Chapter 28: United Kingdom Accreditation Service

Chapter 28: United Kingdom Accreditation Service

Role and function

¶28.1 The United Kingdom Accreditation Service (UKAS) assesses, against nationally and internationally accepted standards of accreditation, those organisations which provide conformity assessment services, such as certification, testing, inspection, calibration and verification.^20612061 See Ukas.com/about-us/about-ukas.

¶28.2 UKAS was established in 1995 as a company limited by guarantee.^20622062 Turner {UKAS0011242/4} page 4, paragraph 7(a). Throughout its existence it has operated under a memorandum of understanding with the government.^20632063 UKAS Memorandum of Understanding (MOU) 2006 {UKAS0011258}; UKAS Memorandum of Understanding (MOU) 2007 {UKAS0011259}; UKAS Memorandum of Understanding (MOU) 2009 {UKAS0011260}; UKAS Memorandum of Understanding (MOU) 2009 {UKAS0011261}; UKAS Memorandum of Understanding (MOU) 2013 {UKAS0011262}; UKAS Memorandum of Understanding (MOU) 2014 {UKAS0011263}; UKAS Memorandum of Understanding (MOU) 2017 {UKAS0011264}; UKAS Memorandum of Understanding (MOU) 2021 {UKAS0011251}.Show full note In 2009 it was appointed the national accreditation body for the purposes of article 4(1) of Regulation (EC) No 765/2008 under the Accreditation Regulations 2009.^20642064 Turner {UKAS0011242/8-9} pages 8-9, paragraph 7(g); UKAS was appointed pursuant to The Accreditation Regulations 2009 {UKAS0011250}. It was, and as at June 2017 remained, the only body exercising that function in the UK.^20652065 UKAS Closing Submissions for Module 6 {UKAS0011447/5} page 5, paragraph 17. It is a not- for-profit organisation^20662066 Turner {UKAS0011242/8-9} pages 8-9 paragraph 7(g). funded primarily by the fees it charges under contracts entered into on its standard terms^20672067 Turner {Day226/174:5-7}; {UKAS0011300}; {UKAS0011301}; {UKAS0011302}. with the organisations it accredits.^20682068 Turner {Day 226/147:5-17}. The MOUs allowed for sponsoring government departments to provide financial assistance in defined circumstances. Those circumstances were not relevant to any matters before this Inquiry. Our findings about its policies and operations are to a large extent based on the evidence of Lorraine Turner, its Accreditation Director.

¶28.3 Since the early 2000s, UKAS has accredited the BBA to certify products and systems under its Agrément certification schemes.^20692069 {BSI00001920}. As an accredited organisation the BBA issued a certificate for Arconic's Reynobond ACM panels.^20702070 {BBA00000047}. UKAS also accredited BRE. It had done so for several years before April 2002, when it first accredited BRE to perform large-scale fire tests in accordance with what became BS 8414.^20712071 {BSI00001726}; {BSI00001924}. As an accredited organisation BRE carried out BS 8414 tests on systems incorporating Kingspan K15 and Celotex RS5000.

Policies and procedures

¶28.4 UKAS accredited organisations to various national and international standards, in particular, a. ISO/IEC 17065, Conformity assessment – Requirements for bodies certifying products,

processes and services,^20722072 {BSI00001920}. which was the standard by which it assessed the BBA's Agrément certification schemes,^20732073 Albon {BBA00010723/7} page 7, paragraphs 19-20. and

202

¶b. ISO/IEC^20742074 ISO/IEC stands for International Organization for Standardization/International Electrotechnical Commission. 17025, General requirements for the competence of testing and calibration laboratories^20752075 {BSI00001726}; {BSI00001924}. which was the standard by which it assessed BRE's performance of BS 8414 tests.^20762076 Howard {BRE00005771/4} page 4, paragraph 18.

¶28.5 For both of those standards the process adopted by UKAS for accrediting an organisation involved checking both the technical requirements of each accredited activity as well as that organisation's processes and safeguards, such as its management systems, internal auditing, environmental conditions and the training of its personnel.^20772077 Turner {Day226/174:11}-{Day226/175:16}; {BSI00001920}; {BSI00001726}.

