The Select Committee on Health´s Report on NICE – Confidentiality Concerns

The House of Commons Select Committee on Health's recent Report on the National Institute for Clinical Excellence ("NICE")¹ contains numerous recommendations to improve the functioning and independence of NICE, many of which have been praised by the pharmaceutical industry, physicians and patient groups alike.

However, some of its recommendations are likely to be of concern to the industry, not least of which the recommendation that "all information which NICE uses in its decision-making process [be] made available for public scrutiny".² The related reasons for this recommendation appear to be:

• the need for clarity and transparency as to the basis on which appraisal decisions are made; and
• the public credibility of NICE, particularly with clinicians, the wider NHS and patient groups.

The Select Committee states that "if industry or others have previously unpublished data which they want to use to support their case then this should no longer be presented to NICE subject to confidentiality". Moreover, if "pharmaceutical companies insist that unpublished data relating to the indication in question remain confidential, this may mean NICE will be forced to withdraw that treatment from the appraisal process". Whilst this is said to carry the risk that guidance on a treatment will be delayed until published information is available, the Committee believes that pharmaceutical companies are unlikely to use this as an "opt-out" clause because, it is said, their refusal to submit their evidence to public scrutiny would be likely to have a negative impact on commissioners' and clinicians' confidence in their product.

This article addresses some of the issues surrounding the recommended prohibition on the confidentiality of data used by NICE in its decision-making processes. It maps the consequences that such a prohibition would have on existing legal principle and administrative and judicial practice, evaluates the Select Committee's reasons for its recommendation in the light of the potential practical ramifications for industry stakeholders, and suggests that there is insufficient justification for such a prohibition.

The most obvious consequence of the prohibition is that it would strip stakeholders of an existing right to confidentiality in respect of information which is commercially sensitive and could otherwise be categorised as confidential. Whilst under English law a sufficient public interest in disclosure may be a defence to what would otherwise be a breach of confidence,³ that is a case-specific question, not a principle that can be applied universally to all information submitted to a particular authority for a particular purpose so as to justify the release of such information to the world at large.

It is true that in Smith Kline & French Laboratories Ltd v The Licensing Authority⁴ the House of Lords held that the Licensing Authority was entitled to cross-refer to Smith Kline & French's ("SK&F") data when a generics manufacturer sought a marketing authorisation and claimed that essential similarity could be demonstrated by reference to the research and testing details that had been supplied earlier by SK&F. In other words, the Medicines Control Agency ("MCA") could refer to the data for a purpose other than that for which it had originally been supplied. However, the issue in that case concerned use of the data by the Licensing Authority for the performance of its statutory functions and duties, not release to the public at large. Indeed, Lord Templeman stated that whilst it was "not unconscionable for the licensing authority to make use of that information in the public interest for the purposes of the Act, … it would be unconscionable for the licensing authority to disclose that information to third parties for other purposes."⁵ Arguably, and as addressed further below, the recommended blanket confidentiality prohibition is not necessary for NICE to perform its functions and pursue its purposes.

The suggested confidentiality prohibition would also take us beyond established administrative law principle as to what types of information consultees are entitled to receive during consultation processes. There is no general rule that one party to a consultation should be given all material submitted by another and no general rule that confidential information may be disclosed as a requirement of consultation processes. Whether such information may and should be disclosed and, if so, how much information should be disclosed, is a case-specific question. Sometimes it may be possible to disclose the gist of information supplied by another, without revealing confidential information, whilst still enabling other consultees sufficiently to comprehend and respond to the matter.

When material not otherwise available to a consultee is of pivotal importance to the decision-maker’s final decision, a decision-maker may take the view that it may properly be disclosed to the consultee, even if some of that information was provided by a competitor and is therefore or might otherwise be considered confidential. However, in the author's view, application of this proposition must depend on the particular industry in question, the subject-matter of the particular consultation and the nature of the particular information considered confidential. In other words, it must be case-specific.

