Upper Tribunal (Administrative Appeals Chamber)

 

 

IN THE UPPER TRIBUNAL Appeal No.  GIA/5435/2014 

ADMINISTRATIVE APPEALS CHAMBER

 

Before: Upper Tribunal Judge K Markus QC

 

 

DECISION

The appeal is dismissed.

 

Representation:

Appellant: Ms Zoe Gannon (counsel)

First Respondent: Mr Adam Sowerbutts (solicitor), not in attendance

Second Respondent: Mr Robin Hopkins (counsel)

 

 

 

REASONS FOR DECISION

 

Factual summary

 

Background

1.    This appeal concerns a request for information relating to the “Safe and Sustainable Review of Paediatric Congenital Cardiac Surgery Services” (“the Review”).  At the time of the request the “Safe and Sustainable Review of Paediatric Congenital Cardiac Surgery Services” the Appellant was a councillor of Leeds City Council and Chair of the Joint Health Overview And Scrutiny Committee (“JHOSC”) for Yorkshire and Humber. 

2.    The Review had its origins in the public inquiry in the care of children receiving heat surgery at the Bristol Royal Infirmary between 1984 and 1985, chaired by Professor Sir Ian Kennedy.  The report of the inquiry made a number of recommendations for the improvement of paediatric cardiac services, including concentrating expertise in a smaller number of specialist units. In 2008 NHS England’s predecessor decided to carry out the Review and it was proposed to reduce the number of paediatric cardiology centres in England from eleven to seven. The Review was a major exercise carried out to determine which should be maintained and which should be closed.  The Review was conducted by the Joint Committee of Primary Care Trusts (“the JCPCT”) which made the final decision. 

3.    The Review was assisted by three specialist groups including an Independent Assessment Panel chaired by Professor Sir Ian Kennedy, also known as “the Kennedy Panel”.  The Panel comprised eight experts from various clinical backgrounds. Its role was to review the existing providers and evaluate their compliance with certain service standards, in essence as to their ability to provide a safe and sustainable service.  

4.    The process of the Review and the role of the Kennedy Panel within that was complex, lengthy and detailed. The members of the Panel were selected in March 2010.  At around the same time the eleven centres were invited to carry out their own assessments of their services against specific criteria.  The individual members of the Kennedy Panel scored the centres based on the self assessments.  The scores comprised 32 separate sub-scores, each on a five point scale against specific assessment criteria. The Panel then visited each centre as a result of which each individual Panel member modified their individual scores for that centre. One Panel member did not visit or score two centres because he had been unwell.  Another member did not visit or score another centre, because he had a connection with it.  It is those modified scores for each centre visited by the individual Panel members which are in issue in this appeal.

5.    After each visit to a centre, the whole Panel met. At each meeting the Panel collectively identified the salient qualitative evidence they had gathered, which was published online.  They also discussed the scores that the Panel should give that centre against the assessment criteria and this resulted in consensus subscores (ie for each criterion) and a total consensus score for each centre.

6.    The subscores were weighted by the Review’s project team. The final weighted scores, but not the unweighted consensus subscores nor the individual Panel members’ scores, were provided to the JCPCT. In December 2010 the Kennedy Panel produced a report which set out the scores and a detailed narrative assessment explaining the Panel’s view as to each centre’s compliance with the criteria.  The report contained declarations of interest for each centre by individual Panel members.

7.    There followed a process including assessment by the JCPCT on configuration options and public consultation. The JCPCT sought further advice from the Kennedy Panel following receipt of the consultation submissions.  That ended the Panel’s involvement in decision-making.  There continued a complex process of review and appraisal by a variety of bodies culminating in a business case and the final decision by the JCPCT.  The business case had considered fourteen options, each for a group of hospitals, assessed against four criteria and ranked on a scale of 1-4.  One of those criteria was “quality”. The Kennedy scores were fed into the assessment against that criterion only.  The highest scoring option was subject to detailed examination involving a wide range of factors and the consultation responses.  The reconfiguration options were finalised during November 2011 and June 2012.

 

The request for information and response

8.    On 29 October 2012 the Appellant sent a request to the Review’s Programme Director, asking for a wide range of information including:

“the individual scores prepared by each of the Kennedy Panel Assessors under each of the assessment criteria for each of the institutions that they assessed”

9.    The request was handled by NHS London, which held the Review documentation.  It provided a great deal of information in relation to other parts of the request.  In relation to the Kennedy Panel, it disclosed all of the requested individual scores. The identities of the panellists were already in the public domain.  The only information which was withheld was the attribution of the individual scores to the relevant individual members.

10. NHS London relied on section 40(2) and section 41 FOIA as the basis for withholding the links between the individual scores and the names of the Panel members. 

11. Following a reorganisation, on 1 April 2013 NHS England inherited responsibility

 

The Information Commissioner’s decision

12. By Decision Notice dated 4 June 2013 the Commissioner decided that NHS England was entitled to withhold the information pursuant to section 40(2) FOIA. 

13. I set out here the First-tier Tribunal’s summary of the Commissioner’s decision, because in its decision the First-tier Tribunal adopted the Commissioner’s reasoning:

“3. The Commissioner was satisfied that the disputed information was the personal data of third parties (DN §28). He therefore went on to consider whether disclosure of the disputed information would be in breach of DPP1 (DN §29).

