Alternative statistical approach to evaluating interlaboratory performance.

Area of Science:

• Clinical Chemistry
• Laboratory Medicine
• Quality Assurance

Background:

• Proficiency testing is crucial for assessing laboratory performance.
• Existing methods may not fully capture the nuances of laboratory errors.
• Establishing objective performance criteria is essential for quality improvement.

Purpose of the Study:

• To develop and validate a new technique for realistic assessment of laboratory performance.
• To establish "state-of-the-art" performance criteria for key blood gas parameters (pH, pCO2, pO2).
• To provide laboratories with tools for accurate self-evaluation and error identification.

Main Methods:

• Utilized interlaboratory results from 129 participants over 18 months.
• Established performance criteria based on established "state-of-the-art" benchmarks.
• Employed two statistical measurement techniques: cumulative percentile rank and mean error analysis (algebraic and absolute).

Main Results:

• The new technique allows for realistic assessment of laboratory performance against established criteria.
• Laboratories can evaluate performance based on state-of-the-art benchmarks, total error, or medical usefulness.
• The method aids in identifying the nature and probable sources of performance errors.

Conclusions:

• The developed technique offers a robust approach to laboratory performance assessment.
• It facilitates targeted quality improvement initiatives by pinpointing specific error types.
• The system can support regulatory compliance and professional quality enhancement when integrated with external standards.