Improving the Bias of Comparator Methods in Analytical Performance Assessments Through Recalibration

• Published in: Journal of diabetes science and technology Vol. 18; no. 3; pp. 686 - 694
• Main Authors: Pleus, Stefan, Eichenlaub, Manuel, Gerber, Thomas, Eriksson Boija, Elisabet, Makris, Konstantinos, Haug, Cornelia, Freckmann, Guido
• Format: Journal Article
• Language: English
• Published: Los Angeles, CA SAGE Publications 01.05.2024
• Subjects: Bias; Blood Glucose - analysis; Blood Glucose Self-Monitoring - instrumentation; Blood Glucose Self-Monitoring - standards; Calibration; Humans; Original; Reference Standards; Reproducibility of Results; Retrospective Studies; analytical performance specification; comparator method; recalibration; continuous glucose monitoring system
• ISSN: 1932-2968; 1932-3107
• DOI: 10.1177/19322968221133107

Background: In analytical performance studies, the choice of comparator method plays an important role, as studies have shown that there exist relevant systematic differences (bias) between laboratory analyzers. The feasibility of retrospective recalibration of measurement results through comparison with methods or materials of higher metrological order to minimize bias was therefore assessed. Method: Existing data from performance studies of continuous and blood glucose monitoring systems were retrospectively analyzed. Comparison with a higher-order method was performed for two different data sets. In both cases, subject samples were measured, and a subset was also measured on a higher-order method. Recalibration based on higher-order materials (standard reference material [SRM]) was conducted for two different data sets containing results from SRM and subject samples. Linear regression analysis was performed for each device separately. Resulting equations were applied to the respective complete data set of subject samples. Bias between devices in a data set across all subject samples was assessed before and after recalibration. Results: Bias between devices was reduced from −3.6% to +0.6% in one data set and from +11.0% to +0.3% in the other by recalibration based on higher-order method. Using higher-order materials, bias was also reduced by recalibration, but mixed results were found: Bias was reduced from −3.1% to −0.1% in one data set and from −4.3% to −2.7% in the other. Conclusions: Recalibration did lead to a decrease in bias and thus can reduce the impact of the choice of comparator method. The procedure should be verified in a prospectively designed setting.