Laboratory error undetectable by customary quality control/quality assurance monitors.

The preanalytical, analytical, and postanalytical rates of laboratory error have not been studied extensively. We evaluated the preanalytical, analytical, and postanalytical components of laboratory error in 438 consecutive samples submitted to a clinical chemistry laboratory for measurement of creatinine concentrations in plasma. We performed red blood cell antigen determinations to establish patient-sample identity, repeated analysis of creatinine in duplicate to detect analytical error, and tracking of patient specimens from receipt by the laboratory to entry of the laboratory result in the patient's information system record. We found a total error rate of 9.36%. A breakdown of the total error rate into its preanalytical, analytical, and postanalytical components revealed error rates of 0.00%, 8.90%, and 0.46%, respectively. These results suggest that preanalytical and postanalytical error, which are not usually detectable by common quality control strategies, are not major sources of laboratory error. Further work is needed to reduce the unacceptably high rate of analytical errors.