PhD

dissertationPhysicians have unmet information needs at the point of care. These needs have been recognized for many years and yet continue to exist despite extensive efforts by educators through Continuing Medical Education (CME), Evidence Based Medicine (EBM) programs, and publications intended to meet those needs. Most of the research done to meet information needs at the point of care has evaluated Medline, accessed via the search engine PubMed, but has had little success as noted by limited use of the application at the point of care and studies showing decreasing usage over time. The work described in this dissertation lowers the barriers to information retrieval by decreasing the time and number of mouse clicks required to find an answer, offering the user after one mouse click potential questions the system can answer, and tailoring the application to meet the user's information needs based on the clinical context under consideration. The application was implemented in late September 2002 and evaluated over a 27-month period. The assumption is: if the barriers to information retrieval are lowered, physicians will use the application more over time. The usage of the application was measured objectively and subjectively through an audit table and Likert questionnaire. The use of the infobutton grew in a linear fashion over the 27-month period by a group of experienced practicing physicians. Physicians who used the infobutton used it more as time progressed. Of the physicians who responded to the Likert questionnaire 87% (20 out of 23) thought the infobutton added significant value to the Electronic Medical Record (EMR). In concluding we agree with Ely who thinks that not only should doctors frame their questions better but the authors really need to frame their answers better. As this study and others have shown the current information need at the point of care is not for better quality information but faster access to quality information that answers specific clinical questions at the point of care