Regulatory and parliamentary authorities appreciate and respect the need for confidentiality in respect of commercially sensitive pharmaceutical data. For example, the MCA's guide on its operation of the Code of Practice on Access to Government Information, entitled Access to Information: A Guide for the Public (October 2000), states that it is unlikely to release information where disclosure "would harm the competitive position of a third party" (similar statements appear in NICE Guidance). There appears to be little doubt that the MCA is minded not to release to the public at large pharmaceutical regulatory data and other information generated in the course of dealings with individual companies where that information is stated to be or is otherwise considered confidential or commercially sensitive. This is certainly the position the MCA has taken in the past, by reference to confidentiality and other exemptions in the Code. For example, in response to one of Social Audit's requests for information regarding certain anti-depressant medicines, the MCA wrote, on 18 December 1997, that it could not disclose the information sought because "disclosure may harm the competitive position of a third party and prejudice the future supply of confidential information, and there is no current risk to public health or safety sufficient to override such interests".⁶

The Parliamentary Ombudsman, when considering complaints against non-disclosure of information by the MCA, has also recognised that in certain situations commercially sensitive information may properly be withheld.⁷

Whilst admittedly the MCA and the Ombudsman have been applying the Code of Practice, which contains exemptions for disclosure on grounds of confidentiality and commercial sensitivity, it was open to them to hold that there was an overriding public interest justifying disclosure. Neither the expert regulator nor the parliamentary watchdog, in the cases before them, has so held. Why should data held by NICE be treated differently and without regard to the facts of individual cases, particularly when NICE is concerned more with health economics than with public safety?

Courts have likewise appreciated the importance to the pharmaceutical industry of confidentiality. For example, in Pfizer Pty Ltd v Birkett,⁸ a case concerning the subsidisation of Viagra, the trial judge in the Federal Court of Australia was "prepared to accept that the estimates of the likely usage of alprostadil, as contained in the [earlier] manufacturer's submission to the [Pharmaceutical Benefits Advisory Committee], was confidential material which the [PBAC] was not at liberty to release to Pfizer." A Full Court of the Federal Court agreed with the trial judge on this issue.⁹ The Court said that the "PBAC will, presumably, often be in possession of confidential information of drug companies and we see no intention disclosed by … the Act that the confidentiality of such information should not be respected". These statements are consistent with the words of Lord Templemen (above).

The Select Committee's recommendation regarding the submission of data in confidence may also be said to cut across the scheme enacted by Parliament in the Freedom of Information Act 2000. That Act, the culmination of years of debate, contains exemptions to disclosure on grounds of breach of confidence, disclosure of trade secrets and prejudice to commercial interests, subject to an overriding public interest in disclosure (which arguably applies even to the so-called "absolute" exemption where disclosure would constitute a breach of confidence, because there is a public interest defence to breach of confidence). The issue of overriding public interest is a case-specific question, not one that can be answered universally without regard to the circumstances of individual cases.

The draft of a Code of Practice to be issued to public authorities by the Secretary of State under the Act states that public authorities should not agree to hold information "in confidence" which is not in fact confidential in nature, that public authorities should not accept confidentiality clauses where this is commercially viable, that any acceptance of such confidentiality provisions must be for good reasons and capable of being justified to the Information Commissioner, and that a public authority should only accept information from third parties in confidence if it is necessary to obtain that information in connection with the exercise of any of the authority’s functions. However, the draft Code does not go that extra step of stating that authorities may impose blanket prohibitions on confidentiality and be justified in refusing to deal with stakeholders if stakeholders wish the confidentiality of their information to be protected.

As a Special Health Authority, NICE will be subject to requests for information under the Act once the individual right of access to information becomes operative in January 2005. In the Act Parliament has sought to balance the competing interests in disclosure, versus retention, of information held by public authorities. Does the Select Committee mean to suggest that either in the interim (e.g., under the Code of Practice on Access to Government Information), or generally, data held by NICE should be treated differently or that stakeholders dealing with NICE should be stripped of protections that the Act would otherwise confer? That would appear to be a consequence of what is recommended.

The Select Committee notes that the MCA, working with the Committee on Safety of Medicines ("CSM"), already conducts detailed assessments of the quality, safety and efficacy of new medicines as part of the licensing process, and that the preliminary summary and committee judgements produced are likely to prove a valuable analytical resource for NICE. Accordingly, it says, "improved regulation of submission of information to NICE should be supplemented by closer working relationships between the MCA and NICE, including the sharing of appropriate summary information prepared for the CSM, in order to prevent duplication and strengthen the quality of NICE's outputs".¹⁰

MCA-derived corporate information would seem to fall within the "all information which NICE uses in its decision-making process" which it is recommended be "made available for public scrutiny". Conceivably that could include at least portions of product dossiers and confidential communications with the MCA and CSM. If so, unforeseen side-effects could include the stymieing of free and frank discussion between manufacturers, the MCA and the CSM, and the infringement of at least the spirit of European Community data exclusivity provisions,¹¹ in that although a generic manufacturer might not be able to rely on an originator's data, it may well obtain an unfair advantage through having early access to it.