4. The Commissioner first considered whether the disclosure of the disputed information would be fair for the purposes of DPP1 (DN §30).In doing so the Commissioner took into account the factors set out at DN §31, namely:

• the individual’s reasonable expectations of what would happen to their information;

• whether disclosure would cause any unnecessary or unjustified damage or distress to the individual concerned; and

• whether the legitimate interests of the public are sufficient to justify any negative impact to the rights and freedoms of the individuals concerned.

5. The evidence before the Commissioner was that the individual Panel members were not, in this context, public facing figures but were independent experts in their particular fields. The Panel itself had a ‘collective’ identity and it was only public facing through its chairman. All of the formal scoring undertaken by the Kennedy Panel and which was ultimately used by the JCPCT was understood by those members to be by consensus of the Panel, and not as members individually (DN §§ 32 -33).

6. The Board therefore maintained that individual Panel members never had any expectation that their individual scores would be published (DN § 34). The Board referred to the Terms of Reference for the Kennedy Panel in support of this contention (DN § 36). It also noted that even the JCPCT, as decision maker, did not itself receive the Panel members’ individual scores (DN §38).

7. In light of this evidence, the Commissioner was satisfied that it was likely that the Panel members would have had a reasonable expectation that the disputed information would not be disclosed under FOIA (DN §39).

8. The evidence before the Commissioner also suggested that disclosure could, in this case, lead to incorrect aspersions, such as allegations of bias, being cast about particular individual named Panel members (DN § 40).

9. Exceptionally, in the circumstances surrounding this case, the Commissioner accepted that disclosure of individual Panel members’ scores linked to individual names could lead to a skewed interpretation or a selective use of data. He accepted that, unusually in this case, there was some merit to the argument that the Board would not be able to provide satisfactory context to aid correct understanding of the disputed information (DN § 41).

10. The Board produced evidence to show the nature of personal attacks which had already occurred publicly in respect of other individuals involved in the Review (DN §§ 42 – 45). It argued that the Panel members may suffer similar attacks if the disputed information was disclosed.

11. In favour of disclosure of the disputed information, the Appellant argued that there was a legitimate interest in the public knowing how each Panel member scored each hospital as he considered that at least one of the Panel members may have had a bias towards a particular centre (DN §46).

12. The Appellant also referred to the possibility of statistical analysis of the scores. He did not explain specifically why personal data, as opposed to the anonymised data set previously released to him, was necessary for such an analysis.

13. The Commissioner accepted that there is a general public interest in terms of the transparency and accountability of public sector organisations and specifically in accessing information about the way a public authority has reached decisions. However, the Commissioner did not consider that any legitimate interest extended to disclosure of the individual Panel member’s names linked to the individual scores they gave (DN §48). Consequently, the Commissioner was unable to conclude that disclosure of the disputed information was necessary to meet such a legitimate public interest (DN §49).

14. The Commissioner therefore concluded that section 40(2) FOIA was engaged by the disputed information and that it would be unfair to the data subjects for their personal data to be disclosed (DN § 50).

15. As the Commissioner determined that it would be unfair to disclose the disputed information,it was not necessary for him to go on to consider whether disclosure was lawful or whether one of the conditions in Schedule 2 of the DPA was met (DN §51). Likewise, it was not necessary for him to go on to consider the application of section 38 or 41 FOIA (DN §52).”

 

The appeal to the First-tier Tribunal  

14. The Appellant did not dispute that the information was personal data.  His case before the First-tier Tribunal was essentially that the scores prepared by the Panel members were not inherently private and that the Panel members were performing a public role, so that a low level of protection was appropriate. He said that it was not reasonable for Panel members to expect that their identities would be kept secret. There was a significant public interest in disclosure, which was to allow the public to assure themselves that there was no risk of the scores being tainted by bias of individual Panel members.

15. The First-tier Tribunal was provided with considerable documentary evidence and witness statements, and heard oral evidence from the Appellant, from witnesses on behalf of NHS England and from a member of the Kennedy Panel (Ms Hildebrand).  The evidence and the submissions of each party were summarised in some detail by the First-tier Tribunal.

16. By a majority the First-tier Tribunal dismissed the appeal. Their reasons were as follows:

“2. The majority view agrees with the Commissioner's Decision and reasoning as set out therein and as set out above in his response to the Grounds of Appeal. We adopt this reasoning.

3. We find that the Kennedy Panel members were not public facing. We accept that they were individual experts in their fields and were engaged in that capacity and as such not in a public facing role. The evidence form Ms. Hildbrand in this regard was compelling and was not challenged, contradicted or rebutted in any way.

4. We accept that disclosure of the disputed information could and probably would put at least some of the Kennedy Panel members at risk of professional and personal embarrassment together with risk of harassment or personal abuse. Again this is based on the evidence referred to above and was not challenged, contradicted or rebutted.

5. We accept on the evidence referred to above and in all the circumstances that there was a legitimate expectation by the Kennedy Panel members that their personal scores would not enter the public domain.

6. The appellant does not dispute that the disputed information is personal data. The test then for exemption under section 40(2) of FOIA is a) would disclosure be fair on the data subjects, b) is disclosure necessary for the purposes of legitimate interests pursued by the appellant and c) is any prejudice to the rights and freedoms or legitimate interests of the data subjects warranted.

7. For the avoidance of doubt we consider that the disputed information contains personal data relating to the panel members. Names are sufficient enough from which to identify someone when set in context such as they are here. Further, whilst the content of the scoring sheets clearly aren't personal data, the fact that these reflect the opinion of a named assessor is, in our opinion.

8. We would also add that we can glean more about a person than just each of their opinions on the hospital units under review, we can also glean that they are sufficiently qualified and knowledgeable to undertake such an assessment. For this reason also we are satisfied that s.40(2) is engaged.

