Failure Mode Effects Analysis of Auto-verification in the Clinical Laboratory
Graduation Date
5-2018
Document Type
Master's Thesis
Degree Name
Master of Science
Department or Program
Clinical Laboratory Sciences
Department or Program Chair
Mary Sevigny, PhD
First Reader
Michael Dacoco, MT (AMT), CLS, MHA, DHA
Second Reader
Maria C. DeSousa, JD, MPA, CLS
Abstract
Today’s hospital-based clinical laboratories are under increasing pressure to simultaneously contain costs, improve quality and reduce turnaround times. Fortunately, advances in analytic and information system technologies allow for automation of some pre-analytical, analytic and post-analytic laboratory processes. This paper addresses the use of computer algorithms to increase the reliability and timeliness of the post-analytical process commonly referred to as verification. Prior to the availability of these technologies, each and every clinical laboratory result required a clinical laboratory scientist to review data before approving the release of the result to the electronic health record. The process of manual verification is inherently time-consuming and prone to considerable inter-rater variability. This paper summarizes findings of a variety of studies that evaluate the risks and benefits of auto-verification. Failure Mode Effect Analysis is used to describe and characterize the risks associated with common clinical laboratory practices. Prothombin Time (PT) is used to illustrate a variety of parameters to be considered before implementing auto-verification including: patient test distribution, individual patient test variability over time, test precision and quality control stability. We piloted an algorithm for auto-verification of PT at two NCAL Kaiser Permanente medical center laboratories over three months. At the time of publication, the pilot sites identified a need to refine the algorithm in order improve efficiency. Study findings identify the need to take a methodical approach to algorithm development and to seek input from all stakeholders to increase the likelihood of a successful implementation.
COinS