Informatics in Medicine Unlocked (Jan 2022)
A multicenter study for establishment and evaluation of auto-verification rules for routine coagulation tests
Abstract
Background & Objectives: Coagulation test results in clinical laboratories are released after manual verification, which increases the laboratory turnaround time (TAT). The purpose of this study was to establish auto-verification rules for routine coagulation tests that can reduce the laboratory TAT as well as ensure the accuracy of results. Methods: Auto-verification rules were established for six routine coagulation tests, i.e., prothrombin time, activated partial thromboplastin time, thrombin time, fibrinogen (FIB), fibrin/fibrinogen degradation product, and D-Dimer. Data were collected from nine centers and the auto-verification rules were designed with reference to CLSI Auto 10A guidelines, which included quality control (QC) status check, consistency check (checking whether sample test assays match the medical order or if results are in correct numerical form), sample quality check, alarm check, critical value check, delta check, limit range check, and logical rules check. The auto-verification rules were then assessed based on passing rate, true-positive rate, true-negative rate, false-positive rate, and false-negative rate. The efficiency of the auto-verification rules was evaluated by TAT. Results: From January 2020 to March 2020, 22,072 coagulation results were collected from nine hospitals for establishing auto-verification rules. The preliminary evaluation of the rules was a simulation validation with data from the rule-generating group, and the passing rate was 69.5% with no false-negative cases. In phase I evaluation, a total of 32,743 results were randomly collected from four hospitals. The overall auto-verification rate was 66.7% (true-positive rate 100.0%; true-negative rate 82.5%; false-positive rate 17.6%; false-negative rate 0%). The phase II evaluation was based on validation of results of actual samples. Auto-verification rules were configured in a middleware (HemoHub) and 14,642 samples results were verified by the auto-verification rules. The overall passing rate was 69.5%. Auto-verification significantly decreased the mean laboratory TAT from 90.8 min to 71.8 min (P = 0.006). Conclusion: This study established auto-verification rules for routine coagulation tests based on Werfen coagulation instrument. Use of auto-verification rules significantly improved the laboratory TAT and workflow. This study provides insights and methods for establishing auto-verification rules.
