United Kingdom Competition Appeals Tribunal

IN THE COMPETITION
APPEAL TRIBUNAL

		25 July 2014

RULING (APPLICATION FOR DISCLOSURE)

Introduction

1. HCA International Limited ("HCA") has brought an application for review under section 179 of the Enterprise Act 2002 to challenge a decision of the Competition and Markets Authority requiring, among other things, HCA to divest itself of some of the private hospitals it owns. The Competition and Markets Authority has taken over functions previously carried out by the Competition Commission, and for ease of reference we refer to them compendiously as "the CMA".
2. This is the Tribunal's ruling on an application brought by HCA in the course of the section 179 proceedings for disclosure to be given by the CMA. We heard argument on this application at a case management conference on 8 July 2014.

Factual Background and the Parties' Positions

3. In April 2012, the Office of Fair Trading made a reference to the CMA for it to investigate the supply of privately funded healthcare services.
4. The CMA proceeded to carry out an extensive investigation, in the course of which it gathered a good deal of information from private hospitals regarding charges they had made to insurance companies and individuals for healthcare services. This information took the form of invoices, recorded in electronic format, for services supplied over the period 2007 to 2011. We refer to this body of material as "the raw data".
5. The CMA analysed the information it gathered, including the raw data, and drew the conclusion (amongst others) that there were adverse effects on competition ("AECs") associated with the structure of the market for privately funded healthcare services in central London, and in particular the concentration of private hospitals owned by HCA in central London. The CMA's conclusion that there were relevant AECs turned critically on its Insured Prices Analysis ("IPA") based upon the raw data. In short summary, the IPA indicated that HCA faced weak competitive constraints from its rivals in central London and, in particular, that it charged significantly higher prices to insurers than other healthcare providers.
6. In its final report dated 2 April 2014, the CMA determined that the remedy for the relevant AECs was that HCA should be required to divest itself of two of its hospitals in London. This remedy plainly represents a serious intrusion upon and interference with HCA's business and property interests.
7. The challenge against the CMA's decision brought by HCA in the Tribunal will involve, amongst other things, a challenge under Article 1 of Protocol 1 to the European Convention on Human Rights to the proportionality of the requirement that HCA divest itself of two hospitals. As part of its challenge, HCA disputes that there was a proper basis in the IPA for the CMA to conclude that there was any significant inflation of prices charged by HCA, or any distortion of the market such as to justify the imposition of the remedy of divestiture.
8. HCA has already indicated various criticisms which it will seek to make regarding the IPA and the CMA's analysis of the data it obtained. It says that there are features of the CMA's final report which give reasonable grounds to suspect that there may have been defects in the CMA's analysis in preparing the IPA. HCA makes the present application for disclosure in order to have access to the data and modelling methodology which were used in producing the IPA, in order to obtain evidence which will assist it to make good the criticisms it is seeking to make in relation to the IPA.
9. In order to understand HCA's application, it is necessary to summarise the process the CMA went through to compile the IPA and extract relevant comparative pricing data by reference to which it could test whether HCA's prices appeared to be inflated beyond what would be expected in an undistorted market. Having obtained the raw data from a number of healthcare suppliers, the CMA "cleaned" the data before using it to input into the IPA. The cleaning process involved: (a) removing certain items from invoices, where those items appeared to be remote from or peripheral to the treatment to which the invoice related; (b) aggregating the remaining items in each invoice (such as the cost of surgery, anaesthetic, medicines and so forth) so as to give one overall figure for the particular procedure in question (say, a hip operation); and (c) excluding from the data set invoices which appeared incredible (e.g. because the values were so far out of line with others as to suggest clerical error).
10. Such aggregated, simplified cost items (which we call "the cleaned data") were then used as the inputs for the CMA's computer modelling to produce the IPA. They were subject to further, different processes of aggregation to give average figures for particular procedures which could provide the foundation for a comparative analysis of whether charges made by particular private hospitals or operators were out of line with those made by others. The cleaned data were subjected to statistical analyses, including the use of regression equations, particular codes for use in the 'Stata' statistical software used to analyse the data, the calculation of coefficients and adjusted R² values. HCA says it does not have access to and is unaware of the content of the raw data, the cleaned data and the full methodology used by the CMA in developing the IPA.
11. The material of which the HCA seeks disclosure is as follows:

