automated generation and presentation of case-relevant medical knowledge objects from a peer generated database during documentation of a clinical encounter.
Prototype software algorithms written in MATLAB (Mathworks Inc. Natick, MA) were implemented for following: (i) recognition of keywords denoting medical condition or facts (for example, left hemiperesis, hyponatraemia, acute mental status changes, etc.) in the electronic clinical encounter documentation (progress note, history and physical etc.) (ii) generation of a set of synonyms or symphrases for the identified keywords from a library hosted in a shared digital database (iii) selecting knowledge objects related to the keywords, their synonyms and symphrases from a library of peer generated knowledge objects present in the same shared database (iv) adjusting the relative rank of the knowledge objects based on the trainee physician’s learning objectives (determined by self or supervisor) and (iv) presentation of these knowledge objects as the clinical encounter is being documented by the physician on a portable computer at the patient’s bedside.
The overall objective of this project was to make access to medical educational resources at the point of care relevant, efficient, seamless and integrated with the work flow of patient care. Further research to determine whether such automatic and relevance adjusted presentation of medical knowledge is beneficial to resident education is currently in progress. Software using the above proof of concept is being developed using the C# programming language.