Variance reduction in analytical processes in the clinical laboratory by digital filtering

Area of Science:

• Analytical Chemistry
• Clinical Laboratory Science
• Stochastic Processes

Background:

• Analytical chemical processes in clinical laboratories exhibit fluctuations.
• Process variations can be modeled as non-stationary stochastic processes with time-varying means.
• Between-run variance is often significantly greater than within-run variance in these systems.

Purpose of the Study:

• To develop and validate a model describing fluctuations in clinical laboratory analytical processes.
• To present a digital filtering procedure for reducing process variance.
• To improve the accuracy of serum sample analysis.

Main Methods:

• Postulated a model for process variations using non-stationary stochastic processes.
• Quantified short-term (within-run) and long-term (between-run) variances.
• Developed a digital filtering procedure to estimate and correct for time-varying means.
• Tested the filtering procedure on inorganic phosphate determination using a continuous-flow system.

Main Results:

• Between-run variance was significantly greater than within-run variance for four analytical systems measuring six serum analytes.
• The digital filtering procedure successfully estimated process means within each run.
• Significant reductions in variance were achieved using the proposed filtering method.
• The procedure demonstrated effectiveness in an experimental environment for inorganic phosphate analysis.

Conclusions:

• The proposed model accurately describes process variations in clinical laboratory analytical chemistry.
• The digital filtering procedure effectively reduces analytical variance by correcting for time-varying means.
• This approach offers a method for enhancing the reliability and accuracy of clinical laboratory testing.