Analytical Performance Evaluation and Comparability of Patient Results Within a Point-of-Care Blood Gas Network

Background: Before an analyzer is replaced by another in a clinical laboratory, the analytical performance of the new device and a comparison with a reference or an already established method must be performed before use on patient samples. In point of care, a large numbers of instruments are used, and it is necessary that all are compared against a laboratory instrument. According to these considerations, the aim of this study was to evaluate 23 blood gas analyzers. Methods: Three types of evaluation were conducted. These included a comparability of RapidLab 1265 (current method) and ABL90 (candidate replacement method), an analytical performance evaluation of ABL90 (precision and bias), and a comparability of the same ABL90 device with the other 22 ABL90. Results: The performance of the candidate method is equivalent to that of the current method. In the analytical performance evaluation of ABL90, estimated bias was higher than allowable bias for only 2 parameters: ionized calcium at 0.8 mmol/L and glucose at 14.49 mmol/L. However, both were less than total allowable error. All the ABL90 devices studied were equivalent to the ABL90 that was compared previously with RapidLab 1265. Just 1 device showed an estimated difference greater than allowable. It was replaced by another analyzer that was also evaluated showing satisfactory results. Conclusions: All analyzers showed results within the acceptance criteria. Studies were carried out in the laboratory before blood gas analyzer installation, before use for patient results. This made it possible to detect technical problems earlier, without altering clinical practice.