Issues in assessing analytical performance specifications in healthcare systems assembling multiple laboratories and...

Area of Science:

• Clinical Chemistry
• Medical Device Performance
• Healthcare Systems Analysis

Background:

• Analytical performance specifications (APS) are typically compared against the intermediate reproducibility uncertainty of a single in vitro diagnostic medical device (IVD MD).
• Healthcare systems often utilize multiple IVD MDs across various laboratories, leading to patient samples being analyzed by different devices, sometimes from different manufacturers.

Purpose of the Study:

• To evaluate the utility of reproducibility uncertainty within a healthcare system for minimizing bias and overall measurement uncertainty.
• To explore methods for quantifying and minimizing measurement uncertainty in multi-laboratory healthcare settings.

Main Methods:

• Utilizing root mean squares deviation (RMSD), calculated as sample standard deviation (SD) and relative SD, to express uncertainties including imprecision and bias.
• Employing variance component analysis (VCA) to quantify the relative contributions of factors influencing measurement uncertainty.
• Leveraging data from commutable control samples, split patient samples, and big-data techniques for bias and uncertainty monitoring.

Main Results:

• RMSD effectively captures both imprecision and bias, making it suitable for expressing uncertainties within a healthcare system.
• VCA provides a method to identify and quantify the most significant contributors to measurement uncertainty.
• Data from control and patient samples are crucial for monitoring bias and measurement uncertainties.

Conclusions:

• Assessing reproducibility uncertainty across a healthcare system is valuable for developing strategies to minimize bias and overall measurement uncertainty.
• Variance component analysis offers critical insights for reducing measurement uncertainty, ultimately benefiting patient care within healthcare systems.