9. Assessment of a data subject's expectations is rooted in the fair obtaining code which is defined under part II of schedule 1 to the Data Protection Act ("DPA"). This is very distinct from the fair processing requirements which require one to satisfy conditions under schedule 2 (and 3 where sensitive personal data is involved) in connection with the processing of personal data. The latter applies to legitimate use (i.e. processing), such as disclosure under FOIA. The former relates to specific information which the data subject must be told at the time their personal data is collected (or, at the very latest, before it is processed).2

10. In relation to fair obtaining, of key importance the data subject must be told :

2.(3)(a)...

(b)...

(c) the purpose or purposes for which the data are intended to be processed,

and

(d) any further information which is necessary, having regard to the specific circumstances in which the data are or are to be processed, to enable processing in respect of the data subject to be fair.

11. Both requirements relate to "fairness" which is why the two are often conflated but they are nonetheless separate tests in their own right.

12. Consequently, it is, in our view, a cardinal requirement of the DPA that data subjects be told what is going to happen with their personal data and that any other use of that personal data will be unfair except in limited circumstances, not applicable here.

13. In this case therefore, given that panelists were explicitly told by Sir Ian Kennedy that their named scores would not be disclosed, they have no expectation that this might occur. Even if we discount Sir Ian Kennedy's evidence we have that of Maria Von Hildebrand who said on oath that it was made clear individual scores would not be made public. We also have a letter from Dame Ruth Carnall which states: "The panel members had no expectation of [sic] that their individual scores would be used. They were only an aide memoir to the individuals.”

14. As well as this we have the fact that all but one panel member, after consultation by Sir Ian Kennedy following receipt of the request, stated in writing that they did not want their individual scores to be disclosed. This is a further indication of their expectations.

15. In light of this, disclosure would fail the test of fairness at the first hurdle because it would constitute a breach of the First Principle. This in itself is sufficient to engage s.40(2) FOIA and thus defeat the appeal.

Schedule 2

16. In the alternative, if we are wrong in the above, we say that disclosure of the requested information (i.e. processing) would be unfair by virtue of the fact that no condition under schedule 2 DPA could be satisfied.

17. The only possibly applicable condition to justify disclosure under FOIA if consent is refused is paragraph 6.

18. Paragraph 6 begins by stating that the processing has to be necessary for the purposes of legitimate interests pursued by the data controller or by the third party or parties to whom the data are disclosed. In this case we must consider the legitimate interests of third parties because the information would be disclosed to the general public, not just Mr Illingworth.

19. It is obviously a legitimate interest to wish to know and to examine the individual scores at the heart of this appeal and to understand how they affected the outcome of the review.

20. It is not, in our view however, necessary to examine the individual scores because they had no impact on the outcome at all. The individual scores were used simply as an aidememoir so that each panelist could take forward their personal scoring into the consensus meeting for discussion. Following discussion consensus scores were agreed upon and it was those and those alone which went forward into the assessment process. These scores have been disclosed. On top of this there were further elements of assessment (such as site visits) which gave rise to the final outcome so even the consensus scores were not determinative.

21. Even if an individual score showed massive bias this would be inconsequential if it wasn't reflected in the consensus scores. Nobody has raised any concerns about glaring inconsistencies in the consensus scores so surely this demonstrates that there are none? If that is right then why is it necessary to examine the individual scores when it is clear that even if there was bias at that stage it wasn't carried forward into consensus, the score that mattered?

22. So far as Councillor Illingworth's argument concerning unconscious bias, even if this did take place it is likely to be reflected by slightly reducing a competing score whereas conscious bias would surely result in significant low scoring. This is something which would be glaringly obvious. In any event we reiterate our point about bias at individual level being inconsequential given that it was the consensus score which mattered.

23. It seems to us that it is not necessary to disclose the individual scores simply to prove a negative (i.e. that there is no bias) because we already know that the consensus scores were the ones that mattered and as the evidence in our papers suggests these show no glaring inconsistencies. Further, we know from the evidence before us and referred to above that there were specific mechanisms in place through which to challenge bias; firstly through the process of coming to a consensus and secondly through the watchful eye of Sir Ian Kennedy.

24. Finally we do not consider disclosure to be necessary because there are other, more appropriate mechanisms through which to have concerns about bias investigated. These include (though there may be others) through Parliament, an approach to the relevant Minister or action through the Courts. Given that such avenues exist we cannot accept that it is necessary to disclose to the general public through FOIA, a method which would trammel the panelists, devoiding them of their rights under the DPA. We suggest that there probably are alternative ways to determine whether bias has affected a major public decision such as in this case.

25. So far as the rights and freedoms or legitimate interests of the panelists it seems to us on all the evidence before us and referred to above that there would be unwarranted prejudice.

26. Unpleasant and inexcusable though this is, it seems to us that anybody becoming involved in a process which leads to an unpopular and highly emotive decision such as in this case must expect to be targeted in such a way. In no way is that statement intended to excuse such vile behaviour, instead it aims to illustrate that this kind of negativity is to be expected.

27. Turning to those rights, freedoms and legitimate interests which we regard as affected, these can be summarised as follows:

a. the right to be told what will happen to your personal data

b. the right to object to processing which is likely to cause substantial damage or

substantial distress

c. prejudice to working relationships across affected hospitals

28. We are of the view that this is not only applicable in relation to an examination of the panelists' expectations, it is relevant to what must happen if, contrary to our views, disclosure was justified. It would be impossible for a data subject to invoke their right to object to processing on the grounds that it would be likely to cause them unwarranted substantial damage or substantial distress if they didn't know that disclosure had been ordered in the first place. Whilst we haven't been addressed on the likelihood of such damage or distress, neither have we seen what was sent to the panelists by Sir Ian Kennedy on this issue. Therefore we cannot know for sure that they were aware of the need to spell out that the risk of damage or distress existed, if indeed it did.