(a) the raw data;

(b) the cleaned data;

(c) full details of the methodology, analyses and various coding values used in the computer modelling to produce the IPA;

(d) the full set of results from each step of the analysis, including all the standard outputs; and

(e) the full set of results from any sensitivity analysis or robustness checks which the CMA performed (for example, regarding the sensitivity of the results to the inclusion of rebates in the analysis).

HCA wishes to have access to all these data and the CMA's computer model in an executable form, so that its economists can test and review the processes adopted by the CMA in producing the IPA.

12. The CMA resists the application for disclosure. It says that the disclosure sought goes well beyond what is required in judicial review proceedings, the principles for which apply in relation to the present challenge. The CMA says that, in reality, all that HCA is seeking to do is to check the CMA's calculations in circumstances where there is no good basis for thinking that there is any defect in those calculations. The CMA contends that the application is an inappropriate "fishing expedition" by HCA on the basis of a mere speculative hope that, if it is given all this information, something might turn up to assist its case. That, the CMA says, is not a proper basis for ordering disclosure in proceedings such as this.
13. The CMA also emphasises that the raw data and the cleaned data contain highly sensitive commercial information regarding the prices charged and pricing and commercial strategies pursued by HCA's competitors in the private healthcare market. The provision of such sensitive commercial information of competitors to a larger and more powerful rival in the market could, if not adequately protected by confidentiality arrangements regarding the handling of the information, in itself be detrimental to maintaining a properly competitive market, by giving HCA an unfair commercial advantage by seeing how its competitors behave. The CMA says that this underlines the need for the Tribunal to adopt a cautious approach to HCA's application, and that HCA has not made out sufficient grounds to justify the Tribunal in ordering the CMA to provide such commercially sensitive information to it. This point was forcefully supported in submissions made on behalf of The London Clinic, a provider of private healthcare services in London which competes with HCA, whose commercial data are included in the raw data and the cleaned data.
14. Furthermore, the CMA says that even though the raw data, cleaned data and details of the methodology used to produce the IPA are all material which is relevant to the production of the final report, it is unnecessary and would be disproportionately burdensome for the CMA to have to provide it. The raw data set and the cleaned data set are both very large. The computer model is complex. The data sets and the model would need to be loaded on to a dedicated server, and the CMA has no server which it can spare from the other public functions it has to carry out in order to provide this facility. Moreover, in view of the commercial sensitivity of the confidential data in question, provision of the data would need to take place in a highly controlled environment in a data room provided by the CMA and supervised by CMA staff. All this would be costly and burdensome to provide.
15. A yet further point made by the CMA is that the volume of the data sought and the complexity of the modelling involved are such that it would take a very considerable period of time for those acting for HCA to undertake the review which it says it wants to carry out, with the result that the timetable already laid down for preparation and hearing of the present challenge could not be met. If an order for disclosure is made, that would involve considerable delay in dealing with the present challenge, which would itself be contrary to the public interest, not least because - pending determination of the challenge - the Tribunal has made an order suspending the requirement that HCA divest itself of two hospitals in central London to remedy the relevant AECs identified by the CMA. If it transpires that the CMA's findings of relevant AECs and its decision regarding the divestment remedy are all lawful and proper, delay in implementation of the remedy will be to the detriment of the public interest in the prompt implementation of that remedy.