A blanket prohibition on confidentiality of data submitted to NICE to enable publication of that data by NICE may also contravene the Human Rights Act 1998 by breaching Article 1 of the First Protocol to the European Convention on Human Rights. Article 1, entitled "protection of property", states that every natural or legal person is entitled to the peaceful enjoyment of his possessions and that no one shall be deprived of his possessions except in the public interest and subject to the conditions provided for by law and by the general principles of international law. This protection does not, however, impair the right of a State to enforce such laws as it deems necessary to control the use of property in accordance with the general interest.

Under the Convention, "possessions" has been interpreted broadly. The European Court of Human Rights has held that "the concept of 'possessions' … has an autonomous meaning which is not limited to ownership of physical goods and is independent from the formal classification in domestic law: certain other rights and interests constituting assets can also be regarded as 'property rights', and thus as 'possessions' for the purposes of this provision".¹² Although novel territory, arguably a right to confidentiality and the interest in keeping commercially valuable know-how away from competitive eyes are assets which can be regarded as "property rights" under the Convention (even if not under domestic law).¹³ Whether an interference with those rights is justifiable depends on the nature of the interference and the grounds for it.

Do the Select Committee's reasons for the recommended blanket confidentiality prohibition – transparency in decision-making and the public credibility of NICE – justify the elimination of stakeholders' right to clothe commercially sensitive information submitted to NICE with confidentiality? Do they justify a departure from existing private and administrative law principle, existing regulatory practice and appreciation of confidentiality and the regime of the Freedom of Information Act? And do they justify what arguably is an infringement of the Convention right to protection of property? In the author's view, they do not.

Is a blanket prohibition on confidentiality necessary for transparency in decision-making? Arguably not. An appraisal decision does not turn on every single item of information submitted to NICE and on every item of information which NICE may (in future) obtain from other regulatory bodies such as the MCA and CSM. Where a particular segment of confidential information is pivotal to an appraisal, NICE may, on existing principle, feel and be justified in disclosing the data to appropriate recipients, perhaps with confidentiality undertakings, on public interest grounds (which can be a defence to breach of confidence). But transparency does not require release of the mass of information which stakeholders may be required to disclose to NICE. Moreover, given its potential volume for some appraisals, the release of all data submitted by all stakeholders might obscure rather than clarify the basis for decisions. Arguably a distinction needs to be drawn between "information NICE uses in its decision-making process" (the Select Committee's wording) and "information on which NICE's appraisal decisions are based". NICE is obliged to "use" all information submitted to it, in the sense that it is obliged to take the information into account, but it is not obliged to (and indeed cannot) base its decisions on all such information. Some data may be more relevant than others and in some cases NICE will have to choose between conflicting data.

Arguably the question of whether confidential data should be released in any given case should be focused on the subset of such data, if any, on which an appraisal decision is based. And even then arguably NICE should ask whether it can reach a compromise with the stakeholders concerned and/or whether the interests of transparency and informed consultation can be served through disclosure of the gist rather than the totality of the information. Transparency in decision-making is a laudable goal but where, in a given case, it is sought in respect of confidential data, there are fairer and more proportionate ways of achieving it.

Is a blanket prohibition on confidentiality necessary for the public credibility of NICE? To the extent that public credibility is a product of transparency in decision-making then, for the reasons above, again the answer would appear to be no. To the extent that public credibility is not a product of transparency in decision-making, arguably the confidentiality of stakeholder data is largely if not entirely irrelevant. The Select Committee has identified other factors affecting the public credibility of NICE, including clinical credibility stemming from the technical quality of NICE's guidance, the inclusiveness of consultees, transparency in NICE's processes, and independence. These factors do not necessarily, if at all, depend upon an absence of confidentiality around commercially sensitive data, and a blanket prohibition on confidentiality does not, and in some instances cannot, ensure they are met.

