OPTIMAL FREQUENCY OF QUALITY-CONTROL ANALYSES IN ENVIRONMENTAL MEASUREMENT

A variety of quality control (QC) analyses are performed with regular frequency in environmental chemical measurements. Typically, an accept able result must be obtained for each QC check before measurement of e nvironmental samples may begin, and each check is periodically repeate d to validate the analyses of intervening samples. Corrective actions are performed if the check is failed, and the intervening batch of sam ples are remeasured. Under such a scheme, excessive costs will occur i f the quality control checks are too frequent (i.e., batch size is too small) or infrequent (i.e., batch size is too large). This paper desc ribes an approach for investigating the frequency of QC checks for min imizing the expected cost of measuring a set of samples. Two specific expected cost models are discussed. It is shown that the empirical est imate of measurement drift can be used along with computer simulation to estimate the batch size that will minimize the expected cost of sam ple analyses. Measurement drift of inductively coupled plasma mass spe ctrometry for Sb-121 is demonstrated with data from standard reference materials repeatedly analyzed between QC solutions. Examples are pres ented using computer simulation and empirical estimates of linear drif t functions for analytical measurements.