Discussion

16. The Tribunal has power to order disclosure under rule 19 of the Competition Appeal Tribunal Rules 2003 (S.I. 1372 / 2003). In considering HCA's disclosure application, we have had regard to the factors relevant to a decision under rule 19 (whether disclosure is required "to secure the just, expeditious and economical conduct of the proceedings") and, by analogy, to the overriding objective in Part 1 of the Civil Procedure Rules – namely, seeking to ensure that the parties are on an equal footing; the desirability of minimising the cost of securing justice; dealing with the case in ways which are proportionate to the amount of money involved, to the importance of the case, to the complexity of the issues and to the financial position of each party; ensuring that the case is dealt with expeditiously and fairly, allotting to it an appropriate share of the Tribunal's resources while taking into account the need to allot resources to other cases; and the desirability of enforcing compliance with rules, practice directions and orders. The need for disclosure must be examined in light of the circumstances of the individual case, taking into account the nature of the decision challenged, the grounds of challenge and the extent of the disclosure which is sought.
17. We have not found the determination of HCA's application for disclosure easy. There is force in the objections made by the CMA and The London Clinic. However, on balance, we have come to the conclusion that an order for disclosure is necessary and proportionate and is required to enable HCA's challenge to the CMA's findings and decision on remedy to be determined fairly and justly.
18. We begin by discussing some of the practical matters to which an order for disclosure would give rise. Through debate at the hearing, in the course of exploration by the Tribunal of the practical options, a broad consensus emerged regarding the way in which disclosure should be handled, if the Tribunal decided that disclosure should be ordered.
19. The first step would be for a data room to be established by the CMA and supervised by it, in which HCA's representatives would have access to a server loaded with the raw data, the cleaned data and the computer model used by the CMA in producing the IPA. We were told that the data sets and the computer model still exist. They would not have to be reconstructed from scratch. HCA has indicated it is willing to provide a server for use in the data room, which would meet one of the CMA's objections about the diversion of its resources from carrying out its other functions.
20. We pressed Ms Rose QC, for HCA, on the time required for HCA's representatives (a team under Dr Mazzarotto, Head of Competition Economics at KPMG) to conduct the review of the data, methodology and computer modelling required to identify any alleged flaws in the process which HCA might then wish to deploy in evidence. Fortunately, Dr Mazzarotto was at the hearing and Ms Rose was able to take instructions directly from him. It should be noted that Dr Mazzarotto already has a good understanding of the nature and volume of the data and the outlines of the methodology used in the computer model, from his previous involvement for HCA in responding to the CMA's investigation, and so was able to give an estimate in which the Tribunal could have confidence. Ms Rose told us, on instructions, that access to the data room for a period of one month would be sufficient. That is an estimate which HCA can expect to be held to, absent very good reason to the contrary.
21. Review of the data and computer model within such a period would not cause the existing timetable for determination of HCA's challenge to the findings and decision in the final report to have to be changed. A further safeguard in relation to the timetable, which would assist HCA in reviewing this material, would be if the disclosure order has built into it a simple procedure whereby HCA could raise technical questions with the CMA in the course of the review to assist it in understanding the CMA's methodology. We invite the parties to seek to agree terms to allow for this.
22. On the basis of the one month period indicated by HCA, we do not consider that the provision of a room by the CMA to serve as the data room and of staff to supervise the KPMG team and enforce confidentiality procedures while they use the data room would be disproportionately onerous for the CMA in the context of these proceedings. We were not given detailed information by the CMA regarding the likely costs for this, nor of any particular difficulties which might arise from proceeding in this way. On the other hand, the matters at stake for HCA (an obligation to divest itself of two high value private hospitals in London) are of considerable weight and importance for it.
23. Those having access to the data room will be bound by strict obligations to protect the confidentiality of the sensitive commercial data of which disclosure is sought, including obligations equivalent to those on persons within Confidentiality Ring 1 (governing access to super-sensitive confidential commercial information) established by the Tribunal's Confidentiality Order of 2 July 2014 in these proceedings. These obligations include an undertaking not to advise any party in relation to any pricing negotiations between any hospital operator and any private medical insurer concerning the price and/or terms and conditions of services supplied to patients of the insurer for a period of two years. In our view, the professionals acting for HCA subject to these obligations can be relied upon to respect them. Moreover, the most recent confidential commercial data which would be made available in the data room date from 2011 and so are now relatively aged. In the circumstances, therefore, we consider that the risk of misuse of highly sensitive confidential commercial data by HCA or its advisers in relation to its competitors is minimal and is within acceptable bounds.
24. If, as a result of its review in the data room, HCA identifies matters which it would wish to introduce into the proceedings by way of amendments to its pleaded case or as evidence, the parties should at that stage discuss practical handling measures which would allow that to be done. If concrete proposals are put forward at that stage, the Tribunal will be in a position to issue appropriate further directions.
25. Accordingly, in our view, none of the practical issues to which the CMA and The London Clinic called attention is sufficient, by itself or cumulatively, to lead to the conclusion that an order for disclosure would be disproportionately burdensome or disruptive and ought to be refused on that ground.
26. We turn, then, to consider the main issue on the application, which is whether HCA's application goes beyond the circumstances in which disclosure may be ordered in judicial review proceedings according to usual principles.
27. Both sides prayed in aid the guidance regarding disclosure in judicial review proceedings given by Lord Bingham in Tweed v Parades Commission for Northern Ireland [2006] UKHL 53; [2007] 1 AC 650, at paras. [1]-[4]:

