The Different Types of Analytical Limits used in CCLAS
For every analytical technique, there are often some practical range limits for the results that might be generated as a result of the test. Where there are some limits to be applied to the results generated for the test, these limits need to be defined so that they can be assessed.
These limits are internal to the analysis processes and to give aid to the analyst or operator. There are different limits that are applied for the reporting of results to the client.
Analytical limits can be of several different types.
Analysis Detection Limits
These limits, sometimes referred to as the practical quantitative limit (PQL), are the extreme result limits that cover the complete analytical range of the particular scheme and analyte. These upper and lower limits can be individually flagged as critical, in which case if they are exceeded, other actions must be undertaken before the sample's result can achieve an analysed state. These detection limits are sometimes dependent upon the sample matrix and, as such, can be changed on the sample scheme analyte.
Refer to Assessing Results against Detection Limits.
Lower Detection Limit and Criticality
Where a sample scheme analyte has a Lower Template Sample Code assigned and its result is below a critical lower detection limit, then the result is flagged for checking. In this case, a workbook operator can trigger the registration of tests from that lower template sample to cater for analysis by a method more suitable to a lower range of result.
Upper Detection Limit and Criticality
Where a sample scheme analyte has an Upper Template Sample Code assigned and its result is above a critical lower detection limit, then the result is flagged for checking. In this case, a workbook operator can trigger the registration of tests from that upper template sample to cater for analysis by a method more suitable to a higher range of result.
Alternatively, if the scheme version analyte has a single Upper Scheme and Upper Analyte defined for upper analysis, and that scheme analyte is registered on the sample and has a Workflow Status of Not Analysed or No Result, then the single sample scheme analyte can be activated.
The limitation to using an Upper Scheme and Upper Analyte instead of a Lower Template Sample Code and Upper Template Sample Code is that only one scheme analyte is provided for the check analysis.
Practical Analytical Limits
Scheme limits are generally used to denote the normal ranges of results and can be defined with warning and failure limits. Results that are outside the acceptable range are flagged. The operator might change their processes and re-analyse the sample, or they might need to acknowledge these failures, so that the results can be validated and hence reported to the client.
Scheme limits can also be used to define acceptance boundaries where true detection limits are not used or, similarly, may be used to define acceptance boundaries where detection limits may be set but are defined as non-critical.
In the case where scheme limits are not set, some analytes on the sample are required to be assessed as surrogate analytes. Surrogate analytes are often a precisely added component, introduced into the sample mixture prior to analysis. This added component would respond to the analytical technique and experience sample matrix effects in the same manner as the naturally occurring analytes or compounds, but the surrogate component would not typically be found in the natural sample. The result of the surrogate analyte determination gives a measure of how well the analytical technique can recover that analyte or compound during the analysis phases, and should be indicative of recovery losses in the naturally occurring analytes.
Surrogate analytes might also be a measure of instrument performance. The surrogate, in this case, would typically be one well suited for analysis with the measuring instrument and then the precise knowledge of the expected result for that surrogate presents an invaluable tool for performance assessment.
Surrogate analytes might be a purely calculated result or an aggregated result based on recovery of other analytes, this giving an overall assessable measure of the performance of the test method.
Surrogate analytes can have limits assigned to them in two ways. Either the laboratory can define their ranges with the scheme version analyte's limit specifications. This gives the most powerful assessment as it has both warning and failing limits. An alternative way is to define the surrogate limits directly within the scheme version analyte properties as an acceptable recovery range, expressed as a percentage.
Surrogate limits can be set as a target value from which a surrogate recovery (as a percentage) is calculated. In addition, the result limit range can be defined.
Refer to Assessing Results against Scheme or Surrogate Limits.