¶28.6 In addition, its internal policies provided for a rolling programme of annual assessment visits,^20782078 General Principles of Assessment of Conformity Assessment Bodies by the United Kingdom Accreditation Service (GEN-1) {UKAS0011246/13} figure 1; The Accreditation Process – Policy and Associated Requirements (ACC 1000) {UKAS0011247/22-23} section 6.8.2. examining each of the organisation's accredited activities at least once over a four-year period.^20792079 The Accreditation Process – Policy and Associated Requirements (ACC 1000) {UKAS0011247/32-39}; Turner {Day 226/181:1-10}. If it found that improvement was needed, by identifying a failure or "non-conformity" with a relevant standard, UKAS required the organisation to propose steps known as "improvement actions" to address the deficiencies. UKAS assessed whether those improvement actions met the problem and had been performed satisfactorily and then decided whether the body had met the accreditation standard. In addition, UKAS could make extra visits, including unannounced visits, outside the rolling programme and did make extra visits to some bodies as a result of the Grenfell Tower fire. However, it rarely made unannounced visits,^20802080 Turner {Day226/188:21}-{Day226/189:11}. so the organisation in question almost always knew when a visit was about to take place.

¶28.7 It was not possible for UKAS to review all the activities of an organisation being considered for accreditation. Each organisation subject to accreditation was therefore required to keep detailed management records, to audit its own performance and to record its more serious non-conformities. Accredited organisations were required to report any significant non-conformities to UKAS,^20812081 General Principles of Assessment of Conformity Assessment Bodies by the United Kingdom Accreditation Service (GEN-1) {UKAS0011246/18} section 4.5.2; Turner {UKAS0011422/11} page 11, paragraphs 8.1-9.1. but if they failed to do so, UKAS had access to the records when carrying out the next assessment and could see how the organisation had dealt with them.^20822082 Turner {UKAS0011422/12} page 12, paragraph 11.1. However, that meant that the effectiveness of its oversight depended to a significant extent on the honesty and integrity of the organisation in question^20832083 Turner {Day226/195:25}-{Day226/197:14}. and its ability and willingness to examine its own behaviour and respond comprehensively to rigorous self-evaluation as well as criticism by UKAS.

¶28.8 UKAS formed committees to advise on technical matters.^20842084 Turner {UKAS0011242/13} page 13, paragraph 10(b). Members were drawn from professional bodies, regulators, customers of accredited organisations and the government.^20852085 Turner {UKAS0011242/13} page 13, paragraph 10(c). The only committee within UKAS that considered matters relating to construction was the Construction Industry Technical Advisory Committee.^20862086 Turner {UKAS0011242/14} page 14, paragraph 11. However, for years before the fire that committee's work focussed on topics relating to infrastructure, for example, building roads and bridges.^20872087 Turner {UKAS0011242/14} page 14, paragraph 11(b). It rarely considered matters relating to fire testing or the fire safety of products used in construction and did not do so at all after 2000.^20882088 {UKAS0010843/4} item 4.5; {UKAS0010844/2} item 4.2; {UKAS0010845/2} item 3.2.

203

¶28.9 UKAS published accreditation schedules that listed all the activities that a particular organisation was accredited to perform. That was the principal way in which the construction industry could find out what functions a particular organisation was accredited to carry out at any time. If UKAS was satisfied that an accredited organisation was failing to meet the standards required for accreditation, it had the power to suspend or withdraw accreditation, either partially or fully.^20892089 General Principles of Assessment of Conformity Assessment Bodies by the United Kingdom Accreditation Service (GEN-1) {UKAS0011246/30} section 6; The Accreditation Process – Policy and Associated Requirements (ACC 1000) {UKAS0011247/24-25} section 7.2. If it had proof of fraudulent behaviour or an accredited organisation deliberately provided false information or deliberately violated accreditation rules, UKAS was obliged to initiate the process for withdrawal of accreditation.^20902090 The Accreditation Process – Policy and Associated Requirements (ACC 1000) {UKAS0010905/19} section 7.2.11; General Principles of Assessment of Conformity Assessment Bodies by the United Kingdom Accreditation Service (GEN-1) {UKAS0011246/30} section 6.5; Turner {UKAS0011422/14} page 14, paragraph 15.1. If UKAS suspended or withdrew an organisation's accreditation, it did not publish the reasons for doing so because it considered itself bound by the confidentiality undertaking in its standard terms.^20912091 Turner {Day226/206:21}-{Day226/207:9}. In cases of fraud or where there were serious concerns about safety UKAS would consider who needed to know about such matters. That might involve notifying the regulator or, if there were no regulator, might involve putting the information into the public domain.^20922092 Turner {Day226/210:7}-{Day226/211:10}. Since the fire UKAS has changed its standard terms to make clear that it is not bound by confidentiality in cases of fraud or a risk to the safety or health of individuals.^20932093 Turner {Day226/211:11-16}; UKAS closing submissions for Module 6 {UKAS0011447/13-14} pages 13-14, paragraphs 40-42. UKAS had no regulatory or enforcement powers.

