Development and validation of autoverification system for biochemistry analysis in Hacettepe University central Laboratory

The autoverification (AV) system in medical laboratories is a process in which computer-based algorithms automatically take actions on laboratory test results without the need for manual intervention by laboratory employees within the framework of predefined rules, or it is a system that enables the evaluation of results with more advanced algorithms. Although it has been a topic of discussion for a long time, there is a lack of standardization in AV systems. There are very few studies in the literature detailing how the rules and algorithms related to AV systems will be implemented. For these reasons, in this study, we aimed to reveal data on the establishment, development and validation of an AV system in our laboratory. AV system software was developed in the Laboratory Result Validation Module, which was designed within the Laboratory Information System in Hacettepe University Hospitals Central and Emergency Laboratories, and integrated with the laboratory module. Validation of the AV system was performed by simulated and real patient results. Defined rules were evaluated for a total of 29 biochemistry tests. The results of 24 115 patients at Hacettepe University Hospitals Laboratories between 1 May 2020 and 1 July 2020 were evaluated by AV system. Out of a total of 793 300 test results, 676 083 (85.22%) results were verified by AV system, while 117 217 (14.78%) test results were not. Among the unverified results, the "Unexpected results" criterion was the highest with 59 413 test results (7.49%). Among 138 029 results unverified by the AV system, unexpected results (24.89%) and delta control (%14.08) were found to be highest in outpatient admissions and unexpected results (%18,15) and serum index assessment (%18.15) in hospitalized patients. 4 710 test results were evaluated by 2 users and AV system to compare expert and AV system verifications. AV system is an important tool to standardize the verification process and increase the reliability of the results. Good initial autoverification rates were obtained with the rules determined in the study. More studies are needed to use AV system more efficiently in wider test panels, to develop more detailed rules and algorithms, and to use AV system more widely.