"1.   … the issue in this appeal is whether discovery of five documents held by the Parades Commission should be ordered for purposes of Mr Tweed's application for judicial review, to the extent that such application turns on a proportionality argument under the Human Rights Act 1998 and the European Convention on Human Rights.

  2. The disclosure of documents in civil litigation has been recognised throughout the common law world as a valuable means of eliciting the truth and thus of enabling courts to base their decisions on a sure foundation of fact. But the process of disclosure can be costly, time-consuming, oppressive and unnecessary, and neither in Northern Ireland nor in England and Wales have the general rules governing disclosure been applied to applications for judicial review. Such applications, characteristically, raise an issue of law, the facts being common ground or relevant only to show how the issue arises. So disclosure of documents has usually been regarded as unnecessary, and that remains the position.

3. In the minority of judicial review applications in which the precise facts are significant, procedures exist in both jurisdictions, as my noble and learned friends explain, for disclosure of specific documents to be sought and ordered. Such applications are likely to increase in frequency, since human rights decisions under the Convention tend to be very fact-specific and any judgment on the proportionality of a public authority's interference with a protected Convention right is likely to call for a careful and accurate evaluation of the facts. But even in these cases, orders for disclosure should not be automatic. The test will always be whether, in the given case, disclosure appears to be necessary in order to resolve the matter fairly and justly.

  4. Where a public authority relies on a document as significant to its decision, it is ordinarily good practice to exhibit it as the primary evidence. Any summary, however conscientiously and skilfully made, may distort. But where the authority's deponent chooses to summarise the effect of a document it should not be necessary for the applicant, seeking sight of the document, to suggest some inaccuracy or incompleteness in the summary, usually an impossible task without sight of the document. It is enough that the document itself is the best evidence of what it says. There may, however, be reasons (arising, for example, from confidentiality, or the volume of the material in question) why the document should or need not be exhibited. The judge to whom application for disclosure is made must then rule on whether, and to what extent, disclosure should be made."