The public credibility in question is not that of stakeholders but of NICE which, the Committee notes, was tarnished by, among other things, its change of guidance regarding the flu drug Relenza. Bodies like NICE, particularly in their infancy, will never have a sparkling public image because they are charged with making difficult resource allocation decisions. No one denies that. But removing confidentiality protections from those who have to submit commercially sensitive data to NICE does not appear to be either an effective or just solution.

One also needs to consider the possible ramifications which the suggested blanket confidentiality prohibition could have for industry stakeholders. It is conceivable that the reception of confidential data by competitors or those ill-equipped to interpret it properly could do significant economic damage to those originators who have no practical choice but to provide information to NICE when their products or technologies are targeted for appraisal. If provided before an appraisal is finished, strategic, ill-informed or otherwise unfair criticism of data may be voiced and echo through print and electronic media to the detriment of a company's share value. This may be more likely if the Select Committee's recommendation, "that for all new technologies, NICE's work programme [be] arranged to facilitate publication of guidance at the time of launch", is implemented, because at that stage data on clinical and cost-effectiveness may be limited and coincidentally skewed (either for or against a product) through insufficient "real life" exposure. Marketing strategies could be affected. The uptake of NICE's final guidance, and its public credibility, might be diminished. A product's reputation might be unfairly affected and in extreme cases the product itself may even be withdrawn. Alternatively, competitors' access to commercially valuable data could give them significant unfair advantage. The issue is not one of hiding data from the public but of recognising that there are legitimate competing interests at stake which need to be balanced rather than severed.

Further aspects of the Select Committee's recommendations on this issue invite comment. The suggestion that a stakeholder's insistence on the confidentiality of unpublished data may force NICE to withdraw the treatment from the appraisal process is somewhat surprising. Confidentiality protection over certain submitted data does not disable NICE from performing its functions. This is not a case like Smith Kline & French Laboratories Ltd v The Licensing Authority (discussed above). NICE's appraisal teams should be sufficiently qualified to critique data put before them without the need for dissemination to and scrutiny from the world at large, particularly if the Select Committee's recommendation that NICE be more involved with other expert regulatory and professional bodies is taken up fully. And as already noted, if, in a given case, specific confidential data is of a pivotal nature, NICE may, on existing principle, feel and be justified in disclosing the data to appropriate recipients. On occasion NICE may have difficult decisions to make in this regard, but that is simply a necessary consequence of the power it wields. In the author's view, there would be no warrant for withdrawing a treatment from an appraisal, which after all will have been requested by the Department of Health, merely because upon submission of data a stakeholder clothes commercially sensitive information with confidentiality or subsequently insists that certain data remain confidential. Indeed, for some treatments targeted for appraisal (e.g., existing and popular yet older treatments or high cost products for which there is a cheaper but similarly efficacious equivalent), the effect of such a withdrawal may be that existing prescription practices will not significantly change, which may be inconsistent with the goals of optimising clinical- and cost-effectiveness.

Finally, it is difficult to reconcile the subtle suggestion that pharmaceutical companies may "opt out" by insistence on confidentiality with the fact that they do not choose to have their products appraised. If they do not submit the requisite data they face the prospect of their products being adversely appraised or, exceptionally (and controversially), blacklisted by the Department of Health. The issue here is not one of supply of information to NICE (that is the subject of a separate recommendation by the Select Committee), but of the reasonable protection of commercially sensitive data.

In the author's view the Select Committee's stated reasons of transparency and public credibility do not justify the blanket removal of confidentiality protection. It is possible that the Committee had something more limited in mind, for example the disclosure of pivotal unpublished cost-effectiveness studies or a narrow and qualified conception of "public scrutiny", but if that is the case, it needs to be made clear. In any event, arguably the melting pot of interests is more appropriately served by considering the issue of release on a case-by-case basis within the rubric of existing private and administrative law principle.

A blanket prohibition by NICE on stakeholders requesting confidentiality over data that is truly confidential may constitute an abuse or power, disproportionality, unreasonableness or breach of the Human Rights Act and hence be subject to invalidation by the courts. Statutory intervention would probably be required for such a prohibition and that is a step which, it is submitted, ought not to be taken lightly.

© reserved.

The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.