29. Any one or more of the panelists may have valid reasons for fearing such consequences and it is their right to be able to raise an objection. Disclosure is a process, thus they must be informed in advance that this is going to happen and have time before the disclosure occurs in which to raise such any objections.

30. So far as prejudice to working relationships with colleagues at hospitals where a low score was given, we endorse and adopt the submissions of NHS England given at para.12. This is not a matter to be taken lightly when viewed in the context of paediatric cardiac surgery where tensions between staff could have severe consequences.

31. In conclusion we say that the processing is not necessary and even if we are wrong on that the prejudice to the panelists is unwarranted when balanced against what would be gained through disclosure of the disputed information.”

17. The reasons for minority view were also set out. In summary, the minority member accepted that panellists did not expect their individual scores to be disclosed but did not agree that the expectation was reasonable.  He decided that the Panel members were public-facing and must have expected considerable scrutiny of their work. He did not accept that the individual scores were no more than an individual aide-memoire.  He thought that the Appellant had a very strong legitimate interest in exploring whether the process had been fair and objective at all stages.  He did not detect any prima facie evidence of bias.  Nonetheless, the information provided did not enable the Appellant to test his hypothesis. He did not accept that the threats to individual if they were identified was as significant as claimed,  nor that release of the individual scores would damage the relationship between clinicians in the centres and Panel members. On balance he concluded that the importance of transparency in the process meant that disclosure was necessary. 

 

Legal Framework

18. The general right of access to information held by a public authority provided by section 1 FOIA is subject to a number of exemptions including that in section 40(2):

 “(2) Any information to which a request for information relates is also exempt information if—

(a) it constitutes personal data which do not fall within subsection (1), and

(b) either the first or the second condition below is satisfied.

(3) The first condition is—

(a) in a case where the information falls within any of paragraphs (a) to (d) of the definition of “data” in section 1(1) of the Data Protection Act 1998, that the disclosure of the information to a member of the public otherwise than under this Act would contravene—

(i) any of the data protection principles

…

(7) In this section—

“the data protection principles” means the principles set out in Part I of Schedule 1 to the Data Protection Act 1998, as read subject to Part II of that Schedule and section 27(1) of that Act;

“data subject” has the same meaning as in section 1(1) of that Act;

“personal data” has the same meaning as in section 1(1) of that Act.”

19. By section 2(3)(f)(ii) FOIA, where 40(3)(a)(i) applies the exemption is absolute.

20. Personal data is defined as follows in section 1(1) of the Data Protection Act 1998 (“DPA”):

“‘personal data’ means data which relate to a living individual who can be identified—

(a) from those data, or

(b) from those data and other information which is in the possession of, or is likely to come into the possession of, the data controller,…”

21. There is no dispute in this case that the disputed information is personal data.

22. Schedule 1 to DPA sets out the data protection principles. The first data protection principle (“DPP1”) is:

 “Personal data shall be processed fairly and lawfully and, in particular, shall not be processed unless—

(a) at least one of the conditions in Schedule 2 is met…

23. The only relevant condition in this case is condition 6(1):

“The processing is necessary for the purposes of legitimate interests pursued by the data controller or by the third party or parties to whom the data are disclosed, except where the processing is unwarranted in any particular case by reason of prejudice to the rights and freedoms or legitimate interests of the data subject.”E+W+S+N.I.

24. In Goldsmith International Business School v Information Commissioner and Home Office [2014] UKUT 563 at [35] – [42], Upper Tribunal Judge Wikeley set out the proper approach to condition 6(1) in the following propositions:

Proposition 1: Condition 6(1) of Schedule 2 to the DPA requires three questions to be asked:

“(i) Is the data controller or the third party or parties to whom the data are disclosed pursuing a legitimate interest or interests?

(ii) Is the processing involved necessary for the purposes of those interests?

(iii) Is the processing unwarranted in this case by reason of prejudice to the rights and freedoms or legitimate interests of the data subject?”

 

Proposition 2: The test of “necessity” under stage (ii) must be met before the balancing test under stage (iii) is applied.

 

Proposition 3: “Necessity” carries its ordinary English meaning, being more than desirable but less than indispensable or absolute necessity.

 

Proposition 4: Accordingly the test is one of “reasonable necessity”, reflecting the European jurisprudence on proportionality, although this may not add much to the ordinary English meaning of the term.

 

Proposition 5: The test of reasonable necessity itself involves the consideration of alternative measures, and so “a measure would not be necessary if the legitimate aim could be achieved by something less”; accordingly, the measure must be the “least restrictive” means of achieving the legitimate aim in question.

 

Proposition 6: Where no Article 8 privacy rights are in issue, the question posed under Proposition 1 can be resolved at the necessity stage, i.e. at stage (ii) of the three-part test. 

 

Proposition 7: Where Article 8 privacy rights are in issue, the question posed under Proposition 1 can only be resolved after considering the excessive interference question posted by stage (iii).