Accreditation of the BBA

¶28.10 UKAS accredited the BBA to issue Agrément certificates for certain products. A schedule^20942094 Albon {BBA00010723/8-9} pages 8-9, paragraphs 23-29; {BBA00010722}; Randall {UKAS0011424/11} page 11, paragraphs 14.1-14.2. was agreed with the BBA, known as "BBA Document 001", setting out thirteen "Technical Specifications", each of which contained short descriptions of categories of products in respect of which UKAS agreed that the BBA could assess and, if appropriate, issue Agrément certificates. For example, Technical Specification 0012 was for "Wall and cladding products and systems", within which both cladding products and built-in cavity wall insulation products were briefly and broadly described.^20952095 {BBA00010722/6-7}.

¶28.11 The Technical Specifications listed core certificates for each product category, known as "Leader" certificates.^20962096 Haynes {BBA00010784/5} page 5, paragraphs 20-21; Randall {UKAS0011424/11-12} pages 11-12, paragraph 14.4. When a product falling within one of the categories came to be assessed, the BBA's project manager was required to follow the process that had been used to produce the Leader certificate. Those processes were not recorded in any separate document; instead, the project manager was expected to follow the steps taken to produce the Leader certificate, as set out in the contract between the BBA and the relevant manufacturer.^20972097 Randall {UKAS0011424/11} page 11, paragraph 14.2.

204

¶28.12 We have not seen any evidence that when it assessed the BBA's competence to issue Agrément certificates UKAS carried out a critical examination of the process by which individual Leader certificates had been produced. Instead, it appears that it considered whether the BBA had followed the steps set out in the contracts with its customers. However, that did not enable it to understand what the BBA was doing to follow those steps. The appendix to the contract made in 2006 between the BBA and Arconic for the certification of Reynobond, for example, consisted merely of short and rather uninformative summaries of the aspects to be considered, such as "behaviour under fire".^20982098 {BBA00008042/11}. It contained no description of the steps the project manager was expected to take to assess and record the product's properties. Accordingly, when it came to assess the work of the BBA, UKAS could not tell what the project manager was supposed to have done and therefore whether it had been done properly or at all. That led to a tendency on the part of UKAS to look only at the BBA's generic documentary processes rather than examine its technical assessment of performance for each type of product to see whether it was sound. In that respect there seems to have been a gap in its assessment of the BBA's Agrément certification scheme.

¶28.13 UKAS's policy relating to the assessment of conformity to ISO/IEC 17065 was to conduct a technical review of all schemes at least once every four years.^20992099 {UKAS0011247/35}. It regarded each category of products (for example, wall and cladding products and systems) as a separate scheme^21002100 Turner {Day227/81:16}-{Day227/82:1}. and accredited the BBA on that basis. Accordingly, although UKAS assessed the Agrément process as a whole, because it was similar for every scheme, it should have assessed the scheme for each of the thirteen Agrément product types separately every four years.

¶28.14 In practice, however, it failed to do so. The BBA's Agrément certification relating to "Wall and cladding products and systems" was assessed just once between 2008 and 2016, in 2009.^21012101 {UKAS0000586/5-6}; Randall {UKAS0011424/13} page 13, paragraph 17.2. UKAS thus failed to meet its own requirements for assessment of the BBA during that period. It is possible that UKAS treated an assessment of one scheme as being sufficient to accredit the BBA for all thirteen Technical Specifications, but there is no evidence that it did and if that was its reasoning, it was unsatisfactory, given the very significant differences between the products and systems covered by the Technical Specifications. Overall, as Ms Turner accepted, UKAS did not pay sufficient attention to the BBA's activities for Agrément certification.^21022102 Turner {Day227/93:10-17}.

