ABSTRACT
BACKGROUND: Results of automated clinical chemistry tests are affected by many factors including analytical variability. In 1976, the College of American Pathologists (CAP) Conference on the analytical goals in clinical chemistry recommended that analytical variability should be less than 1/4 of the appropriate biological variability to improve distinction between normal and diseased populations. This study is conducted to evaluate whether automated clinical chemisty analyses performed in our laboratory is in conformance with the CAP's recommendation. METHODS: Routine chemistry and electrolyte tests were performed using Hitachi 747 automatic analyzer on 22 healthy volunteers. Blood samples were obtained from the volunteers' same vein twice in one week interval to study the total variability. Serum samples from 12 subjects were tested in duplicate immediately after blood collection for within-run analytical variability; and samples from another 10 subjects were repeated after 6 hours for within-day analytical variability. Within-run analytical variability plus within-day analytical variability make total analytical variability. Biological variability was defined as the difference between total variability and the analytical variability. Finally, ratios of analytical and biological variabilities were calculated. RESULTS: The ratios of analytical and biological variabilities of uric acid, glucose, and K were less than 0.25. But ratios of BUN, PO4, alkaline phosphatase, total bilirubin, AST, cholesterol, ALT, Cl, and protein exceeded 0.25. The ratios of Na, Ca, albumin, CO2, and creatinine could not be calculated. CONCLUSIONS: It is suggested that the analytical processes of the automated clinical chemistry tests be improved so as to be in conformity with the CAP's recommendation.