 

 

Discussion  

Ground 1: Fairness

25. It appears from paragraphs 12 – 15 of the majority’s reasons that they thought that disclosure would necessarily be unfair where the data subject was told that the information would not be disclosed.  If so, they would have been wrong.  However, I do not consider there was any material error of law here. It may be, as Mr Hopkins submits, that in that part of its reasons the tribunal was simply drawing a distinction between the fair processing requirements of the DPA (in paragraphs 1 and 2 of Part II of Schedule 2) and substantive fairness which was dealt with under the heading “Schedule 2”.  But even if it was not, its findings in relation to condition 6(1) inevitably meant that disclosure would be unfair.  In addition to its conclusion that the panellists had a legitimate expectation that their named scores would not be disclosed, the tribunal had found as a matter of fact that disclosure would probably put at least some of the panellists at risk of harassment, abuse, and personal and professional embarrassment.  Even though at paragraph 26 the tribunal’s view was that the panellists could expect some harassment and abuse, in the light of the tribunal’s view as to the irrelevance of the information sought and that any bias would have been inconsequential, it is not surprising that the tribunal attached such significance to the strongly held expectations of the panellists that their names would not be disclosed.  In any event, even if the tribunal approached fairness in error of law, that was not material in the light of its finding that no Schedule 2 condition applied.  As I explain below, I find no error of law in that finding which meant that the disputed information could not be disclosed. 

26. Ms Gannon also submits that the First-tier Tribunal failed to give any or proper weight to the fact that the Panel members were an essential and high level part of a public review process as a result of which they must have had some expectation that their decision-making process would be open to public scrutiny. She relies on the observations of the Information Tribunal in House of Commons v Information Commissioner and Norman Baker EA/2006/0015&0016 at paragraph [78]:

“…we find that when assessing the fair processing requirements under the DPA that the consideration given to the interests of data subjects, who are public officials where data are processed for a public function, is no longer first or paramount. Their interests are still important, but where data subjects carry out public functions, hold elective office or spend public funds they must have the expectation that their public

actions will be subject to greater scrutiny than would be the case in respect of their private lives.”

27. In addition she says that the expectation that their work would be publicly scrutinised was enhanced by the fact that the High Court subsequently quashed the decision made by the JCPCT following the Review which was then abandoned. 

28. The problem with these submissions is that they fail to distinguish between the public role of the Panel and the role of the individual members.  The work of Panel was made public and could be scrutinised.  The consensus scores and the very detailed report of the Panel’s deliberations were published. The individual scores were also published, but not attributed to identified panellists.  It was the non- attribution of the scores to them which the Appellant challenged.  But in that regard the First-tier Tribunal found as a fact that the individual Panel members were not public facing (paragraph 3). The tribunal adopted the Commissioner’s reasoning which included that the Panel had a collective identity and was only public facing through its chairman. The tribunal relied on the “compelling” evidence of Ms Hildebrand (one of the Panel members), which had not been challenged.  She had explained that she was an expert Panel member but not in any public facing capacity, and that although the consensus scores were public the individual scores were “irrelevant and a distraction to the ultimate decision”.  That is a finding of fact which was open to the tribunal on the evidence.  On the basis of that finding of fact, the approach in the House of Commons case is irrelevant. 

 

Ground 2: Condition 6(1)

i) The role of the individual scores

29. At the heart of the Appellant’s case before the First-tier Tribunal and repeated before the Upper Tribunal are two main contentions:

a)    The initial individual scores played a very important part in the Review and its outcome; and

b)    Disclosure of those scores by reference to the individual panellists would help to identify whether any individual scores were tainted by actual or apparent bias.

30. The First-tier Tribunal did not find that the individual scores were tainted by bias but, in any event, it rejected the contention at (a) and so it followed that, even if there was bias in the initial scores, it was irrelevant to the consensus scores which were what mattered.

31. I have described the process adopted by the Kennedy Panel.  The Panel members produced their own scores for each centre at an early stage of the process.  When they came together after each visit, the Panel collectively identified the relevant evidence and came to a collective view including consensus scores.  There was no evidence that that the panellists attempted to adhere to their individual scores or to persuade others to do so.  The consensus scores were not set by averaging or other calculation involving the individual scores.  The initial individual scores reflected the individual panellists’ private preliminary views, but there was no evidence that they had any other bearing on the consensus.  The Panel’s report of December 2010 explained that they had not sought to compare centres.  Each was scored on its merits, not relative to other centres.  There was evidence before the First-tier Tribunal (see the letter dated 18 January 2013 responding to the Appellant’s request for an internal review) that “throughout the decision making process, the JCPCT has had no regard to the individual Panel members’ scores” and that those scores were not even disclosed to other Panel members.  Indeed, the NHS had explained that the individual scores had been located and transcribed from their original form (handwritten notes held in storage boxes) in order to respond to the request for information.

32. Moreover the evidence as to the nature of the Review process showed that, as the Review progressed, the individual scores became increasingly remote from and irrelevant to the conclusions and the consequent decision. The Panel had no role in the weighting of the consensus scores. The weighted scores fed into only one of the four criteria used by the Review to identify the proposed options. Each option comprised a group of four hospitals and they were tested by reference to the application of the criteria to each group rather than to the individual hospitals. Although the option which was finally selected comprised centres which were scored the highest by the Kennedy Panel, the evidence was tested rigorously by the JCPCT not only on the basis of the Panel’s scores but a variety of other factors and in particular the consultation responses of patients and clinicians.  This was set out in considerable detail in the Decision Making Business Case, which was a very detailed 218 page document explaining how the JCPCT made it final recommendations.

33. The First-tier Tribunal was well aware of this context.  It formed much of the basis of the submissions and evidence which it had summarised in the earlier part of its decision.  The above is highly material to the First-tier Tribunal’s approach to the appeal and its conclusion that the individual initial scores were irrelevant to the consensus scores. 

