Random errors in haematology tests: a process control approach.

Abstract

Random errors are usually not detected by classic quality-control methods, because these can occur not only during the analytical procedure itself, but during any step of the entire process of generating test results: from patient specimen collection till the final result is being charted. Random errors therefore require a variety of methods for prevention and detection. Process control enables the analysis of the causes and categories of random error, while it also assists in the design and implementation of methods for prevention and detection. In this paper the various aspects of such an approach are discussed. Among the suggested measures that will reduce the frequency of random error are: specimen labels with machine-readable codes, application at the site of specimen collection; instrument interfacing to a laboratory computer system; and streamlining of data entry. Recommendations for random error detection systems include parameter-flagging criteria, internal consistency checks of test results, and delta checking. These measures need to be part of a systems approach to be most successful, and will result in a significant reduction, although not in a complete elimination, of both the incidence of random errors and the frequency with which test results containing random errors will be reported.