Current issues of ACP Journal Club are published in Annals of Internal Medicine


Editorials

Using Best Evidence in clinical practice

ACP J Club. 1998 Mar-April; 128:A15. doi:10.7326/ACPJC-1998-128-2-A15



The 1998 edition of Best Evidence has just been released. We are proud of it, but given our involvement in assembling it, reluctant to boast. We will tell you about its benefits so you can decide for yourself whether it might help you in providing evidence-based medical care.

What is Best Evidence?

Best Evidence is the electronic presentation of all issues of ACP Journal Club (ACP JC) (1991 to 1997) and Evidence-Based Medicine (1995 to 1997). More than 1300 articles, carefully selected for scientific merit and clinical relevance, are represented by enhanced structured abstracts and expert commentaries. Best Evidence also includes 54 editorials from ACP JC and notes from Evidence-Based Medicine. All entries in ACP JC in 1991 and 1992 have been examined by the editorial staff, and outdated files have been archived; the remaining entries have been revised to reflect current editorial standards. Electronic links connect similar articles, letters, corrections, and updates. Best Evidence includes over 3000 listings from the Other Articles Noted pages of ACP JC. These articles passed our selection criteria but could not be abstracted because of space constraints.

Best Evidence is distributed by the American College of Physicians (800-523-1546; www.acponline.org), the BMJ Publishing Group (44-171-383-6260; www.bmjpg.com/data sjindex.htm), and the Canadian Medical Association (800-663-7336; www.cma.ca). It will also be available on the Internet in 1998 through OVID Technologies (www.ovid.com) as part of its new service, Evidence Based Medicine Reviews.

Why would you use Best Evidence?

The most important use of Best Evidence is for solving clinical problems in "real time." It also supports teaching, presentations, and quality-improvement activities. The articles in Best Evidence are both methodologically sound and important to the practice of clinical medicine. The articles are primary studies or systematic reviews on the treatment, prevention, diagnosis, cause, prognosis, clinical prediction, economics, or quality improvement of common or relatively common medical conditions.

ACP JC and Evidence-Based Medicine are published bimonthly and include articles from approximately 85 journals. These articles are selected to keep readers up to date on major advances in medicine that have been published since the previous issue, with the goal of increasing current awareness. Best Evidence, by contrast, is a compilation of evidence collected over 7 years, with weeding of outdated items to provide a database of current best evidence.

How would you use Best Evidence?

Best Evidence is available on compact disk and can be installed on 1 computer (it is about 23 megabytes in size) or on a network for multiple users. Because the database is relatively small (1300 citations vs 9 000 000 citations in MEDLINE), searching is simple and very fast. For example, searching for studies on "amiodarone" in Best Evidence takes about 15 seconds, which includes the time to start the program; type in the search term; and retrieve 22 citations, all of which provide high-quality evidence on the clinical use of amiodarone. The same computer hooked up to the Internet with a fast link takes about 6 times as long and retrieves 1051 citations (using PubMed, www.ncbi. nlm.nih.gov/PubMed). In addition to a few trials, these 1051 MEDLINE citations include animal and laboratory studies, case reports and case series, letters to the editor, news items, and editorials. A more complex strategy is needed to refine the search on MEDLINE (1), and examination of the individual items to discern the most valid and relevant studies can be time consuming and frustrating, given the lack of detail in standard abstracts.

Single terms are usually sufficient for searching in Best Evidence, but sometimes multiple terms are needed. For example, "heparin" retrieves 109 citations, whereas "low molecular weight heparin" retrieves 38 citations. As another example, searching for "gold" for rheumatoid arthritis retrieves 46 citations, but of the first 13, only 3 discuss therapeutic agents for rheumatoid arthritis, 1 deals with gold miners, 1 deals with optimal shoulder repair, and the other 8 describe "gold standards" in diagnostic studies. A much better search would be "gold and rheumatoid," which retrieves 8 citations. The Best Evidence search engine permits use of "and," "or," "near," and "not."

Best Evidence also has an index for searching. Index terms are best for topics that are clearly defined. For example, an index search for the combination trimethoprim-sulfamethoxazole takes only 6 keystrokes because the index search automatically provides the best match with the initial keystrokes; a full-text search for the drug combination takes many more keystrokes and good spelling skills.

Additional features

Users can copy all or part of an item to a disk or paper, annotate articles, and bookmark items of interest (e.g., articles that were used for rounds presentations or for specific needs). History and Go Back buttons allow backtracking. Go To, Search, and Index buttons facilitate retrieval, and button choices in the Contents provide chronological listings of items. Tables and figures are included and enhanced for easier review, and Help is always available.

In creating Best Evidence, we've tried to produce an efficient resource for supporting clinical decisions with current evidence from research. When Best Evidence is judiciously combined with clinical experience and the needs and expectations of patients, you'll find the result is a successful formula for evidence-based decision making.


Reference

1. McKibbon KA, Walker-Dilks CJ.Beyond ACP Journal Club: How to Harness MEDLINE for Therapy Problems ACP Journal Club. 1994 July-Aug;121:A-10.