34. The Appellant did and does not accept that the initial scores played such a limited role in the process.  But in any event, he said in the First-tier Tribunal and repeated here that his concern was to scrutinise the scores for possible bias by individual panellists.  If a panellist was biased in favour of or against a particular centre, that might have affected their input into the collective discussion and the consensus scores which could not be ignored in the light of the decision of Nicola Davies J in the Save our Surgery judicial review.  Ms Gannon says that, even if any bias could not be detected from the consensus scores themselves (which by their nature did not single out contributions of individual panellists), the initial individual scores might reveal bias.  It follows, she says, that the tribunal was wrong to say that the individual scores did not matter simply because there was no evidence of bias in the later scores.

35. I reject this submission.  There was no evidence of bias on the part of Panel members at the initial stage.  The Appellant’s case was speculative.  In that context, it was understandable that the tribunal should not explore such allegations but instead consider the possible relevance of any intitial bias even if it existed. In particular:

a)    The tribunal pointed out that no-one had raised glaring inconsistencies in the evidence.  There was enough information to identify a possibility of bias, if that had existed.  The panellists’ backgrounds, and their declarations of interest in respect of each centre, were public. That provided a starting point for assessing a risk of bias. 

b)    Even though the initial scores were anonymised, it would have been possible to look for something unusual in the allocation of scores such as whether a (albeit anonymised) panellist was awarding unusually high or low scores for a centre. It is telling that the minority member analysed the individual scores in the light of the Appellant’s suggestions of bias but could not find any prima facie evidence of bias.  In addition, the NHS asked one of its senior statisticians to examine the data for evidence of bias and he found none. I note that the minority member thought that the statistician may not fully have addressed the Appellant’s concern as to the way in which bias may have operated but nonetheless it adds to the overall picture as seen by the majority, which was that there was no evidence of bias. 

c)    If the consensus scores were radically different from the majority of the initial scores but were more consistent with the score of one unusually high or low scoring individual, that might have indicated that bias had been carried through to the consensus scores.  But no such suggestion was made.

d)    When the Appellant complained about the Review to the Secretary of State for Health in November 2012, he did not allege bias in the initial individual scores. The Independent Reconfiguration Panel, which advised the Secretary of State in April 2013 on the proposals for change, did not mention any concern about bias on the part of the Kennedy Panel members.  The grounds of claim in R (Save our Surgery Ltd) v JCPCTs did not involve allegations of bias in the Kennedy Panel. Ms Gannon sought to explain this on the basis that the initial scores had not been disclosed at that point, but the Appellant’s case is that the anonymised initial scores did not enable him to scrutinise for bias and therefore disclosure of those scores is not relevant to his ability to allege bias.

e)    Ms Gannon says that the Appellant had set out a prima facie case as to bias, in his written closing submissions to the tribunal.  In those submissions he set out the affiliations between a number of the panellists and some hospitals, all of which was already in the public domain.  The submissions were addressed by the NHS before the First-tier Tribunal However there was nothing there to advance a coherent theory of bias. 

f)     In R (Royal Brompton and Harefield NHS Fondation Trust) v JCPCT [2012] EWCA Civ 472 the Court of Appeal dismissed a claim that one of the Panel members had been influenced by bias because, after the publication of proposed options in the consultation document, he lent his name to a public campaign to save one of the units (at which he had previously worked.  The Court said at [138]:

“It is one thing... to object to a proposed outcome to which you have yourself contributed.  It is another thing, and a very serious one, to have attempted by stealth to obviate that outcome.  To deduce the second from the first in the absence of firm evidence requires an assumption that the individual concerned was prepared to forfeit professional objectivity in favour of partisanship. That is not an assumption which in our judgment the fair minded and objective observer would be prepared to make.  She would look for evidence of the assumed link, and in the present case she would find none.”

I recognise that the Appellant in this case submits that disclosure was needed in order to detect bias which cannot be identified, if it exists, from the disclosed information.  Nonetheless, the above observation of the Court of Appeal illustrates just how far the Appellant’s case is from a coherent theory of possible bias.  In the Royal Brompton case the Court was considering a panellist who had a previous affiliation with a hospital which he subsequently supported in a public campaign, but that did not get close to establishing bias particularly in the light of the professional role of the individual.  All the Appellant has asserted in the present case is the professional affiliations of some panellists with hospitals or their specialisms, all of which have been publicly declared.  It gets nowhere near to suggesting even a risk of bias and it is not surprising that the First-tier Tribunal did not see the need to explore it further.

36. The tribunal also found that, even if an individual was biased, there were other mechanisms in place to reduce the risk of bias affecting the process.  This was a reference to the Commissioner’s submissions that the consensus scores were not an average of the initial scores, that the consensus scores were weighted independently, and the JCPCT did not see the individual or consensus scores.

37. Ms Gannon says that the NHS was clearly concerned that individual bias at the initial stage might taint the process, because it asked a panellist not to participate in the initial scoring process because of the possibility of bias.   This is merely a challenge to the tribunal’s evaluation of the evidence but does not demonstrate an error of law. In any event, the approach to that Panel member’s role could equally support the Respondent’s case.  The Panel member in question took part in the post-visit discussions and contributed to the consensus scores.  The concerns about bias cannot have been great.

