5/8/99

In the fall of 1996, the Department of Medicine reorganized Journal Club. We worried that the format of traditional journal clubs only taught appraising medical evidence. However, physicians must first find evidence in order to appraise it and finding evidence is no easy task. Studies have shown that MEDLINE is very difficult and time consuming for physicians to use. AFTER two one-hour training sessions, physicians, as compared to a reference librarian, only find about half of the articles relevant to their question. At the same time, physicians find twice the articles unrelated to their question (1). As a consequence, physicians frequently do not even try to answer questions that arise during clinical practice (2).

Thus we placed a computer in the main medicine ward and clinic of the Audie Murphy VA Hospital. The computers housed the best software available for finding medical evidence - MEDLINE (with Ovid Client), ACP Journal Club, Cochrane Library, Scientific American Textbook of Medicine, and the Internet. The computers monitored how often each program was used. An extensive Windows help file provided additional instruction (download). After several months, a brief qualitative review of the logs revealed that nobody used the computers.

We decided that simply placing a menu of resources that was displayed when someone clicked the small 'Start' button in the corner of the screen was inadequate. We used Visual Basic® to create a interface (a Windows shell) that was always visible on the computer's monitor, guided the physician to the best resource, and provided online help. After two years of upgrades we had an improved product that managed resources over the local area network and the internet.

The new interface greatly increased use of the computers. Some of the computers had 200-300 'hits' per month. However, several more problems arose.

The computer logs showed physicians were only using MEDLINE and underusing resources such Database of Abstracts of Reviews of Effectiveness and Scientific American. Conversations and observation of residents doing searched showed they had difficulty in using MEDLINE. These problems led to the creation of a course on medical informatics. The course emphasized using constructing search questions, choosing the best resource, and more specific concepts such as MEDLINE filters and systematic reviews.

Second, the Windows® interface limited expansion of our project. As the interface improved, it had more complex interactions with the Windows operating system. Our software already excluded Mac computers, and now the software was difficult to deploy with some computers that used Windows. The administatration of our teaching hospitals were reluctant to install our custom software, but were very willing to link to our web pages. Thus we created a web version of our clinical resources (Clinical.UTHSCSA.edu). The importance of using the Internet to expand the number of computers with access to our resources outweighed the loss of the excellent Windows versions of Ovid Client and Scientific American.

We replaced this Windows-based interface with UT-Browser, a customized web browser.

We quantified the impact of our efforts in a quasirandomized controlled trial. The informatics class for our third year medical students began January, 1998. As this was half way through the academic year, we taught half of the students by the end of the year in July 1998. As compared to the students who did not receive the class, the 'taught' students had improved knowledge of how to search. However, neither self reported or measured use of the computers show the class increased the amount of searching done by the students (details of presentation at the Clerkships Directors of Internal Medicine, Denver, CO 9/98).

Medical informatic now has excellent information readily available, however, considerable experience and effort are needed to become adept at quickly locating evidence during the time constraints of patient care. We believe most physicians do not have the time to develop their searching skills. The next step in medical informatics is to make the access of medical evidence easier and faster. Parallel meta-searching is an important part of the solution. Parallel meta-searching means the physician enters the search query one time, and the computer then decides where to search, formats the search for each resource, and organizes the results in a single, easy to read document.

The Department of Medicine has now developed SUMSearch for this purpose (SUMSearch.UTHSCSA.edu) to achieve this purpose.

References:

1. Haynes RB, et al. Online Access to MEDLINE in clinical settings: a study of use and usefulness. Ann Intern Med 1990;112:78-84.

2. Covell DC, et al. Information needs in office practice: are they being met? Ann Intern Med. 1985 Oct;103(4):596-9.

3. Wilczynski NL et al. Reasons for the loss of sensitivity and specificity of methodologic MeSH terms and textwords in MEDLINE. Proceedings - the Annual Symposium on Computer Applications in Medical Care: 436-40. 1995.

4. Williamson JW, et al. Health science information management and continuing education of physicians. A survey of U.S. primary care practitioners and their opinion leaders. Ann Intern Med. 1989 Jan 15;110(2):151-60. PMID: 2909205; UI: 89075473.