Regression analysis in interlaboratory surveys: a case study with cholesterol and triglycerides

Area of Science:

• Clinical Chemistry
• Laboratory Medicine
• Biostatistics

Background:

• Interlaboratory variations in clinical chemistry analyses can impact patient care.
• Accurate laboratory results depend on reliable standardization and methodology.

Purpose of the Study:

• To introduce and validate a new survey design using regression analysis for assessing interlaboratory performance.
• To identify and quantify systematic errors in cholesterol and triglyceride measurements across laboratories.

Main Methods:

• A novel interlaboratory survey design employing regression analysis was developed.
• Fifty New Zealand laboratories participated, performing cholesterol and triglyceride analyses.
• Regression analysis was used to compare laboratory results against target values, identifying systematic errors (non-unity slope, non-zero intercept).

Main Results:

• Significant interlaboratory variation was observed for both cholesterol (CV 8-27%) and triglycerides (CV 13-113%).
• Systematic differences accounted for 30% of cholesterol and 40% of triglyceride variations.
• Regression analysis confirmed simple (standardization, blank correction) and complex (nonlinearity, nonspecificity) systematic errors.
• Graphical displays aided in diagnosing faults and detecting standardization differences.
• Enzymatic cholesterol methods showed lower precision compared to colorimetric methods.

Conclusions:

• The proposed regression analysis survey design is effective for detecting and characterizing interlaboratory variations and systematic errors.
• This approach facilitates improved laboratory quality control and standardization.
• Method comparison highlights potential precision issues with enzymatic cholesterol assays.