38. Ms Gannon also says that the Appellant relied not only on possible bias but also on a broader interest in understanding the role of individual panellists in the process, for instance whether panellist’s scored their own clinical specialism differently to other areas, or whether any individual exerted a disproportionate influence on their colleagues.  These are simply other ways of putting the bias point.  The Appellant’s written submissions to the tribunal were concerned with bias.  The tribunal cannot be criticised for failing to detect another factor.

39. I now turn to the Appellant’s remaining grounds of appeal. 

 

ii) Necessity

40. Ms Gannon submits that the Tribunal erred when considering whether disclosure was “necessary”.  At the heart of her this is her disagreement with the First-tier Tribunal’s conclusion that the individual scores “had no impact on the outcome at all” (paragraph 20). As Ms Gannon fairly acknowledges, this is a challenge to findings of fact by the First-tier Tribunal. However, she says that this finding was in error of law because it contradicted the evidence which showed that the individual scores played a very important role in the Review and its outcomes, as explained by Ms Hildebrand who said that “the individual scores were used to reach a consensus score”. 

41. This submission takes Ms Hildebrand’s comment out of context.  It is clear from the First-tier Tribunal’s reasons that the tribunal understood the process adopted by the Kennedy Panel and therefore that the individual scores were the starting point for the Panel’s discussions. The point being made by the tribunal at paragraph 20 was that those scores were ultimately irrelevant. That is consistent with Ms Hilderbrand’s evidence, which described the process that I have explained earlier. For the reasons which I have explained, that was a conclusion which the tribunal was entitled to reach on the evidence.

42. Ms Gannon is wrong to say that this conclusion cannot stand in the light of the judgment Nicola Davies J in R (Save our Surgery Ltd) v JCPCT and another [2013] EWHC 439 (Admin), a decision on a judicial review of the JCPCT’s decision, in which the Judge found that the Kennedy Panel’s scores on Quality were an important factor in the final decision and that the subscores could not be described as no more than “underlying workings” but were an important component of the Quality assessment. She was not here talking of the initial individual subscores which are the subject of this appeal, but the Panel’s consensus subscores on the individual criteria which led to the final consensus scores.  Consistently with that decision, the First-tier Tribunal in this appeal accepted that the consensus scores were important: they were the scores “that mattered”.

43. The Appellant also submits that the tribunal applied the wrong test for “necessity” because it set the bar too high and failed to consider whether disclosure would make “furthering the purposes of a legitimate interest more effective”. This is a reference to paragraph [23] of the judgement of Baroness Hale in South Lanarkshire Council v Scottish Information Commissioner [2013] UKSC 55.  But the Supreme Court did not say that disclosure will always be necessary where it makes furthering the purposes of a legitimate interest more effective. The Court confirmed the well-established test of necessity where, as here, Article 8(2) of the European Convention is engaged (see Goldsmith proposition 5): the measure must be the least restrictive for the achievement of a legitimate aim. The First-tier Tribunal’s approach here, in particular at paragraphs 21-24, was consistent with that.

44. Overall the tribunal’s approach was consistent with the guidance in Goldsmith.  It identified the interest for which disclosure was sought, at paragraph 19. Subsequent paragraphs show that the tribunal understood that the particular focus of that interest was to identify possible bias by panellists. It considered the necessity of disclosure for the purpose of that interest (paragraphs 20-24).  There was nothing in its approach to indicate that the tribunal applied a test of absolute necessity.  On the contrary, it is clear from the tribunal’s reasons that it did not consider that there was any necessity for disclosure: it was not necessary to disclose the scores to prove a negative because that could be discovered from the available information.  The tribunal considered the interference with the rights of the panellists and then, consistently with proposition 7 in Goldsmith, addressed whether disclosure was unwarranted in the light of that.

 

iii) The nature of the data

45. The First-tier Tribunal adopted the Commissioner’s submissions.  The Commissioner had submitted that, on the basis of the disclosed information, it was “entirely possible for a member of the public to track the individual scores given by a specific Panel member (albeit in an anonymous form) for each hospital assessed.”  Ms Gannon says that that was wrong and shows that the tribunal had not understood the nature of the information disclosed. 

46. The First-tier Tribunal was provided with an explanation of the way in which the disclosed scores had been presented, both in the Information Commissioner’s decision notice and in the review letter referred to above.  Because two Panel members had not provided initial scores on some centres, it was decided not to use a numerical cipher which referred to the same Panel member in every assessment because that would enable a person to identify those two individuals.  There is no reason to suppose that the tribunal misunderstood this.  The Commissioner’s submission which the tribunal recorded was that the tracking of scores by a specific Panel member was possible “for each hospital assessed”, ie on a hospital by hospital basis.  That was correct. 

47. That is sufficient to dispose of the allegation that the tribunal misunderstood the evidence. But, had I had any doubt, it would have been dispelled by the following.  First, the tribunal had seen both the anoymised and the withheld attributed scores and so the randomisation of the scores would have been obvious to it.  Second, in his written closing submissions to the First-tier Tribunal the Appellant had addressed the Commissioner’s submissions about tracking and the minority member addressed this in some detail including the Appellant’s contention that it was not possible to compare an individual panellist’s score across centres.  It strains credulity to conclude that the members would not have discussed these issues.  I remind myself of the observation of Lord Hope in Jones v First-tier Tribunal and Criminal Injuries Compensation Authority [2013] UKSC 19 at [25]:

“It is well established, as an aspect of tribunal law and practice, that judicial restraint should be exercised when the reasons that a tribunal gives for its decision are being examined. The appellate court should not assume too readily that the tribunal misdirected itself just because not every step in its reasoning is fully set out in it.”

