BMTA offers the following comments on the draft UKAS document ‘TPS 72 UKAS Policy on Decision Rules-Pass/Fail Criteria taking Uncertainty of Measurement into Account’. The comments have been collated following a meeting organised by BMTA on 10 July 2019 and are the views of the 25 people representing 17 testing laboratories and 3 calibration laboratories that were present. There was particular concern about the requirement that measurement uncertainties must be included in all laboratory test reports whether the customer wanted them or not.
- Para. 4.4: This paragraph requires that where decisions rules for deciding conformity with a standard encompass expressions of uncertainty of measurement proposed by a laboratory’s customer are not considered appropriate i.e. do not comply with ISO17025:2017 the laboratory will need to discuss this with their customer to ensure that the outcome is fit for purpose. The agreed decision rule employed for the customer must be clearly detailed in the contract review and in the report or certificate following testing/calibration. The paragraph then states that if a simplified report is requested, then the decision rule applied will still need to be fully recorded in the laboratory documentation for the testing/calibration that has been performed. The laboratory shall not agree on any decision rule that will have a detrimental impact on the integrity and outcome of the test or specification.
- The BMTA members pointed out that many of their customers were suppliers of various consumer products including textiles. The suppliers were contracted by retailers who were in effect second party specifiers. The numbers of individual items to be tested could be very large. The retailers (second party specifiers) set the specifications for the product concerned and were only interested in a simple pass/fail result. In the majority of instances the retailers did not understand the concept of uncertainty of measurement, of possible false acceptances or possible false rejections, and in many cases simply did not care. The suppliers who were the test laboratory’s customers were under commercial pressure to supply only those products which met the retailer’s specification (i.e. passed) and in many cases had no desire to be informed about measurement uncertainty as it was irrelevant to them. Any acknowledgement of PFAs or PFRs would simply complicate life for them. If a laboratory suggested that PFAs or PFRs were retested their customers might see this as the laboratory creating work for itself and increasing the cost of testing.
- Testing laboratories thus face a dilemma; if they comply with the requirements of para graph 4.4 of TPS 72 they risk losing considerable business from their supplier customers who
could move their business to overseas laboratories in countries whose NABs took a less rigorous approach to accreditation to ISO 17025:2017. This would result in a net loss of business for UK laboratories when they are already enduring the uncertainties of Brexit. Alternatively, if the laboratories accede to their supplier customers’ demands for simple pass/fail test reports they risk losing their UKAS accreditation. Most suppliers require their testing work to be undertaken by an accredited test laboratory on their ‘approved accredited laboratories’ lists. If a laboratory loses is accreditation it will lose the suppliers business. Either way the test laboratory loses.
- This might also result in significant reputational damage to UKAS who in extreme cases might be accused of putting British laboratories out of business.
- A number of BMTA members highlighted the difficulties of determining the uncertainty of measurement especially in testing textiles where because of the very large range of textile types involved, multiple tests on each type of textile would be required which would increase the cost of testing significantly. Again, there is a risk that the laboratories’ customers would simply see this as the labs trying to create business for themselves, would not pay for it and would take their business elsewhere.
- Microbiological testing laboratories considered that the concept of uncertainty of measurement was not relevant to their reports as quantification was either not possible, very difficult and subjective or not required. For example, customers requiring testing for legionella in hot water systems only needed to know whether it was present or not. The measurement uncertainty thus becomes more of a question of detection limits. A similar situation pertains with microbiological testing of food.
- A possible solution to the Laboratories’ dilemma might be afforded at the Contract Review stage. Laboratories should have a discussion with their customer about the measurement uncertainty of the test in question and this discussion should be fully documented and signed by both parties. The laboratory customer should then state clearly whether they wished to have a simple unconstrained pass/fail test report not withstanding the possibility of PFAs or PFRs. This should also be documented and signed by both parties. It should also be stated clearly on test reports that measurement uncertainty had not been taken into account as agreed with the customer. In this way the Laboratory customer (suppliers) would demonstrate that they had clearly discussed, understood and accepted the risks associated with PFAs and PFRs. BMTA members present considered that it was an immense task to educate their customers about measurement uncertainty and its implications and a yet greater task to educate retailers who were the Laboratories’ customers’ customers.
- Guard banding as mentioned in paragraph 5.2.2 might prove helpful in providing a simple test/fail report provided that it is agreed with the Laboratory’s customer.
- It is important to remember that it is the Laboratory that is being accredited not its customers. Laboratories’ businesses should not be put at risk on account of what their customers choose to do with their test reports. It is not possible for test laboratories to have an in depth knowledge of their customers’ businesses to the extent that they could advise their customers on the level of risk to the customer’s business arising from application of a particular decision rule.
- Members present thought that it would be more helpful if TPS 72 was not published but that a technical guidance document on the subject supported by a number of case studies be published instead. The test laboratory cannot advise specifically or insist on the decision rule to be applied by its customer but can explain the concept of uncertainty of measurement and can direct customers to relevant guidance.
The Consultation process
BMTA Members would have appreciated receiving an email from UKAS alerting them to the TPS Consultation before it started. Most labs do not look at the UKAS website every day and many of those present were unaware that the consultation had commenced. Furthermore, the date of closing of the consultation is not stated on the document neither is the address of to whom comments should be sent (apart from the general UKAS Office address).