16 October 2014

A disapointing policy, far from the "full transparency" announced by the EMA in November 2012. Redacting to uphold commercial confidentiality gives companies the upper hand on the content of clinical reports that are made publicly available.

> Click here to download the joint statement (pdf, 689 Ko)

Joint Statement

BRUSSELS - On 2 October 2014, the European Medicines Agency (EMA) finally adopted its policy on access to clinical data (1). From 1 January 2015 onwards, the general public will be able to "view on screen" selected parts of trial reports submitted in support of marketing authorisations, and academic and non-commercial researchers will be allowed to download this data. The policy, unfortunately, also gives pharmaceutical companies the upper hand in deciding the contents of the clinical reports by allowing them to redact data on the grounds of "commercial confidentiality". The implementation of the policy will need careful monitoring.

For more than a decade, the EMA has failed to comply with a key measure of the European Freedom of Information Regulation (Regulation (EC) N°1049/2001), adopted in 2001: to set up a register of documents that it holds (2).This makes it very difficult for citizens to determine which document to request, leading to endless exchanges with the EMA before documentation is provided. Growing criticism regarding the EMA's failure to correctly respond to documentation requests, along with increasing concerns within the research community about clinical data being withheld by pharmaceutical companies, has made it unavoidable for the EMA to finally start a process towards more transparency (3).

In November 2012, the EMA announced that it would proactively publish clinical trial data and enable access to full data sets by interested parties by January 2014 (4). The idea, reflected in the EMA's June 2013 draft policy, was that interested parties would no longer need to invoke Regulation (EC) N°1049/2001, when exercising their fundamental right to access documents held by the EMA (3).

However, in May 2014, at the same time that negotiations of the bilateral trade agreement between the EU and the USA were taking place (5,6), the EMA shared documents that indicated a watering down of its 2013 draft policy on proactive access to clinical data by:

• Imposing wide "view-on-screen only" restrictions on the use of the data, even for researchers;
• Proposing restrictive "terms of use" conditions, forcing data users to enter into legal agreements with pharmaceutical companies and;
• Proposing "redaction principles", allowing censorship by pharmaceutical companies under the guise of protecting "commercial confidentiality" (7,8).

In June 2014, following a public outcry calling on the EMA not to backtrack on its previous commitments, the Agency announced the removal of the screen-only restrictions for academic and non-commercial researchers, but did not address other criticisms (8-10). By July 2014, the EMA agreed to postpone the adoption of its policy on access to documents until early October 2014, "in order to improve it further" (11).

Access to selected parts of clinical reports, no access to anonymised individual participant data

Public access to regulatory data is essential to minimise the effects of publication bias, a practice where "positive" results on a new drug are made available while "negative" results are being withheld. Unfortunately, the EMA has recently abandoned its plans, announced in November 2012, to routinely require pharmaceutical companies to submit all original clinical trial data in a format that would allow the EMA to re-analyse the data (12). This means that a medicine can still be approved by the EMA on the basis of incomplete evidence as it was the case, for instance, with the anti-viral drug oseltamivir (Tamiflu°) (13). Consequently, proactive access to all relevant clinical trial data will not be granted to researchers, nor to the public.

• Pierre Chirac, coordinator of the Medicines in Europe Forum, comments:
  "It seems that the EMA has found an easy solution to avoid having to release much clinical data: by just not requiring it from pharmaceutical companies in the first place…"

Only reports of trials submitted in support of centralised marketing authorisation procedures fall under the scope of the EMA's policy (1). No proactive access to clinical reports will be retrospectively granted for medicines currently marketed (1).

Clinical overviews, clinical summaries and selected parts of clinical study reports (including the protocol and protocol amendments, sample case report forms and documentation of statistical methods) – not the full clinical study reports – will be released (by 1 January, 2015 – with a one-year delay – for new medicines, and by 1 July, 2015, for new indications of already marketed medicines).

Since clinical study reports will not be published in full and information can be redacted (see below), there is a risk that "from now on clinical study reports will be written and structured in such a way as to withhold vital details of a pharmaceutical drug's effects or present them in the best possible light—mirroring the unreliability of trial results published in journals (in their potential to be affected by reporting bias)" (14).