¶28.15 For its part, the BBA did not comply with its obligation to keep UKAS informed of important developments as part of the continuing accreditation of its certification activities. Despite becoming aware in 2014 that there had been a very serious misstatement about fire performance in the certificate relating to K15,^21032103 The certificate provided that K15 could be used in accordance with Approved Document B paragraph 12.7, see Chapter 22; {BBA00000178}. the BBA neither reported it to UKAS as significant non-conforming work nor recorded it as such in its own records so that UKAS could discover it on the next assessment.^21042104 {UKAS0000977}; There was no mention of this significant non-conforming work in UKAS's subsequent management review: {UKAS0000977/4-5}. It should have been clear to the BBA that the information was relevant to its accreditation and needed to be shared with UKAS in accordance with its publicly available principles of assessment and standard terms of contract.^21052105 General Principles of Assessment of Conformity Assessment Bodies by the United Kingdom Accreditation Service (GEN-1) {UKAS0011246/18} section 4.5.2; {UKAS0011300/3-4} section 2.7; Turner {Day226/193:12}-{Day226/194:6}. We note that UKAS has now put the matter beyond any doubt by amending its standard terms.^21062106 UKAS Closing Submissions for Module 6 {UKAS0011447/12-15} paragraphs 39-46; UKAS Phase 2 statement {UKAS0011448/8} paragraphs 34-36.

205

¶28.16 The steps taken by UKAS after the Grenfell Tower fire to review its assessment of the BBA shed significant light on the efficacy of its previous practices. On 12 July 2017, UKAS made an extra visit to the BBA to consider the Reynobond certificate.^21072107 Randall {UKAS0011424/14-15} pages 14-15, paragraph 18.1. That visit prompted a further visit^21082108 {UKAS0001077/4}. on 15 August 2017 which examined the Reynobond certificate in greater depth and considered the BBA's work in reissuing several certificates relating to ACM material after the fire.^21092109 {UKAS0001083}.

¶28.17 At around the same time, UKAS received correspondence from a person who alleged that the Reynobond certificate was not up to date, that the BBA lacked competence, that it had failed to heed previous cladding fires and that it did not attach sufficient importance to fire performance generally.^21102110 {UKAS0001382}. The UKAS assessors who attended the BBA had known of those allegations before the visits in July and August 2017.^21112111 Randall {UKAS0011424/15} page 15, paragraph 18.2.

¶28.18 When they visited the BBA in July 2017, the UKAS assessors identified only two respects in which remedial action in relation to the Reynobond certificate was considered mandatory: changes to statements concerning the use of the material on buildings over 18 metres in height (including the removal of the picture of a tall building on the front of the certificate) and a change in the name of the holder from "Alcoa" to "Arconic".^21122112 {UKAS0001077/4-8}; {UKAS0001077/7}; {UKAS0001077/8}. The BBA had already identified those errors and at the time of the visit had made arrangements to correct them in the forthcoming reissue of the certificate.

¶28.19 The assessors also examined the BBA's conformity with ISO/IEC 17065, not only at the date of the visit but also over the period since the certificate relating to Reynobond had been first issued in January 2008. However, despite declaring in their report that the statements made in the certificate were supported by appropriate test data,^21132113 {UKAS0001077/4}. they failed to notice that the fire performance certificates provided by Arconic at the time of the initial assessment related only to the product in rivetted form, not in cassette form. They also failed to notice that the description of the product blurred the distinction between the form in which it left the factory (flat sheets) and the form it would take in use, which required that it be cut to shape and drilled for rivets or fabricated into cassettes. Those were significant failings, because in these particular assessments UKAS's specific role was to consider whether statements made by the BBA about the performance of the product were complete and accurate. The assessors noted that Arconic had failed to respond to the BBA's requests for information when the certificate was reviewed between 2013 and 2015, even though the BBA's own processes required responses,^21142114 Randall {UKAS0011424/20} page 20, paragraph 28.1; {UKAS0001083/6}. but failed to take any further steps to understand whether there were any problems or risks with continuing certification. Ms Turner admitted that Arconic's failure to respond should have prompted UKAS to examine more fully the merits and robustness of the BBA's decision to continue to certify Reynobond.^21152115 Turner {Day227/117:11}-{Day227/118:1-5}; {Day227/128:13-20}; {Day227/131:18}-{Day227/132:2}.

206

¶28.20 The file that UKAS reviewed in the course of those assessments contained correspondence from which UKAS could have seen that the BBA was prepared to negotiate with Arconic about the wording of the certificate. For instance, it contained correspondence in November 2007 in which Claude Wehrle of Arconic negotiated the additions to the draft BBA certificate for Reynobond with Hamo Gregorian.^21162116 {MET00055859}; {BBA00008042/515-517}. It should have been plain to UKAS that, if manufacturers were allowed to influence statements made in certificates, the quality of certificates might be undermined. We consider that UKAS should have identified the willingness of the BBA to discuss the wording of its certificates with manufacturers and at least have questioned whether it affected the independence, quality and accuracy of the certificates it issued.

