Clinical decision limits as criteria for setting analytical performance specifications for laboratory tests

Clinical laboratory testing involves inherent biological, preanalytical, and analytical variations that influence test result interpretation. Established knowledge includes the use of biological variation to guide analytical performance specifications. However, a gap remains in how to incorporate clinically significant differences into these specifications. Prior research has shown that diagnostic variation can be used to estimate acceptable performance criteria. Yet, no prior work had resolved how to integrate statistical power and repeated measurements into these specifications. This uncertainty drove the need for a new approach to derive performance specifications that align with clinical decision-making. The relevance of this work lies in improving the accuracy of test result interpretation in clinical settings. Understanding how variations contribute to diagnostic outcomes is essential for optimizing laboratory performance. This paper addresses the need to refine analytical performance standards using clinical decision limits.

The study aimed to derive clinically acceptable analytical performance specifications (CAAPS) for laboratory tests using clinical decision limits. The specific problem addressed is how to incorporate biological, preanalytical, and analytical variations into performance criteria. The motivation stems from the need to ensure that laboratory results are clinically meaningful and interpretable. The researchers propose a method that integrates diagnostic variation with statistical power and repeated measurements. This approach allows for a more accurate estimation of acceptable analytical performance. The study focuses on six specific measurands to demonstrate the feasibility of the method. The goal is to provide a framework that aligns analytical performance with clinical decision-making. This work contributes to the field by offering a novel tool for setting performance specifications based on clinical significance.

The study used biological, preanalytical, and analytical variation coefficients (CV_I, CV_PRE, CV_A) to calculate diagnostic variation (CV_D). The reference change concept was applied to determine clinically significant differences (CD) between measurements. A formula was used: CD = z * √2 * CV_D. Clinically significant differences for six measurands were obtained from international guidelines. CAAPS were calculated by subtracting CV_I and CV_PRE variances from CV_D. Modified formulae were introduced to account for statistical power (1-β) and repeated measurements. The calculations were performed for urine albumin, plasma sodium, pancreatic amylase, and plasma creatinine. The method also considered the need for replicate measurements in blood HbA_1c and low-density lipoprotein cholesterol.

The derived CAAPS for urine albumin was 44.9%, for plasma sodium 0.6%, for pancreatic amylase 22.9%, and for plasma creatinine 8.0%. These values were calculated using z = 3, α = 2.5%, and 1-β = 85%. For HbA_1c and low-density lipoprotein cholesterol, replicate measurements were necessary to meet CAAPS for patient monitoring. The CAAPS were compared with analytical performance specifications based on biological variation. The results suggest that the CAAPS model provides a more clinically relevant framework. The method accounts for statistical power and repeated measurements, which is a novel approach. The study demonstrates that CAAPS can be tailored to specific measurands. The findings highlight the importance of integrating clinical decision limits into performance specifications.

The authors propose that CAAPS models offer a new tool for assessing analytical performance specifications in clinical laboratories. The usability of these models depends on the relevance of clinically significant differences and the feasibility of repeated measurements. The study shows that CAAPS can be derived for specific measurands using diagnostic variation and statistical power. The findings suggest that replicate measurements are necessary for certain tests like HbA_1c. The authors suggest that integrating clinical decision limits into performance specifications improves their relevance. The study does not propose new drug targets or future research directions. The conclusions are based on the derived CAAPS values and their comparison with existing specifications. The authors emphasize the importance of aligning analytical performance with clinical outcomes.