Proficiency testing in a medical-needs context.

Proficiency testing by State and Federal agencies is an ongoing activity of clinical laboratories and an occasional source of anxiety; strict statistical evaluations of "snapshot" laboratory values are inappropriate and medical-needs criteria should be used. The quality of laboratory results largely depends on available technology. Fortunately, for most of the common clinical chemistry analytes there has been a steady reduction of imprecision during the past 20 years. Proficiency testing may have been the stimulus for this improved performance. Medical-needs criteria differ, depending on the testing goal. For proficiency testing, population screening criteria are appropriate, e.g., the College of American Pathologists fixed criteria for the common tests. Stricter criteria are needed for short-term (inpatient) and long-term (outpatient) monitoring of laboratory data. Explicit proficiency-testing limits are given here for nine of the common clinical chemistry tests for each of the three medical-needs criteria described above. The limits consider total error--i.e., bias from the believed correct value, and imprecision. Rather broad limits are acceptable for the commonly performed enzyme tests when used for screening purposes.