OMB Winter Mtg.-February 5-6, 2015

97

commercial analyst concentrations, without concern for covering a wide range of analyte concentrations. So in practice, we find that PT data have a more limited range than would a designed CS data set. In the past, the purpose of proficiency tests was to verify the correct implementation of a method in a laboratory. As a part of laboratory accreditation, the laboratory must demonstrate proficiency vs. other laboratories or against traceable standards or reference materials. Similarly, Method Validation was a process used prior to implementation to assure the method could be used in more than one laboratory and be shown to have a reasonable level of laboratory reproducibility. ISO 5725 is the standard used to design validation experiments and to analyze the results of multi-lab validation studies. Table 1 is a summary of the traditional differences between the Validation process and Verification.

Method Development  Method Validation  Method Verification

Table 1 Validation

Verification

After Method Development

After Validation

Pertains to the Method

Pertains to the Laboratory

Validation can be

Generally requires multiple laboratories, but can be done as a single laboratory with a certified RM

Single Laboratory Multiple Laboratory

ISO 5725

ISO 13528

Validation Experiments/ Collaborative Studies Strict Control of Method

Verification Experiments/ Proficiency Tests (PT) Loose method control

Statistical Output Reproducibility Repeatability Trueness or Bias

Statistical Output: z-score for each lab