28. Ms Rose emphasised that HCA's challenge will involve challenging the rationality of the CMA's findings that the relevant AECs existed and the proportionality of the divestiture remedy which it has decided to impose. In relation to the latter, HCA's case is that the remedy constitutes a major interference with its business and property and accordingly it requires robust justification by reference to the underlying data fed into and analysed in the IPA, which was used as the foundation for making those AEC findings and as warranting that remedy. HCA had not been able to review the inputs, methodology and modelling used to produce the final version of the IPA, and in order to have a fair opportunity of making out its case it needs access to that underlying material. The general description of the methodology and the summary of the results set out in the CMA's final report are not adequate to ensure that HCA's challenge would be disposed of fairly. Ms Rose submitted that HCA's application cannot be characterised as a speculative "fishing expedition". The information is not sought simply with a view to checking calculations in relation to which there is no particular reason put forward by HCA to suggest that they are flawed. On the contrary, HCA has identified particular features of the IPA set out in the final report which at the least give rise to significant questions regarding the data and methodology used to support what appears in that report.
29. Ms Smith QC, for the CMA, on the other hand, emphasised that the volume and nature of the material for which disclosure is sought in this case are utterly different from the materials at issue in Tweed. In particular, she objected to the request for disclosure of the raw data. Those data were not directly input into the IPA. The cleaned data had been used at that stage. Further, Ms Smith pointed out that HCA had been able to formulate its grounds of challenge by reference to the final report, which suggested that it did not really need to have access to all the underlying data and information now sought under its disclosure application.
30. In our judgment, this last point for the CMA involves the implicit acknowledgement - which is, in our opinion, warranted by the points of challenge to the IPA which have already been identified by HCA in its notice of application under section 179 and on the present application for disclosure - that HCA's disclosure application is not a mere speculative "fishing expedition". We were impressed by the explanations given by Roger Witcomb, who was Chair of the CMA Group which considered the private healthcare market investigation, in his witness statement for the CMA in meeting many of the points made for HCA in an expert report of Dr Mazzarotto, and in pointing out that HCA is already able to make the criticisms of the IPA it has indicated it wishes to make on the basis of the information set out in the CMA's final report. However, we do not consider that Mr Witcomb's explanations were sufficient to meet the thrust of HCA's case for disclosure. HCA has raised a serious case regarding data input and possible methodological flaws in the CMA's analysis which the CMA will be called upon to answer in the course of these proceedings. HCA is not seeking disclosure simply on the off-chance that it might throw up some hitherto unsuspected error of calculation by the CMA. It wishes to have disclosure to assist it to make good an arguable case which it has already set out and advanced.
31. In the circumstances of this case, we consider that fairness between the parties does require that HCA have access to the material covered by its disclosure application and that this is not outweighed by any of the countervailing practical issues we have reviewed above. The IPA was absolutely critical as the basis for the CMA's findings of relevant AECs and thus for its decision to impose the divestment remedy. Arguable grounds of attack on the IPA have already been identified by HCA and such an attack will inevitably be centre stage in these proceedings. The CMA may have good answers to that attack, and will of course submit that in making the inevitable evaluative judgments required at each stage in taking the raw data, refining them into the cleaned data then analysing them as it did, it is entitled to the benefit of a substantial margin of appreciation or evaluation. But in our view, this does not mean that in this case HCA should be practically disabled from making the best case it can by being deprived of information about the underlying data (including the raw data) and a clear picture of the methodology employed by the CMA. HCA should not have to meet such difficulties as it may face in showing that the CMA has acted unlawfully in refining the data and then in choosing to analyse it as it did while subject to the practical disadvantage that it does not actually know the facts regarding what the CMA has done to construct the IPA.
