ALACC FAQ
Updated June 15, 2007
Question 1:
The batch size is, defined in the document for quantitative methods as "usually 20 samples or less", is that a maximum number that the lab must meet or alternatively can the size of the batch be determined as appropriate for the type of testing activity the laboratory is conducting?
ALACC 5.9 states:
For quantitative methods, accuracy and precision should be demonstrated to be acceptable; at a minimum, the laboratory shall run a quality control sample (QCS) with each batch of samples (usually of 20 samples or less). If a QCS is not available the lab must choose an alternate control sample such as a sample duplicate or reference material with each batch of samples (usually of 20 samples or less).
For qualitative tests controls, such as QCS's, should be included to demonstrate that the test worked. The suitability of the controls used should be justified by the lab.
Answer:
The ALACC guide in section 5.9 does not intend to define the size of a batch. The intent of this statement was to explain that in a batch, of whatever number of samples, Quality Control Samples (QCS) shall be included at a frequency defined, documented and justified by the lab. For many test methods, the QCS is included every 20 samples.
The lab must determine for itself the maximum number of samples that can be analyzed in one batch, or run contiguously between QCSs. The lab must document and justify its decision.
Labs use different terminology to define the number of samples that are either analyzed together in a group, such as samples on an ELISA plate or are run contiguously, such as liquid samples presented to an instrument. Some labs use the term batches, other labs use the term lots. Any term is acceptable; but, the lab must define the term unambiguously.
Question 2:
For qualitative tests, is it imperative that "controls such as the QCS...be included" or can it be left up to the lab to decide what is appropriate for a given test and whether controls are necessary?
Answer: Controls are necessary for qualitative tests. It is up to the lab to decide what is appropriate for a given test. This decision must be documented and justified. If no Quality Control Samples exist, the lab must do the best it can do.
Section 5.10.8 "Format of Rand Certificates" should state 5.10.8 "Format of Reports and Certificates."
4.13.2.1 Previously the section for Technical Records (formerly 4.12.2) stated that "All records shall be traceable to the analysis performed on a particular sample...". The new revision states that "the chain of custody for all records shall be unbroken from the time a laboratory takes responsibility for the sample until the result is reported to the customer..." The question thus concerns the intent of this clause and whether this is meant to be referring to all records or just to those records that pertain to the sample?
This section, 4.13.2 deals with Technical Records, and the ISO 17025 wording is "The records for each test or calibration shall contain sufficient information to facilitate, if possible, identification of factors affecting the uncertainty and to enable the test or calibration to be repeated under conditions as close as possible to the original.", so 4.13.2 deals with the records of the test. Section 4.13.1 deals with General Records. Therefore, 4.13.2.1 is meant to refer to those records that pertain to the sample.
Also in Section 5.9: For determining the accuracy and precision associated with quantitative methods, paragraph two states that a QCS must be run if one is available or that an alternate control sample (such as a sample duplicate or Reference Material) if one is not. Going on, paragraph four states that replicate tests shall be used if suitable Reference Material is available, and the sixth paragraph states that statistical process control charts shall be prepared for the QCS sample analyses and for the duplicate analyses if duplicates are used. Does this mean that both a QCS and a duplicate sample or a Reference Material replicate be run with each batch of samples or is a single QCS or RM acceptable?
If possible both a QCS and a duplicate sample should be run with a batch of samples because the QCS does not capture the variability unique to the sample(s). If either a QCS or sample duplicates are not available, the lab must explain this, demonstrate they tried their best to obtain/develop a suitable QCS or use sample duplicates and explain what they are doing instead.
Table 1 Since temperature holding equipment, in particular incubators often differ greatly in volume, should there be a point below which, where the volume of the unit is sufficiently small that it becomes irrelevant to require that they undergo temperature uniformity and stability studies? Similarly, for most uses of freezers, should there also be a lower limit at which their volume renders the determination of uniformity of temperature irrelevant?"
The last note in Table 1 addresses this point.
Due to size, use or design, some refrigerators and freezers may not need to be temperature and stability calibrated annually and may not need the demonstration of temperature uniformity and stability. In these cases the laboratory should justify the reason for an alternate frequency of calibration or for not demonstrating temperature uniformity and stability"
Yes, due to size, a point below which, where the volume of the unit is sufficiently small, the lab does not need to demonstrate temperature uniformity and stability. It is up to the lab to determine this size and justify not performing the temperature uniformity and stability studies.
ISO 17025 Section 5.2.1, first paragraph, second sentence reads, "When using staff who are undergoing training, appropriate supervision shall be provided."
and ALACC states
"Personnel shall have demonstrated their competency prior to working on customer samples."
Can actual samples be used in training or when demonstrating competency?
In the ALACC statement, the phrase "working on" means the analyst performs the analysis of customer samples independently with the intention of reporting the results. Ideally, during the training and demonstration of
competency no customer samples would be used; however, in some cases customer samples may be used. Whenever a trainee is performing tasks on customer samples, the trainee must do so under the supervision of a trained analyst and it is the trained analyst who takes responsibility for these
tasks. The lab must be able to defend to an auditor that the trainee performing the tasks does not impact the analysis.
Examples of cases where a trainee may perform tasks on customer samples
Simple tasks such as pureeing plant material such as lettuce; the task has a distinct endpoint that the trained analyst can observe and ensure the trainee has performed correctly.
Entering sample sets into a computer, the trained analyst can verify the information is correct and entered correctly before the analysis takes place.
|
