Refining quality control strategies in highly automated laboratories: experience in the integration of multistage statistical designs and risk management.

Abstract

Introduction: The ISO 15189:2022 standard considers both the robustness of analytical methods and the risk of erroneous results in the quality control plan (QCP) design. Westgard et al.'s nomogram recommends quality control (QC) rules based on sample run size to ensure that the maximum expected number of unreliable patient results remains below one. This study aimed to implement a standardized, risk-based QC strategy across multiple analyzers without integrated on board QC, ensuring practical quality assurance.

Material and methods: Thirty-two biochemistry parameters on Alinity c systems and three on Cobas Pro systems were included. The analytical performance of each parameter on each analyzer was assessed using sigma metric. Workload requirements were considered to determine the desired run size. Based on the "sigma metric statistical QC run size nomogram" proposed by Westgard et al., a multistage bracketed QCP was designed for each parameter. When multiple designs were available, the simplest QC rule was prioritized.

Results: Seven QCPs were initially established for 35 parameters. In the absence of automation, practical adaptations based on sigma metrics were implemented. Additionally, to streamline management, the QCP covering the greatest number of parameters per analyzer was prioritized, which ultimately resulted in the adoption of only two general QCP. Only 4 individualized QCP were required to cover 10 parameters with lower sigma values.

Conclusions: This approach demonstrates the feasibility of implementing a refined QC strategy for parameters with sigma ≥ 4 in a highly automated laboratory, ensuring consistent quality assurance and efficient resource allocation for higher-risk tests.