48. In the light of these factors, the First-tier Tribunal’s decision cannot reasonably be read as meaning that it misunderstood the way in which the materials were presented.

 

iv) Bias

49. In addition to the submissions about bias which I have already addressed above, Ms Gannon submits that there was no evidence to support the tribunal’s decision that even if an individual panellist had been unconsciously biased, this was “likely to be reflected by slightly reducing a competing score whereas conscious bias would surely result in significant low scoring. This is something which would be glaringly obvious…”.  She says that it is not clear why unconscious bias would result in less pronounced divergence of scores than conscious bias and that, in any event, the tribunal did not appreciate that ultimately the consensus scores were separated by only one mark.

50. There was no need for evidence on the point.  Here the tribunal was explaining part of its reasoning underlying its assessment of the Appellant’s case. That was a matter for its own judgment by reference to common sense.  There is nothing irrational in the view that a consciously biased attempt to influence the scores would result in greater divergence than an unconscious one.

51. In any event, in the light of the tribunal’s principal conclusions as to the irrelevance of any bias at the initial stage, this aspect of the appeal does not get the Appellant anywhere. 

 

v) Alternative measures

52. An alternative measure must be the least restrictive means of achieving the legitimate aim (proposition 5 in Goldsmith). Ms Gannon submits that the tribunal misdirected itself at paragraph 24 in finding that there were alternative measures for investigating concerns about bias. She says that this shows that the tribunal considered whether there were alternative measures for investigating bias rather than whether there was a means of disclosure which would be less restrictive. 

53. This submission is misconceived. The question is whether there is a less restrictive means of achieving the aim, which in this case was to detect bias. If there was a possibility of giving disclosure which would achieve this in a way which was less restrictive of the data subjects’ rights, that would be relevant.  

54. The only possibility identified by Ms Gannon was to disclose anonymised but trackable scores. This was not realistic for the reasons explained by the NHS on internal review and endorsed by the minority tribunal member: it would have revealed who at least two Panel members were and so would have heightened the risk of identification of the others (because of the reduction of the residual pool).   There was no error of law by the tribunal in not identifying that as an alternative measure.  Ms Gannon also submitted that the tribunal could have considered the possibility of disclosing the trackable scores for the centres which all panellists visited. Even though anonymised, it would be possible to compare the scores for the same individuals across the centres and that could have revealed any startlingly different score for any one centre. That was not a suggestion made to the tribunal but the minority member did consider and dismiss the possibility. As he said, “the genie is out of the bottle”:  if the Appellant were to be given the non-randomised information for the other six, he could quickly identify the scores of those who missed some visits because they were absent for different centres.

 

vi) Legitimate interest

55. At paragraph 18 the tribunal said that it had to consider the legitimate interests not only of the Appellant but also the general public.  That is not a correct approach to the question of necessity, but the interests of the public may be relevant when considering whether processing was unwarranted by reason of the prejudice to the rights and freedoms or legitimate interests of the data subjects: see GR-N v IC and Nursing and Midwifery Council [2015] UKUT 449 AAC at [20]-[24].  Despite the error in the tribunal’s expression of the test, this was not reflected in the substance of its decision.

56. Ms Gannon submits that the tribunal failed to consider the specific interests of the Appellant, in his capacity as Chair of the JHOSC.  The problem with this submission is that has no traction in this case, where the tribunal had found that the information requested could not assist in detecting bias or otherwise understanding the Panel’s work.  The strength of the Appellant’s interest in detecting bias, arising from the capacity in which he sought to discover it, could make no difference because the tribunal had found that the information was irrelevant and disclosure of the information would not assist in detecting bias. 

57. The Appellant now relies on his statutory responsibilities to investigate and scrutinise the Review procedures. He did not do so before the First-tier Tribunal and so the tribunal cannot be criticised for failing to consider it.  This makes it unnecessary for me to explore in detail the relevant statutory provision, regulation 5 of the Local Authority (Overview and Scrutiny Committees Health Scrutiny Functions) Regulations 2002 which were the relevant regulations in force at the time.  It suffices to say that I accept Mr Hopkins’ submissions that this regulation does not assist the Appellant. The obligation is to provide such information as the overview and scrutiny committee may “reasonably require in order to discharge its functions”.  In the light of the First-tier Tribunal’s conclusions as to the irrelevance of the information sought, the disputed information cannot be said to have been reasonably required by the committee.  Moreover, the regulations do not require the provision of confidential information relating to and identifying a living individual unless the individual consents or it is disclosed in a form from which the identity of the individual cannot be identified.

 

vii) Prejudice

58. The tribunal relied on its finding that there would be a risk of prejudice to panellists’ working relationships.  Ms Gannon submits that there was no or insufficient evidence upon which the tribunal could have made that finding.

59. In its closing submissions to the tribunal the NHS relied on the risk of professional problems and disgruntled colleagues. There was evidence before the tribunal in this respect. The tribunal referred to some of it at paragraph 30. This included the internal review decision by NHS London.  The tribunal referred to some of the evidence at paragraph 48e of its summary of the NHS case, where it referred to “incorrect aspersions being cast about a particular individual panel member”.  This appears to be a reference to NHS England’s letter to the ICO of 17 May 2013.

60. I accept that the evidence about prejudice to working relationships was not strong, but the case for disclosure was very weak as the tribunal had found.  Not much was required to find that disclosure was unwarranted.

 

Conclusion

61. For the reasons given I dismiss the appeal.

 

 

 

Signed on the original Kate Markus QC

on 22 June 2016 Judge of the Upper Tribunal