¶28.21 UKAS also considered the technical competence of Valentina Amoroso^21172117 {UKAS0001077/8}. and Prayer Nkomo, both of whom had been involved in the reviews of the Reynobond certificate. However, it did not record any assessment of the competence of Hamo Gregorian and Brian Haynes, who had worked on the original Reynobond certificate.^21182118 Randall {UKAS0011424/18} page 18, paragraph 23.1. That was a significant oversight, because one purpose of the visits was to assess the validity of the certificate for the whole of the period from its initial issue in 2008. That required a more thorough investigation into whether those involved in the initial certification had the necessary skills and had properly assessed the product's performance at the outset.

¶28.22 The Lead Assessor for UKAS on the visits to the BBA in both July and August 2017 was Cary Randall, who had previously been employed by the BBA on Agrément certification until 2012.^21192119 Randall {UKAS0011424/3} page 3, paragraph 5.1 Indeed, his own work on the BBA's certification of a cladding product had been assessed by UKAS in 2009, the only assessment of the "Wall and cladding products and systems" scheme it had carried out.^21202120 {UKAS0000586/5-6}. Mr Randall had also worked for a time for Geoffrey Gurney, who had provided technical approval for the Reynobond certificate.^21212121 {UKAS0001077/3-5}; Albon {BBA00010723/6} page 6. His appointment created a potential conflict of interest, as UKAS was aware. However, it decided that it was not inconsistent with its policy for Mr Randall to assess the BBA because five years had passed since his employment by it and its policy required only a two-year interval.^21222122 Turner {UKAS0011242/18} page 18, paragraph 12(a).

¶28.23 Although the potential for bias was to some extent mitigated by the fact that Mr Randall was assisted by individuals who had no connection with the BBA,^21232123 Mr Randall was assisted by Technical Assessor Dan Patterson (for the July review) and observed by Sam Giles (for both): Randall {UKAS0011424/15} page 15, paragraphs 19.1-19.2; {UKAS0001077}; {UKAS0001083}. we consider that the decision to use him was unwise. He was not merely assessing the BBA's current performance; his task included the assessment of previous work, some of which had been done while he was still an employee of the BBA. Although Mr Randall had no involvement in the BBA's original certification of Reynobond or in its reviews of the certificate, his involvement in similar work for the BBA and his employment by it during the period under consideration would have made it difficult for him to examine its work critically.

207

¶28.24 The assessments of the BBA carried out by UKAS in 2017 resulted in almost no significant criticism. In our view UKAS should have examined the history of the BBA's work on the Reynobond certificate more critically and, given that the assessment specifically considered the entire history of the certificate from 2008 onwards, should have considered whether the BBA's work was sufficiently robust for the whole of the period from that date until the Grenfell Tower fire. That was all the more so in the light of the fact that the fire had involved a product that had been certified by the BBA and that it had received highly critical allegations about the BBA's past performance. Having noticed that the certificate had failed to state that the product was not suitable for use on buildings over 18 metres in height, the BBA should have been required as part of the assessment process to ascertain why the certificate had been wrong and why the error had been allowed to persist for so many years. UKAS should have made critical findings and should have considered suspending the BBA's accreditation until it had done so.^21242124 Turner {Day227/133:9}-{Day227/134:10}. The fact that UKAS failed to scrutinise the BBA's processes in that way and does not appear to have considered imposing any sanction of that kind indicates to us that its assessment of the BBA was too lenient.

Accreditation of BRE

¶28.25 UKAS accredited BRE to perform BS 8414 tests in accordance with ISO/IEC 17025. The tests of systems incorporating Kingspan K15 and Celotex RS5000 were carried out under that accreditation. It also accredited BRE under ISO/IEC 17025 to perform many other activities and each assessment generated lengthy and detailed reports covering a variety of subjects.^21252125 Turner {Day226/176:11-14}. Any weaknesses identified by UKAS in BRE's systems must therefore be understood in the context of the full report.^21262126 Turner {Day227/19:24}-{Day227/20:10}; {Day227/21:25}-{Day227/21:1-7}. However, there were some weaknesses that should have been of particular concern to UKAS but which it failed to identify and its failure to do so indicates defects in its assessment process.

