OMB Winter Mtg.-February 5-6, 2015

170

Draft 6. Note: This document is a draft suggestion only and is not an approved document. Please do not distribute to laboratory staff for use. Laboratory On-going Competency: Laboratories shall participate in at least one PT event annually for each test, type of test/method, and/or technique on the scope of accreditation. Competency by Alternative Means of Evaluation: For those tests/methods and/or techniques which are not covered by relevant and available external PT schemes, the laboratory shall demonstrate competency by an alternative means of evaluation, justifying its actions. Alternative actions, listed in suggested order of preference, are:  Participate in a round robin: interlaboratory test performed independently several times  Testing of a blinded well characterized LCS  Performance of an ILC with other accredited testing labs  Performance of a comparison with another test method/technology  Address the elements identified in ISO/IEC 17025 Section 5.9 including the evaluation criteria of the data 5.9.2 Quality Control Sample Acceptability The laboratory shall have procedures that define the acceptability of quality control samples. A laboratory’s SPC data shall be reviewed on a regular basis that ensures it is reporting reliable data. The laboratory shall define what constitutes a trend in the SPC data and investigate trends where necessary. Corrective action shall be initiated when controls do not meet the established acceptability criteria. When using multi-analyte test methods the probability exists that at least one analyte in each run will be out of limits; however, that does not necessarily mean there is an error and assignable cause for the result. The laboratory’s criteria should be balanced: avoiding unproductive corrective actions for statistically random events, yet not so broad as to ignore correctable, non-random errors. The laboratory’s criteria may take into consideration the fact that for some multi-analyte methods, some analytes behave better than others (i.e. exhibiting less variance and/or higher mean recovery) and that an analyte’s variance may increase as the concentration of the analyte decreases. There are several approaches that could be taken to set limits that take into account the probability of an out of control result in multi-analyte methods.  Assign the analytes to groups that have similar analytical characteristics or chemical structure. An example of this is low-molecular-weight ketones that tend to be lost in the sample preparation process or organic acids that may be poor performers in certain extraction processes. Then the quality control and control charting procedure can be designed to track a representative analyte of each of these groups.  Set limits that take into the account the probability that an out of control result will be encountered, such as the approach set forth for environmental testing laboratories in the 2003 National Environmental Laboratory Accreditation Conference (NELAC) Standard (Appendix D, section D.1.1.2.1.e). Proficiency Test Sample Acceptability The laboratory shall evaluate PT results when they are received. Most external PT providers issue acceptability limits and criteria; if issued, the laboratory shall use the PT provider’s criteria to evaluate the results. If the PT provider does not issue acceptability criteria or the laboratory is performing proficiency testing by the alternative means described above, then they shall have procedures that define the acceptability of the results. The following information provides guidance to assist the laboratory in evaluating the results. The assigned value can be established by one of the following four options:  Consensus of the majority of the laboratories, providing that a sufficient number of laboratories participate in a particular round (set). When using a consensus value, outliers may be removed for statistical analysis after an error analysis has been shown to be inconclusive  The SPC ranges for blinded LCSs  Fortification value of prepared samples  Assigned result from a previous round (set)