Bringing evidence to the point of care.

Evidence-based medicine calls for the integration of our clinical expertise and our patients' values with the best available external evidence (1). To accomplish this, we must translate our information needs into answerable questions and then seek the best information with which to answer them (2). Critics of evidencebased medicine have appropriately suggested that its practice may require time and resources unavailable to busy clinicians (3). Davidoff and colleagues suggested that a general physician would need to read 17 articles each day to keep up with the current medical literature (4). Clinicians have little time to set aside for keeping up-to-date or for reading in between seeing patients, and self-reported weekly reading times reflect this: Consultants report reading from 30 to 45 minutes each week, while house officers report from 0 to 20 minutes (5), Further, some clinicians don't have quick access to the evidence and may have to travel several floors, blocks, or even miles to visit their local library to find it. Proponents of evidence-based medicine are meeting these challenges in various ways. The development of the Cochrane Library, along with the growing number of evidence-based journals of qualityand relevance-filtered secondary publications (e.g., Evidence-Based Medicine and Evidence-based Health Policy and Management) and the creation of "best evidence" sections in a number of established journals (e.g., Journal of Pediatrics and Journal of Family Practice), have resulted in better accessibility of high-quality evidence for clinicians. The Cochrane Library is available on CD and on the Internet for electronic searching. Family physicians have expressed great interest in having patient-education materials, drug information, and overviews of treatment recommendations available on computers (6). Housestaff have stated they like having access to high-quality evidence at the bedside but would prefer to have it available for use within 30 seconds (7). These surveys suggest that we need to find ways of speedily delivering high-quality evidence to the point of care. Several groups are working to find ways to provide quick access to the evidence. We recently used an evidence cart (7) that included a notebook computer containing Best Evidence (a CD containing ACP Journal Club and Evidence-Based Medicine), the Cochrane Library, MEDLME, and a collection of 149 one-page summaries of critically appraised topics (CATs) created in response to clinical questions posed by previous members of our clinical teams (8). The cart (housed in a seminar room between our wards) was used on our clinical and teaching rounds and whenever a member of the clinical team posed a question. The resources that could be accessed the fastest (we could find specific 1-page summaries in under 12 seconds) were used most frequently and most often resulted in a successful search. Our speed of access to evidence using these resources was probably more rapid because these resources were developed by otir clinical teams for this purpose and, therefore, their contents were well known to us. However, such customised information is being developed in a number of settings. We could access specific entries in Best Evidence in 23 seconds (about the upper limit of time we were willing to devote to searching during our rounds) and found that we could answer 16 questions using this resource on die ward in the time it took us to answer just one using the library (4 floors away). MEDLINE was successful at answering our questions (59% of the time) but was so slow (mean time for a successful search was 90 seconds) that most searches were completed outside of rounds. Similarly, the evidence in the Cochrane Library was superb, but it took too long to find during rounds. The evidence found altered the clinical approach of at least 1 team member 48% of the time. During the month after removal of the cart, the team iden.•••§'•'.: ^ . I