A scheme for the evaluation of methods in clinical chemistry with particular application to those measuring enzyme activities. Part II: analysis of data and performance assessment.

Recommendations are made concerning the editing of day-to-day reproducibility data for establishment of precision in the evaluation of a clinical chemistry method. The concept of diagnostic and equalized diagnostic indices is introduced together with formulae for their generation from the usual precision and accuracy data acquired in method evaluations. These indices allow direct comparison of data obtained from procedures for measuring enzyme activities which employ a variety of experimental conditions and units in their protocols. The equalized diagnostic index permits assessment of the suitability of the normal range assignment. Permissible limits of variation (PLV) and permissible limits of discrepancy (PLD) have been developed empirically from detailed examination of data from method evaluations and proficiency testing surveys in the published literature. The application of the diagnostic indices and the two permissible limits of criteria have been illustrated using data from the assessment of 19 kits measuring CPK activity. The inconsistency of the correlation coefficient in method comparisons is confirmed.