The how and why of an accuracy base for proficiency testing programs

The National Bureau of Standards (NBS) has been involved for more than ten years in development of definitive methods for measuring important inorganic and organic analytes in human serum, including calcium, potassium, chloride, cholesterol, glucose, urea, uric acid, and creatinine. These methods are carefully designed and thoroughly tested isotope dilution mass spectrometric procedures that produce results of high precision and accuracy. Definitive methods are used to certify concentrations in reference serum pools, which are distributed among clinical laboratories as a means of assessing the accuracy of the routine methods. The College of American Pathologists operates a major proficiency testing program for clinical laboratories. The National Bureau of Standards works with the College of American Pathologists to assign values for analyte concentrations in proficiency testing samples where possible, and these values are used to assess the overall accuracy of the participants (grand mean) as well as the accuracy of peer groups (those that use the same methodology and/or instrumentation). For the analytes studied to date, the grand mean values have generally been close to the definitive values, but some of the peer group results have shown significant biases. To illustrate the definitive method development process, a modified and improved definitive method for cholesterol measurement is described. The modified method uses a cholesterol-13C3 as the labeled internal standard, capillary gas chromatography for sample introduction, and a novel means of electrically switching between the ions being measured. Results on standard reference material 909, a lyophilized human serum, demonstrate the precision and absence of significant measurement bias attainable with the new method.