Last but not least, the Agency decided to postpone the release of individual participant data (IPD) until further notice: "the Agency will not request applicants/MAHs to submit IPD for the sole purpose of publication of IPD. The Agency will first undertake a targeted public consultation (…) on the various aspects in relation to IPD to provide clarification" (1). The fact that the EMA has abandoned its plans to re-analyse clinical data however only reinforces the need to provide anonymised individual patient data to researchers, academics and health technology assessment (HTA) bodies so that they can do valuable reanalysis and secondary research.

"Terms of use": two levels of access and less legal threats for researchers

The new proactive policy establishes two levels of access (1):

• For general information use by the general public: a basic user profile with clinical data available after simple registration, but only view-on-screen mode, with prohibition to "download, save, edit, photograph, print, distribute or transfer the Clinical Reports";
• For academic and non-commercial research purposes: an upgraded user profile with downloadable clinical data. Users will have to identify themselves and to provide an address in the EU (a).

The EMA policy forces users to agree to the Agency's terms of use to access the data (1). Users are required to "acknowledge that the Clinical Reports are protected by copyright or other intellectual property rights of the Applicant/MAH and can be considered commercially valuable when used for commercial and regulatory purposes", despite the fact that the notion of clinical trial data being protected by proprietary rights is highly controversial on the grounds of health-related human rights (1,15).

Compared to the draft terms of use from May 2014, the statement, "when used for commercial and regulatory purposes", is welcomed, as is the deletion of the definition of the "information owner" referring to pharmaceutical companies (1). Another welcome change is that researchers are no longer required to sign confidentiality agreements (b)(1). Researchers performing secondary analyses are "encouraged" to provide the EMA with copies of the article prior to publication, but the Agency does not require control over how the data is used, contrary to the very restrictive approach taken in the EMA's draft revision of its policy on access to pharmacovigilance data (c)(16).

Redaction to uphold commercial confidentiality gives companies the upper hand and negotiating power behind closed doors

According to the EMA, its new policy is "designed to guard against (…) breaches of intellectual property rights that might disincentivise future investment in R&D" and "the Agency respects and will not divulge CCI [commercially confidential information]" (1). CCI is very broadly defined by the EMA as "any information contained in the clinical reports submitted to the Agency by the applicant/MAH that is not in the public domain or publicly available and where disclosure may undermine the legitimate economic interest of the applicant/MAH", and grants the pharmaceutical industry wide and far-reaching data protection (1,15).

Despite concerns raised by researchers that redaction can hinder the interpretation of data (e.g., in the interpretation of a serious harm narrative) or delay access to information of public interest (e.g., redaction of results on exploratory endpoints) (14,17), companies will be allowed to propose that "sections of the clinical reports, may be considered CCI and, therefore, may have to be redacted (…)" (d) (1, annex 3).

The final policy specifies that "the final decision will rest with the EMA" (1), but, at the same time, makes clear that "in case of disagreement with the Agency's final decision on the redaction, the applicant/MAH will be given a defined period prior to the publication to seek an interim injunction from the Court"… (18).

Moreover, despite the EMA's claim that this final policy is "without prejudice to Regulation (EC) No 1049/2001 [European Freedom of Information Regulation]"(1), the Agency explains that "there will be no difference in the understanding of CCI in the Agency's assessment of the documents held by the Agency that are requested through ‘access to documents' or that will be proactively published by the Agency" (18). The new policy will, therefore, undoubtedly have an impact on the application of this Regulation.

• According to Jörg Schaaber, President of the International Society of Drug Bulletins:
  "This is particularly worrisome. Since March 2013 when two pharmaceutical companies took the EMA to court to oppose data disclosure, some of our members experienced a regression in the EMA's response to their requests for documents. There were delays in response and data delivery, a shift in procedures and a change of tone."

Implementation: EMA under close watch

Any judgment on whether the EMA policy meets the needs of European citizens will only be possible once it is rolled out. For the time being, concerns remain, especially in regard to pharmaceutical companies' right to redact some data, a practice that allows them to control what data – which is regulatory data created for public interest use in order to protect consumers – is released (e,f).

We look forward to the implementation of the "policy's second phase" (access to individual patients' data). It remains to be seen whether the EMA will be able to uphold high standards of transparency, which the European Commission (particularly the Directorate General for Enterprise) actively opposed in the recently adopted Clinical Trials Regulation (19).

