Grant: Engage Grants Program

MD Physician Services Inc. (MDPS), delivers by email, synopses of clinical research called InfoPOEMs?. To transform simple reading of the InfoPOEMs? into an actual learning experience, research in education states the importance of having the reader(learner) reflect on the value of his reading by answering some questions. While questions on the content only test short-term memory, questions on the value of the information for clinical practice can stimulate reflective learning. Researchers at McGill developed the Information Assessment Method (IAM), an individual brief reflective e-learning activity specifically geared toward the continuing education of health professionals. Since 2006, MDPS includes the IAM method with the InfoPOEMs? delivered. Collaborative environments (Web 2.0) show how individual knowledge can benefit other members of a community. MDPS has already implemented a mechanism for finding the ""views of others"" by gathering statistics on the InfoPOEMS rated by the readers. MDPS posts on CMA (Canadian Medical Association) website evaluations of specific InfoPOEM using a ""thumbs up or thumbs down"" graphic. MDPS wishes to move toward the sharing of textual comments from physician raters, but this requires a more complex analysis than computing statistics. MDPS does not have the human resources for reading, gathering and digesting such comments to give back to their members. This proposal suggest that human resources could be, to some extent, simulated by automatic text categorization. This research project is to develop a method, based on a supervised machine learning approach, which could learn to categorize valuable from non-valuable comments from a subset of comments annotated by medical practitioners, and then infer the categories of future comments. As we anticipate this to be a difficult task, confidence estimation must be included to allow MDPS to share the comments that the automatic method would categorize as valuable with a high enough confidence.