Project Members: Saverio Maviglia and Gil Kuperman

The volume of medical knowledge continues to expand at a pace that makes it impossible for the individual clinician to keep up. Furthermore, difficulty in accessing information can result in errors; one study evaluating the systems failures associated with adverse drug events found that better information could have prevented half of serious medication errors, and that poor dissemination of drug-specific knowledge accounted for the single largest proportion of medication errors (29%). While this information was available somewhere (either on paper or electronically), it was not at the providers’ fingertips. We believe that information technology should anticipate clinicians’ needs, and bring the information they require to the point of care. To address these issues, we are developing an application extender called KnowledgeLink, the goal of which is to provide “just-in-time” context-sensitive information retrieval.

Within relevant clinical applications such as order entry, lab results, and problem lists, KnowledgeLink will consist of specially designated “information buttons” that will initiate automated queries into various web-based medical resources already licensed by Partners. The queries will be generated dynamically based on the clinician’s current activity within the application. For instance, in order entry, the button could display information about dosing, indications, warnings, pregnancy effects, or toxicology for the medicine currently being prescribed. The same button in the problem list application could initiate automated literature searches to review diagnostic or therapeutic guidelines, natural history, or epidemiology of a selected disease. A similar function while reviewing lab results could retrieve a differential diagnosis of a highlighted set of labs. The queries are context-sensitive in that the topic (“digoxin”, for instance) is automatically inferred and does not have to be re-specified. In this way, answers to queries likely to arise during a clinical activity such as ordering medications or reviewing lab results can be obtained exactly at the time such queries arise and with minimal interruption to the activity itself.