CHANGES IN ISO/IEC 17025:2017 – SAMPLING UNCERTAINTY?

« Back to Previous Page
17
0

Am I wrong or ISO/IEC DIS 17025:2016 proposes in clause 7.6.2 that laboratories performing sampling MUST evaluate “sampling uncertainty”? I...

Please to read the entire article.

Marked as spam
Posted by Alexis Oramas Pérez (Discussions: 1, Comments: 18)
Replied on March 14, 2017 12:00 am
115 views
0
Private comment
Please ignore this post. It is a replay of the one above, posted by mistake. I don´t see any option to delete it. Sorry. Debate is developing in the post above.
Marked as spam
Posted by Alexis Oramas Pérez (Discussions: 1, Comments: 18)
Replied on March 14, 2017 8:00 pm
0
Private comment
1. I don't think that it is a discrimination. It is about responsibility and intended scope of the test result. If a laboratory tests an incoming sample, they are only responsible for the results of "that" sample. If a laboratory is responsible for the result of the whole population (for example the pollution level of a lake), of course they have to think about the effects of sampling.
2,3. If we are talking about uncertainty, we can not expect that assessors and laboratories can be wholly prepared for it. There is no true way of estimating uncertainties, we have only some accepted conventions. Like ordinary GUF or MCMC. It is about awareness and level of understanding.
Those methods which are available today will be a good beginning point for most of us after this clause is effective. As we start to think about sampling uncertainty and its statistical nature and discuss it together then we will develop better methods in time.
Marked as spam
Posted by Abbas Dundar (Discussions: 0, Comments: 4)
Replied on March 14, 2017 8:00 pm
0
Private comment
(cont.)
There is no need to worry about the audits. Good assessors always knew what to ask/demand and when. (Bad auditors can be bad on any of the clauses.)
4. Sorry I do not know anything about this.
Marked as spam
Posted by Abbas Dundar (Discussions: 0, Comments: 4)
Replied on March 14, 2017 8:00 pm
0
Private comment
Dear Abbas:

Imagine two different first-party labs, in different factories, each of which carries out sampling in the respective production line of its factory and then analyses those samples, being both accredited for the two activities. Now, according to the DIS, they will have to evaluate sampling uncertainty, as well as that for the actual measurement.

Then, one of those two labs has the “clever” idea of making an arrangement with the production manager so, from now on, the production department is to do the sampling and then take the samples to the lab to be analysed. So, this lab can claim it doesn´t carry out sampling, thus it isn´t compelled to evaluate sampling uncertainty and simply continues evaluating measurement uncertainty.

Meanwhile, the other lab, which continues to carry out its sampling, is now compelled to evaluate them both. Do you get the picture? Isn´t this unfair?

“Isn´t this some kind of discrimination?”

Hope you get now my point.

Best regards,
Alexis

Marked as spam
Posted by Alexis Oramas Pérez (Discussions: 1, Comments: 18)
Replied on March 15, 2017 8:00 pm
0
Private comment
Alexis if the other lab party did the sampling and allegedly they claimed and reported it that they did not do the sampling then my question is, who did the sampling? This is now a question of traceability. All samples should be accompanied by a Chain of Custody signed by the person who did the sampling and the people received the samples for analysis or during transient so that it could be verified and track down.

The laboratory should have 2-types of calculation templates for reporting sampling uncertainties. One is for laboratory and 2nd is for the client or independent/subcontractor. The quality of sampling may vary between own laboratory sampling technique and the contractor involve. To be able to control the Quality Assurance of the sampling the laboratory should have guidelines on how the sample to be collected so that variation of the sampling technique and uncertainties could easily be identified and limit to an acceptable value. Been doing this type of reporting since 2011
Marked as spam
Posted by Ricardo Peralta (Discussions: 0, Comments: 9)
Replied on March 18, 2017 8:00 pm
0
Private comment
I think everyone needs to back up a step. In general "sampling uncertainty" is important in, for instance, wet chemistry, where the homogeneity of a batch and/or the homogeneity of samples drawn from a batch can have profound effects on the analytical result.

