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


Evidence-based medicine

ACP J Club. 1991 Mar-April;114:A16. doi:10.7326/ACPJC-1991-114-2-A16

An internist sees a 70-year-old man whose main problem is fatigue. The initial investigation reveals a hemoglobin of 90 g/L. The internist suspects iron deficiency anemia. How might she proceed?

The way of the past

When faced with this situation during her training just a few years earlier, the internist was told by the attending physician that one ordered serum ferritin and transferrin saturation and proceeded according to the results. She now follows this path. If both results come back below the laboratory's lower limit of normal, she will make a diagnosis of iron deficiency anemia, and investigate and treat accordingly. If both results are above the laboratory's cut-off point, she will look for an alternative diagnosis. If the results of the tests conflict, she can proceed according to her own clinical instincts, ask a more senior colleague or local hematologist how the results should be interpreted, or consult a textbook.

The way of the future

The internist asks herself whether she knows the diagnostic properties of the tests she is considering ordering and realizes she does not. She turns to the microcomputer in her office, which has a modem and inexpensive software to link by telephone to MEDLINE. She conducts a quick, computerized literature search, using the indexing terms “iron deficiency anemia” and “sensitivity and specificity,” and retrieves seven citations at a cost of $0.79. When she surveys the titles, one appears directly relevant (1). She faxes the citation to the library at the local hospital and picks up the article when she does rounds the next morning. She reviews the paper and finds that it meets criteria she has previously learned about validating a diagnostic test (2) and that the results are applicable to patients like hers.

The study shows that she should order a serum ferritin level, but not transferrin saturation, which is less powerful and adds no useful information. She also finds that her laboratory's normal range for the test is misleading. The internist estimates the pretest likelihood of iron deficiency and orders the test. When the result is available, she uses data from the article to determine the sensitivity and specificity associated with the serum ferritin value obtained, calculates the post-test probability of iron deficiency, and then decides on further management.


The way of the future described above depicts an important advance in the inclusion of new evidence into clinical practice. Clinicians were formerly taught to look to authority (whether a textbook, an expert lecturer, or a local senior physician) to resolve issues of patient management. Evidence-based medicine uses additional strategies, including quickly tracking down publications of studies that are directly relevant to the clinical problem, critically appraising these studies, and applying the results of the best studies to the clinical problem at hand. It may also involve applying the scientific method in determining the optimal management of the individual patient (3).

For the clinician, evidence-based medicine requires skills of literature retrieval, critical appraisal, and information synthesis. It also requires judgment of the applicability of evidence to the patient at hand and systematic approaches to make decisions when direct evidence is not available. The primary purpose of ACP Journal Club is to help make evidence-based medicine more feasible for internists by extracting new, sound clinical evidence from the morass of the biomedical literature so that practitioners can get at it.

Gordon H. Guyatt, MD, MSc


1. Guyatt GH, Patterson C, Ali M, et al. Diagnosis of iron-deficiency anemia in the elderly. Am J Med. 1990;88:205-9.

2. Sackett DL, Haynes RB, Guyatt GH, Tugwell P. Clinical Epidemiology, a Basic Science for Clinical Medicine. 2nd ed. Boston: Little, Brown and Company. [In press for 1991].

3. Guyatt GH, Keller JL, Jaeschke R, et al. The n-of-1 randomized controlled trial: clinical usefulness. Our three-year experience. Ann Intern Med. 1990;112:293-9.