Statistical procedures for determination and verification of minimum reporting levels for drinking water methods

Area of Science:

• Environmental Science
• Analytical Chemistry
• Laboratory Science

Background:

• Accurate quantitation is crucial for environmental monitoring and regulatory compliance.
• Existing methods may lack precision and accuracy at low concentrations.
• The U.S. EPA's Office of Ground Water and Drinking Water sought a robust quantitation procedure.

Purpose of the Study:

• To introduce and define the lowest concentration minimum reporting level (LCMRL) quantitation procedure.
• To detail the methodology for determining the LCMRL, ensuring high confidence in analytical results.
• To present a simpler verification procedure for laboratory performance at a given minimum reporting level (MRL).

Main Methods:

• The LCMRL procedure involves processing multiple replicates across a concentration range.
• Data are plotted (measured vs. true concentration) and analyzed using ordinary or variance-weighted least-squares regression.
• Ninety-nine percent confidence prediction intervals are calculated, intersecting with 50%-150% recovery data quality objectives to determine the LCMRL.

Main Results:

• The LCMRL is defined as the highest concentration where future recovery is predicted between 50% and 150% with 99% confidence.
• The procedure accounts for both precision and accuracy in quantitation.
• A verification method using seven samples confirms laboratory performance at the MRL if the prediction interval falls within recovery limits.

Conclusions:

• The LCMRL procedure provides a statistically sound method for establishing the lowest reliable reporting level in analytical chemistry.
• The method is flexible, allowing adjustments to data quality objectives and confidence intervals.
• The LCMRL determination is primarily for method development, with a separate verification procedure for ongoing quality assurance.