With justification, physicians have taken pride in their diagnostic acumen. Few diseases now escape detection in careful clinical examinations coordinated with appropriate laboratory investigations. To assure continuing success, the results of laboratory tests, procedures that play progressively greater roles in our evaluations of patients, must fulfill the following three criteria: (1) be precise and reproducible, (2) indicate "normal" ranges for defined populations, and (3) by identifying a present condition, predict outcomes for patients.
Methods of quality control, when applied with diligence, guarantee that the first criterion is met for most in vitro types of measurements. However, clinicians are sometimes unaware that results from a number of the more complicated studies, eg, catheterization of vessels, are compared with data from relatively few and perhaps not entirely normal persons.
The second criterion is often ignored. Physicians care for people who are sick and, in many cases, have already been treated with a