In contrast, if you are performing lot sampling of, for instance, electronic power supplies that are being produced on an assembly-line basis, the "sampling uncertainty" can essentially be ignored- if and only if -the samples are obtained by following a statistically valid sampling plan - which can be anything from a MIL-STD-105 / ISO 2859 -type plane, an FDA Binomial singe or two stage plan, etc.

Discriminatory? Not at all. The issue is whether or not a sampling uncertainty is RELEVANT. That is why the draft Standards says "... A laboratory performing SAMPLING or testing activities SHALL EVALUATE MEASUREMENT UNCERTAINTY...". Some Labs do and some don't - depending on what kind of testing they do, and why, and for whom.
Marked as spam
Posted by 56cdf1806453f0.12856495@limsforum.com (Discussions: 0, Comments: 11)
Replied on March 18, 2017 8:00 pm
0
Private comment
Further to my last post.
As for the example of the two Labs given by Alexis Oramas Pérez, the distinction is crucial - and is not in any way discriminatory. The question is who is responsible and for what are they responsible. Note that, in the production of both Information Technology Equipment and in the production of vials of Insulin, a manufacturer will often send production samples to an independent third party lab for testing and assessment of results.

If the independent Lab is responsible for the design and implementation of sampling plan, then Sampling Uncertainty (if relevant) may indeed be the responsibility of the Independent Test Lab. (Or not, as per the contract for the work). If, on the other hand, the manufacturer simply sends samples for testing and analysis, the Independent Test Lab is NOT ever responsible for the computation of "Sampling Uncertainty".
Marked as spam
Posted by 56cdf1806453f0.12856495@limsforum.com (Discussions: 0, Comments: 11)
Replied on March 18, 2017 8:00 pm
0
Private comment
Further to my previous post. I am very sympathetic to concerns that the methods for estimating "uncertainty" - regardless of whether we are concerned with sampling uncertainty or measurement uncertainty or compliance uncertainty.

That said, as a practical matter, test labs can and should use whatever "cookbook" models (modified as appropriate to the particular Labs' exact test processes and equipment) are available in Normative Standards or published Guides. I would not spend too much time worrying about the fact that these models are imperfect (or worse than imperfect). You can only do what is practical to do given the state-of-knowledge available when you do your work.
Marked as spam
Posted by 56cdf1806453f0.12856495@limsforum.com (Discussions: 0, Comments: 11)
Replied on March 18, 2017 8:00 pm
0
Private comment
How can one get access to the DIS please?
Marked as spam
Posted by Paul Hayes (Discussions: 0, Comments: 2)
Replied on March 18, 2017 8:00 pm
0
Private comment
The essential condition of sampling is "the sample must be a true representative of the bulk.If not true, nearest to true. This leaves a room for uncertainty. The clause 7.6.2 is in this connection.Besides, the environmental conditions, the holding time, transportation, packaging , all add to uncertainty.The uncertainty , thus, in sampling needs serious consideration
Marked as spam
Posted by Mahesh Gupta (Discussions: 0, Comments: 6)
Replied on March 19, 2017 8:00 pm
0
Private comment
Thoughts following on above discussion. if the lab pass on the responsibility for sampling to the production line, since the scope of the lab is to test for release of result related to manufacturing process, aren't the lab still responsible to ensure the quality of the work made by whom they 'subcontracted' it to?
as already mentioned, it comes back to relevance.
Marked as spam
Posted by Anna Annerberg (Discussions: 1, Comments: 3)
Replied on March 19, 2017 8:00 pm
0
Private comment
1. The labs not carry sampling activities but are involve in testing keep in view that sample has significant impact on test results.
The lab should develop and implement the sample acceptance or rejection policy and should critically examine the condition of sample. In case sample is not in good condition the lab should request for re sampling to the client or should clearly mention the sample condition on sample receiving invoice and also mention on test report as well.