Analytical Accuracy Limits
Analytical techniques are designed to measure precisely and accurately the parameters of interest. There are many opportunities for mistakes to be made, so some quality control checking mechanisms are required. The insertion and analysis of blank and standard QC samples along with the client’s samples are one of these QC mechanisms. If these QC samples give the ‘right’ result for the test, then there is more confidence that the client’s samples are also reporting the right result.
blank QC samples comprise materials for which the result for a particular test is either zero or very low. These types of samples are used to check contaminations and drifts at low values. Sometime the blank QC samples are measurements on the base reagent mixtures that are used in the test (to check for base contaminations), and might not be a literal ‘material’ at all, an empty crucible, for example. Generally a number of blank QC types are created and used to identify where contamination is being introduced. These may include method blanks, equipment blanks, field blanks, trip blanks, and instrument blanks. Information from blanks can be used to enhance the accuracy of other standards when placed in the same batch and in certain cases, blank data is used to correct QC and real samples prior to assessment.
standard QC samples are materials where the results for a particular test is already known (and certified), so that if the laboratory gets the same result, then errors are managed. Standards are used to check for bias and gross errors.
spike QC samples are similar to standard QC samples in that they have known results, but are actually a mixture of a client’s sample and a standard QC material, to check impacts on the sample matrix with respect to the ability to get the right result (as standard QC materials may not match the sample matrix exactly). A precisely known quantity of spike or standard QC material is added to a selected sample at some point prior to analysis. The timing of this addition affects the interpretation, as mentioned, for the case of blank assessment. The use of spike QC samples is primarily to identify sample matrix and transport effects, however, as with standard QC samples, the overall intent is to provide an assessment of data accuracy.
Some laboratories use standard QC materials for their spike QC samples as well, so there is an option that, once a standard QC material is defined, it can also be used as a spike QC material as well. Other laboratories keep these materials and definitions independent.
A standard qc material is comprised of lots that generally represent different batches of material (which need to be tracked independently as they might have slightly different limits) or might represent different aliquots/containers of the material which might need to be tracked and identified. Different containers of the same material can become contaminated or changed individually, so it is important to known which container of the material was used.
QC samples can also be used as calibration samples whose results assist with the calibration of the test method.
Note: QC samples can be used for multiple schemes.
QC materials, especially standard and spike QC materials can be manufactured or purchased and can be quite expensive (and their analysis is not typically included in the specific charges to a client), so the laboratory should not analyse more QC samples than are necessary to fulfil the quality control requirements of the laboratory and the client. Likewise, if less material can be analysed per test (as long as it does not compromise the tests), then this reduces the costs and extends the life of the QC samples. For this reason, the typical weights/volumes used for QC samples is configured independently to those weight/volumes for the client's samples.
Note: QC materials (like materials supplied by the client) might be hazardous so QC blanks, standards and spike materials can be associated with hazards for reporting purposes.
Refer to Assessing the Accuracy of Analysis.
Analytical Precision Limits
In most cases, analytical results should be reproducible. Precision limits for reproducibility is definable per analyte. The results for duplicate and replicate QC samples are checked against the defined limits. These precision limits are the normal ranges of the results and can be defined with warning and failure limits. Duplicate or replicate QC sample results that are outside the acceptable range are flagged. The operator might change their processes and re-analyse the sample, or they might need to acknowledge these failures, so that the results can be validated and hence reported to the client.
Refer to Assessing the Precision of Analysis.
Product Limits
Customers submit samples for analysis where the samples are of a specific material type, that is, of a known a product that may have client-defined or internationally recognised specifications. Often the client is seeking a compliance statement of their product to a particular international standard. The compliance certificate is issued by the testing laboratory after a series of tests are completed.
Often a single product material might have quite different specifications based on the country, intended use, or seasonal changes.
Within CCLAS, a product is a known material type that is associated with one or more specifications. Each specification has a set of limits defined by scheme, version, analyte and unit.
At sample registration, a product code can be assigned to a sample, depending on how the client has identified the sample. The product's primary and other specifications are then attached to the sample. The primary specification is used in QC assessment, and the primary and other specifications on the sample are assessed for reporting.
Although the responsibility of a product failing specifications is the responsibility of the client, it is important for the laboratory to be internally notified of any potential primary specification failure before results are reported, so that all internal checks can be performed, as clearly the client is not expecting the specifications to fail. These internal check procedures are dependent upon laboratory and product, and are not covered in this process documentation. If at the end of these internal checks, there are no changes to the results, then they are reported as they are, and do not reflect on the testing laboratory. But if the internal checks are not performed, and the client receives a failing specification on their report and then asks the laboratory to double-check their results upon which the results are amended, then this does reflect on the laboratory's performance.
If samples have more than one specification attached, then as results are captured, all results are assessed against all attached specifications.
When a report request is created, the reporting specification, and its assessment, can be selected from the available specifications.
Refer to Assessing the Product Quality of Samples.