¶28.26 To meet the requirements of ISO/IEC 17025, BRE was required to conduct internal audits which assessed its own performance and recorded failures and corrective action.^21272127 {BSI00001726/19-20} section 4.14.1-4. That was particularly important, because assessments by UKAS could only involve a sample of BRE's work.

¶28.27 Between 2012 and 2016, assessors repeatedly raised concerns about the quality of BRE's auditing. In 2012, it made several criticisms of BRE processes, principally, that its audit procedures were not recorded, that there was no comprehensive management of audits for the entire year, that the few audits of testing that had been carried out had not been reported to the Quality Manager and that no corrective action had been recorded.^21282128 {UKAS0004239/4}; {UKAS0004239/11}. In its report of 24 July 2012 UKAS told BRE that it would not renew its accreditation unless it was able to demonstrate at a further visit that management control was being exercised and that the management system was being implemented.^21292129 {UKAS0004239/6}. At the next visit BRE was able to satisfy UKAS about those matters^21302130 {UKAS0004253}. and in 2013 it demonstrated some improvement,^21312131 {UKAS0004269/8}. but in its assessments in 2014 and 2015 UKAS recorded delays in carrying out and responding to audits.^21322132 {UKAS0004315/3}; {UKAS0004315/11}; {UKAS0004339/3}. In 2015, in particular, UKAS reported that BRE had failed to record as non-conforming work certain deficiencies that had been noted in audits.^21332133 {UKAS0004339/3}; {UKAS0004339/7}. UKAS assessors noted that the BRE had not investigated the scale and effect of that non-conforming work, so it was not possible to understand whether it was necessary to take immediate action, for example, by halting work.^21342134 {UKAS0004339/7}; {UKAS0004353}.

208

¶28.28 It is clear from the assessments made by UKAS that throughout the period BRE's internal auditing processes were inadequate and UKAS should have been aware of that.^21352135 Turner {Day227/20:11}-{Day227/21:15}. Although it is now not possible to determine whether there was a single root cause of that underperformance,^21362136 Turner {Day227/21:15-23}. UKAS should have identified the problem and required BRE to investigate it.

Assessment of BS 8414 testing

¶28.29 In the four years between 2012 and 2016, UKAS witnessed only one test in 2016 as part of its assessment of BRE's competence to conduct BS 8414 tests^21372137 UKAS Assessment Report 2012 {UKAS0004239}; UKAS Assessment Report 2013 {UKAS0004269}; UKAS Assessment Report 2014 {UKAS0004315}, where BS 8414 was assessed by checks on equipment and discussion only (see {UKAS0004315/28-29}); UKAS Assessment Report 2016 {UKAS0004364}. and there are no records of its having witnessed any such tests during the period between 2008 and 2011. UKAS planned to witness BS 8414 testing every year from 2011 to 2016^21382138 Visit Plan 2011 {UKAS0004200/5}; Visit Plan 2012 {UKAS0004230/3} which planned to assess all reaction to fire tests; Visit Plan 2013 {UKAS0004265/12}; Visit Plan 2014 {UKAS0004313/7}; Visit Plan 2015 {UKAS0004329/3}. and although we accept that there might have been difficulties in arranging to observe a full-scale test, nonetheless UKAS failed to meet its own policy^21392139 {UKAS0011247/39}. of witnessing a live test once every four-years.^21402140 Turner {Day227/8:8-9}.

¶28.30 In its assessment in 2016, UKAS recorded the following findings and requirements for improvement in relation to BS 8414 testing: a. that BRE did not maintain a formal record of the procedure for carrying out BS 8414

tests, but relied on informal documents that did not form part of the quality assurance system;^21412141 {UKAS0004364/21}.

¶b. that the test area and control rooms in the burn hall used for BS 8414 tests were untidy and disorganised, with equipment and test materials not being stored appropriately.^21422142 {UKAS0004364/21}. BRE was therefore not meeting the ISO/IEC 17025 standard in those respects and improvements were required if BRE was to maintain accreditation.

¶28.31 In failing to witness BS 8414 tests for many years UKAS was depriving itself of the ability to monitor BRE's performance effectively and was unable to say if BRE was conducting the tests accurately and consistently.

209

Subsequent assessments of BRE

¶28.32 As in the case of the BBA, certain events that occurred after the fire show that longstanding defects in the assessments by UKAS of BRE continued in the years after the Grenfell Tower fire and proved stubbornly resistant to correction.