32. There are two further matters which support our conclusion that disclosure should be ordered. First, Ms Rose relied on the decision of the Court of Appeal in R (Eisai Ltd) v National Institute for Health and Clinical Excellence [2008] EWCA Civ 438, in which it found that fairness in the conduct of a consultation by the National Institute with industry on the calculations and analysis to be used as a foundation for a cost/benefit assessment when working out what pharmaceuticals could be recommended for purchase by the NHS required the National Institute to make available to pharmaceutical companies a fully executable (not merely read-only) version of the computer model it had constructed to use for this purpose, so that they could make effective representations on the key question of the robustness or reliability of the model (see [36]), in particular in relation to the sensitivity analysis affecting the model (see [44]). In that case, the Court recognised that the claimant had been given a good deal of information which had allowed it to make substantive comments, but found that nonetheless it had been unfairly hampered in making representations by not having available to it an executable version of the computer model (see [49]-[53]). Ms Rose submitted that in the present case, similarly, fairness required that the CMA should give the disclosure sought by HCA.
33. Against this, Ms Smith correctly pointed out that the point at issue in Eisai related to the requirements of fairness, as a substantive principle of public law, in the course of a consultation conducted by the National Institute. She submitted that the judgment in Eisai had no bearing on the issue before the Tribunal, which relates to the operation of the principles governing disclosure in the course of judicial review litigation.
34. We recognise that the public law principle of fairness or natural justice is distinct from the principle of fairness in the course of the conduct of litigation. The contexts in which they operate are significantly different. In the latter case, a court or tribunal has to strive to find a balance between competing considerations specific to the litigation context and reflected in the factors which bear upon the overriding objective, including the desirability of containing the cost of dispute resolution through hostile litigation.
35. However, in our view, the principles of fairness in the two situations also have some significant family resemblance, despite the difference in context. In the particular circumstances of this case, we consider that the resemblance between what was at stake in the Eisai case and the purposes for which HCA seeks disclosure in these proceedings is substantial. HCA wishes to be able to check the critical modelling conducted by the CMA, in particular with respect to the sensitivity analyses employed by it, just as the claimant wished to do in Eisai in relation to a similarly critical computer model. The Court of Appeal recognised in Eisai that the claimant would be hampered in making representations with full force and effect on that issue without access to the relevant computer model and held that this would be unfair. Similarly, we consider that HCA would be hampered in presenting its case with proper force and effect in these proceedings without having access to the underlying data and the modelling, and that this would be unfair to it in the context of this litigation. In our view, therefore, Eisai does lend material support to HCA's application for disclosure.
36. Secondly, proportionality analysis for the purposes of the Human Rights Act 1998 requires that the evidential basis adduced by a public authority to justify an interference with Convention rights should be "relevant and sufficient": see BAA Ltd v Competition Commission [2012] CAT 3, [20(5)]. The greater the interference with Convention rights, the more robust and reliable the evidential basis relied upon to justify that interference may be required to be. Critical examination of such evidential basis in the process of adversarial litigation may make a significant contribution to ensuring that the evidential basis is sufficiently robust and reliable to justify the measure taken on the basis of it. The Tribunal might find itself hampered in examining whether the serious interference with HCA's rights in issue in these proceedings is lawful and proportionate in circumstances where (as HCA contends, and which will be a matter for argument at trial) HCA had not had a full opportunity to address the evidential basis for the measures decided on by the CMA in the course of its investigation nor an opportunity to subject that evidential basis to full and informed critical scrutiny subsequently in the course of the litigation to challenge that interference.
37. In our opinion, there is force in the points made by Dr Mazzarotto in his witness statement, dated 7 July 2014, that it will be important in the context of the present case to be able to examine carefully the basis for the CMA's conclusion that the relationship between market share and prices is causal, and that issues regarding whether - in constructing the IPA - prices have been correctly measured and comparisons have been performed on a suitable like-for-like basis (even allowing for a significant margin of evaluative judgment on the part of the CMA at each stage in the process of investigation and analysis) are capable of being illuminated by HCA being given access to the raw data, the cleaned data and information about the methodology used by the CMA in its analysis.

Conclusion

38. For the reasons given above, HCA succeeds in its application for disclosure. Since there are significant practical issues which will need to be sorted out to give effect to the order for disclosure, we invite the parties to seek to agree a form of order. If there are areas of disagreement about the terms of the order to be made, they should be addressed by way of written representations and the Tribunal will resolve them on the basis of such representations.

The Honourable Mr Justice Sales (Chairman)	Dermot Glynn	Clare Potter
Charles Dhanowa O.B.E., Q.C. (Hon) (Registrar)	Date: 25 July 2014	Date: 25 July 2014