Machine learning-based error detection in the clinical laboratory: a critical review.

Abstract

Laboratory test results play a crucial role in the modern medical decision-making process. As such, errors in any phase of the testing process can have substantial clinical and operational impacts. While the development of increasingly robust quality assurance systems has enhanced the reliability of laboratory results, opportunities for improvement still exist. Machine learning approaches offer the potential to evaluate complex patterns and discriminate physiological variation from laboratory errors. In this work, we critically evaluate the current state of published machine learning solutions to laboratory errors, while highlighting unmet needs and potential barriers to widespread implementation.