(PORTLAND, OR) --- Busy physicians who need the best information aren't turning to their computers or the Internet for help. That is the conclusion of William R. Hersh, M.D., of Oregon Health Sciences University, who is publishing two articles in the Oct. 21, 1998, Journal of the American Medical Association. Hersh states that while there is great potential for on-line medical information to improve physician and patient decision-making, we still have a ways to go.
He says that despite the proliferation of electronic information retrieval systems specifically designed for physicians, physicians aren't using them. Physicians are using systems, such as Medline, for only a fraction of their information needs. "Current information technologies are too time-consuming for busy clinicians," says Hersh, who heads the medical school's Division of Medical Informatics and Outcomes Research. "They must break out of their routines to look up something, track down a publication, which may or may not be in electronic form, and read it to assess its value. And typically, there are multiple studies on a clinical finding." At an average of 30 minutes per question, going to the computer on a routine basis is impractical, especially at the point of care, according to the researchers.
When physicians do seek on-line information, their searches turn up only one-fourth to one-half of the relevant articles on a topic. "The variation in language and terminology make it difficult to find everything," Hersh says. "This is true even for the most experienced professional searchers." He says that physicians more often rely on colleagues, textbooks and medical journals. But he notes that textbooks are notoriously out of date and many scientific journals are not relevant to clinical practice.
Hersh and his colleague David Hickam, M.D., also of OHSU and the Veterans Affairs Medical Center, write that despite the growing number of products available at modest cost on computer networks and CD-ROM, it is not clear how well physicians use them, whether they are cost-effective or how often they lead to better clinical decisions. They advocate the development of better systems as well as further research to discover how to best build those systems.
While an information retrieval system is not needed to answer every clinical question that arises, clearly these systems are not playing a major role in meeting overall physician needs, the authors write. They say that more work needs to be done to make electronic information easier for physicians to use. Perhaps less surprising is the Oregon researchers conclusion about the quality of information on the World Wide Web.
In a separate article in the same issue, Hersh and colleagues Paul N. Gorman, M.D., and Lynetta S. Sacherek, M.L.S., both of OHSU, point out that material published on the World Wide Web is either of limited use to clinicians or of poor quality. They found that 89 percent of Web pages did not apply to the question that prompted the search. Furthermore, basic indicators of quality, such as the Web page author or the author's credentials, were missing. Only 1 percent of pages disclosed financial or other conflicts of interest.
The group believes that clinicians are much better off relying on traditional information sources, such as peer-reviewed medical literature and textbooks, many of which are becoming increasingly available on the Web. In the future, Hersh would like to see more systematic and exhaustive reviews of the medical literature available to physicians electronically. A group of researchers or clinicians could collect and review all the studies on a single subject, such as the safety of diagnostic ultrasound in pregnancy. This single review would then be posted on-line, and a midwife or obstetrician could feel relatively confident that all the reputable studies done on the subject would be reflected in the review.
Note: Reporters may call Dr. Hersh directly to arrange an interview at 503-494-4563 or send an email to firstname.lastname@example.org.