2. True the accreditation bodies should also communicate the labs about the competencies of their assessors and technical expert; I observed that some time the technical expert may not have grip over Measurement of Uncertainties, Method validation, Method Verification, etc.. but they are working from long time with accreditation bodies.
Lab should inquire about assessors competencies before the assessment; as the primary goal is to improve the quality of testing not only assessment.

Marked as spam
Posted by Muhammad Farooq (Discussions: 0, Comments: 9)
Replied on March 19, 2017 8:00 pm
0
Private comment
3. Method for calculation depends on the nature of test; which factors should be included while a lab is going to calculate uncertainty. GUM guide is also good. Review it
4. details should be given about changes but its depends upon the international bodies.

You are not mistaken; you have highlighted the fact truly.
Marked as spam
Posted by Muhammad Farooq (Discussions: 0, Comments: 9)
Replied on March 19, 2017 8:00 pm
0
Private comment
Evening group,does any one have a copy of the ISO17025:2017 (previous Measurement traceability) but now metrology traceability.I will appreciate very much.
Marked as spam
Posted by Morne Oerson (Discussions: 0, Comments: 15)
Replied on March 19, 2017 8:00 pm
0
Private comment
My email morne1532@gmail.com
Marked as spam
Posted by Morne Oerson (Discussions: 0, Comments: 15)
Replied on March 19, 2017 8:00 pm
0
Private comment
Isn't that document subject to copy write legislation?
Marked as spam
Posted by Eugene Organ (Discussions: 0, Comments: 89)
Replied on March 20, 2017 8:00 pm
0
Private comment
I think that as long as the laboratory is the contractor to the production and not subcontracted the production for the sampling,
we can see the production as an customer and there of we have no party in the sampling.
Our responsibility in this case is to ensure that the product is in order and fulfill the requirements
in the contract between the laboratory and the customer.
Marked as spam
Posted by Magnus. Skattberg (Discussions: 0, Comments: 1)
Replied on March 20, 2017 8:00 pm
0
Private comment
I think it may be safest to see what the FDIS says and then see how accreditation bodies interpret it. This wording you pick out in the DIS looks like a carry-over from all other references to testing labs being changed to 'sampling and testing...'. So it may or may not remain as currently worded. Even if it does, however, it does not necessarily represent a change. 17025:2005 sec 5.4.6.3 requires that "all uncertainty components which are of importance in the given situation shall be taken into account". Eurachem's view would be that if the intent is to estimate a bulk property after sub-sampling, the sampling uncertainty contributes and any uncertainties associated with sub-sampling are expected to be considered (though by 'considered' I include reducing to negligibility via homogenisation) . Hence the existing 5.4.6 already implies a need to consider sampling if the lab is doing any. That is not causing much trouble. So what will matter is how accreditors interpret the requirement.
Marked as spam
Posted by Stephen Ellison (Discussions: 1, Comments: 5)
Replied on March 21, 2017 8:00 pm
0
Private comment
As Stephen stated, if there is a significant uncertainty due to sampling process, it should be included in this ALL (5.4.36). In this way we can not say SU is a new uncertainty component. What makes 2017 different is the special emphasis on the sampling uncertainty. That might be coming from assessors audit experiences. Especially if there is a wide ignorance between the laboratories on SU.
Marked as spam
Posted by Abbas Dundar (Discussions: 0, Comments: 4)
Replied on March 22, 2017 8:00 pm
0
Private comment
Sampling uncertainty can be estimated by using replicate (samples) method. Similar to the bulk PT sample prepared and subsamples distributed. Homogeneity/ inhomogeneity data / study can give SD of sampling

ANOVA can be used to determine homogeneity of sub samples
IUPAC also gives a strong formula and simple statistics to work out between samples Standard Deviation
Marked as spam
Posted by mushtaq ahmad (Discussions: 1, Comments: 16)
Replied on March 23, 2017 8:00 pm
« Back to Previous Page