¶28.33 In September 2017, Sam Giles, Head of Construction and Physics in the Operations Department, reported to his seniors on the three bodies accredited by UKAS that had been connected with the refurbishment of Grenfell Tower, the BBA, BRE and Exova.^21432143 {UKAS0010859}; {UKAS0010860}. The report described the additional visits and other investigations that had been conducted by UKAS by that time. In the case of BRE, they included reviewing the records of the BS 8414 tests that DCLG had commissioned in July 2017, including the tests on systems which reproduced the external wall of the tower.^21442144 {UKAS0004456}. Although its review was based only on the records, UKAS identified a number of deficiencies, including inadequate staff training records, the absence of any formal contract between DCLG and BRE and inconsistencies in the recording of technical findings.^21452145 {UKAS0004456/3-4}. Mr Giles suggested that in the light of what it had discovered UKAS needed to take a tougher approach to BRE, particularly with regard to implementing required improvements.^21462146 {UKAS0010860/2}.

¶28.34 However, it appears that there was no change in the existing state of affairs. On 23 October 2020, Kingspan wrote to BRE to tell it that it had withdrawn several BS 8414 test reports on K15 because the material used in the tests had not been representative of the product it was selling.^21472147 {KIN00024104}. Despite the fact that the letter was reported in the specialist construction press shortly afterwards,^21482148 On 5 November 2020, Inside Housing magazine published an article about this letter: https://www.insidehousing. co.uk/news/kingspan-withdraws-insulation-fire-test-admitting-it-is-not-representative-of-product-on-market-for- 15-years-68461. UKAS became aware of it only on 17 March 2021 when Sam Giles noticed the article and circulated it within the organisation.^21492149 {UKAS0011430}.

¶28.35 UKAS immediately arranged an extraordinary assessment of BRE, which was conducted in May 2021 by John Leadbeater, Senior Assessment Manager and Technical Focus Person for fire testing,^21502150 Leadbeater {UKAS0011426/4-5} pages 4-5, paragraphs 3.2-3.3. with responsibility for leading on fire testing.^21512151 Turner {Day227/33:6-13}; Leadbeater {UKAS0011426/5} page 5, paragraph 3.3. He found no evidence that the tests had been performed incorrectly, but he did find evidence of weakness in complying with documentary requirements and internal audit procedures.^21522152 {UKAS0004773}. He recommended that if similar non-compliances were noted on the next routine assessment, there might be grounds for partial or full suspension of the BRE's accreditation for those activities.^21532153 {UKAS0004773/7}.

¶28.36 BRE objected to various aspects of the report^21542154 {UKAS0004774}. and as a result UKAS withdrew it and replaced it with a less critical version that made it clear that the findings had no bearing on the current performance of BRE.^21552155 {UKAS0004778}; {UKAS0011435}; Leadbeater {UKAS0011426/20} page 20, paragraph 18.1. That was done without consulting Mr Leadbeater, who was not made aware of the decision to withdraw the original report.^21562156 Mr Leadbeater's annotated comments: {UKAS0011435/3}; Turner {Day227/56:13}-{Day227/57:11}.

210

¶28.37 In the end UKAS decided to avoid any potential conflict by withdrawing the revised report and leaving all potential criticisms until the next assessment of BRE, which was imminent.^21572157 Leadbeater {UKAS0011426/20} page 20, paragraph 18.1; Turner {Day277/57:24}-{Day227/58:8}. It was conducted by Mr Leadbeater over several days in July and August 2021, as a result of which he made numerous adverse findings and identified a number of areas for improvement.^21582158 {UKAS0011327/35-45}; {UKAS0011314} rows 107-120.

¶28.38 One consequence of withdrawing Mr Leadbeater's first report was that his observations about the historical deficiencies of BRE were subsumed in an assessment of how BRE was performing in July 2021.^21592159 {UKAS0011327}. Similarly, after that assessment discussions about BRE's performance turned on how UKAS could prevent a recurrence of the failures, a matter that UKAS proposed to make the subject of requirements for improvement.^21602160 {UKAS0011331}. Although an organisation's current competence is, rightly, the main question for UKAS, a process that fails to take account of historic performance risks overlooking significant trends. An organisation could, and on the evidence before us did, propose measures to deal with problems as though they were of an isolated nature when they were symptomatic of more general shortcomings. BRE should have conducted a searching analysis of both the root cause of the deficiencies and their effects, and if it did not, UKAS should have required it to do so.

Conclusions

¶28.39 UKAS accepted in its closing statement that its dealings with the BBA and BRE showed that its systems were in need of improvement.^21612161 UKAS Closing Submissions for Module 6 {UKAS0011447/18} page 18, paragraph 56. It also told us that it had introduced a number of changes to its policies and procedures that were intended to rectify the defects that had been identified.^21622162 UKAS Closing Submissions for Module 6 {UKAS0011447/8-17} pages 8-17, paragraphs 25-53. In October 2022, we received further documents from UKAS indicating that it was continuing to make efforts to address weaknesses.^21632163 {UKAS0011454}; {UKAS0011455}; {UKAS0011456}; {UKAS0011457}. We take account of those matters when we come to make our recommendations. Nonetheless, the evidence of the way in which UKAS handled the accreditation of BRE and the BBA in the period before the Grenfell Tower fire has led us to the following conclusions: a. UKAS did not follow its own policies in relation to the frequency of assessing activities.

Its assessment of the BRE's performance of BS 8414 tests between 2008 and 2016 was inadequate, with just one test being witnessed in 2016.^21642164 Turner {Day 227/7:5-18}. It also failed to carry out a proper assessment of the BBA's Agrément schemes, assessing the certification of only a few categories of products sporadically between 2008 and 2015.

¶b. It failed to identify a consistent failure on the part of BRE to audit its work adequately in the years 2012-2016. UKAS should have required BRE to examine the underlying cause or causes of those repeated auditing failures. Similarly, when other failures were identified, it failed to require the BBA or BRE to complete a comprehensive investigation into their cause and possible wider effects.

¶c. When assessing the BBA after the fire it failed to identify important errors in the issuing and review of the Reynobond certificate, which resulted from a failure to examine the information available in the assessments with sufficient rigour. It was also

211

unwise of UKAS, after the Grenfell Tower fire, to allow a former employee to assess the BBA's Agrément Scheme.

The relationship between UKAS and accredited bodies

¶28.40 Although we accept that UKAS took its role as national accreditation body seriously, the reliability of its assessments depended on the honesty and co-operation of the organisations it was accrediting. It was therefore left to the organisations themselves to identify weaknesses, to report them to UKAS and to implement improvements in good faith. UKAS left a great deal to trust. Although a degree of trust was essential, UKAS should have taken a more searching, even sceptical, attitude to the bodies it accredited.

¶28.41 Sam Giles identified that as a potential problem in 2017 after the Grenfell Tower fire. After an assessment of the BRE, the BBA and Exova (conducted in July and August 2017), Mr Giles concluded in September 2017 that some organisations might be displaying a degree of arrogance in dismissing findings made by UKAS.^21652165 {UKAS0010859}; {UKAS0010860/3}. However, despite Mr Giles' observations, his comments did not appear to prompt a reconsideration of the relationships between UKAS and some of its accredited bodies, at least at that time. For example, UKAS assessors might have been instructed to adopt a less trusting view of the organisations subjected to routine assessment and to push them to consider in greater detail the cause and effect of any critical findings that UKAS made.

¶28.42 UKAS was sometimes too willing to accept the competence of an organisation without having reviewed all the relevant documents^21662166 {Day227/144:25}-{Day227/145:5}. and sometimes too quick to minimise or ignore its deficiencies or to accept superficial improvements as adequate. That was demonstrated by events both before and after the fire. The effect of failing to require more thorough investigations into the causes and effects of shortcomings was that opportunities to improve the work of organisations engaged in testing and certifying products were lost.

¶28.43 UKAS's powers in the face of recalcitrant or defective bodies are surprisingly limited. The most it can do in the face of unsatisfactory conduct on the part of an accredited organisation is to suspend or withdraw accreditation. It has no legal powers or role in enforcement. Moreover, the fact that accreditation is voluntary presents a particular challenge, in that, subject to market pressures, a certification body can simply decline to be accredited by UKAS at all.

¶28.44 That seems to us to be a thoroughly unsatisfactory state of affairs. A voluntary system of accreditation, where the accrediting body has no power of enforcement other than suspension or withdrawal of accreditation, in the end simply leaves the designer and building owner having to trust the certificate or test report and classification. It deprives them of the confidence that the relevant body has itself adopted rigorous methods which are wholly free from the commercial pressures inherent in relationships with manufacturers.

213