We also call on the EMA to improve its transparency practices by:

• Proactively providing a numbered, standardised table of contents for clinical study reports (including a list of appendices and attachments containing information on study design, conduct or results), so researchers can identify relevant additional information that they might wish to request using Regulation (EC) No 1049/2001;
• Finally setting up a register of documents that it holds, as requested by Regulation (EC) No 1049/2001, to make it easier for citizens to exercise their rights;
• Increasing transparency over pharmacovigilance data, which are crucial in protecting patients from avoidable harm (16).

Association Internationale de la Mutualité (AIM), Medicines in Europe Forum (MiEF), Health Action International (HAI) Europe, International Society of Drug Bulletins (ISDB), Nordic Cochrane Centre (NCC)

©AIM, MiEF, HAI Europe, ISDB, NCC October 2014

Notes:
a - Users will have to give their "name, date of birth, passport or ID card number, expiry date of the document; for juridical persons, the affiliation and position within the organisation" (ref. 1). And the EMA specifies that "it should be noted that Courts may require the EMA to disclose the identity of the users who do not comply with the terms of use to the marketing authorisation holders (MAHs)/applicants" (ref. 18).
b - These confidentiality agreements, foreseen in the May 2014 draft terms of use, would have made it particularly easy for pharmaceutical companies to challenge researchers in court simply for violation of the EMA’s policy on terms of use (ref. 8).
c -  In the revision of its 2011 policy on the access to the European pharmacovigilance database, EudraVigilance, the EMA sets up restrictive conditions for granting access to researchers (e.g., the signature of confidentiality agreements). The EMA also demands to "view any publication resulting from EudraVigilance data before submission (…) [and that] any issues raised by the Agency (…) must be addressed to the satisfaction of the Agency before submission for publication" (ref. 16). 
d - The EMA’s policy even gives incentives to companies to redact data up front by classifying a number of sections as "may be commercially confidential information", and by providing "ready-to-use" justifications for redaction (refs. 1,20). 
e - Clinical study reports (CSRs) are not third-party documents owned by the marketing authorisation holder. Even if prepared by the marketing authorisation holder, using data collected from clinical trial participants, patients and healthcare professionals, CSRs comply with regulatory requirements as to content and format and are a key component of marketing authorisation procedures. These data are, in essence, regulatory data, created for public interest use. When clinical study reports are received at the EMA, they become a "document held by the Agency" and Regulation (EC) N°1049/2001 applies.
f - The EU Ombudsman has already announced that she will monitor how the EMA deals with access to documents requests, and that she will verify whether the redaction of information by the EMA and the terms to which users are required to sign up before gaining access to documents are justified (ref. 9).
 
References:
1- EMA "European Medicines Agency policy on publication of clinical data for medicinal products for human use" POLICY/0070 (EMA/240810/2013); 2 October 2014: 22 pages.
2- "Regulation (EC) No 1049/2001 of the European Parliament and of the Council of 30 May 2001 regarding public access to European Parliament, Council and Commission documents" Official Journal of the European Communities 31 May 2001: L 145/43-L 145/48.
3- AIM, HAI Europe, ISDB, MiEF "EMA's 2013 policy on access to clinical-trial data: Transparency in the public health interest" Joint Submission of comments on Policy 0070 on publication and access to clinical-trial data; September 2013: 15 pages.
4- EMA "Access to clinical-trial data and transparency - Workshop report" 19 December 2012: 6 pages.
5- The U.S. Chamber of Commerce – Global intellectual property center "Heading in a different direction? The European Medicines Agency's Policy on the Public Release of Clinical Trials Data"2014 www.theglobalipcenter.com: 33 pages.
6- Murray J "Clearing the Air on Transparency Decisions, and on Industry Demands to the Juncker Commission" openmedicineeu.blogactiv.eu; 1 October 2014.
7- AIM, HAI, ISDB, MiEF, NCC "Backpedalling on EMA's "proactive publication of clinical-data" draft policy: Was it all just a window-dressing exercise? Who or what is the EMA afraid of?" Joint Press Release; 20 May 2014: 4 pages. > Pdf
8- AIM, HAI, ISDB, MiEF, NCC "EMA's new policy on access to clinical data: About to privatise pharmaceutical knowledge? The proof will be in the pudding" Joint Press Release; 24 June 2014: 4 pages. > Pdf
9- "European Ombudsman reaction to EMA's 12 June 2014 statement issued after its Management Board meeting" www.ombudsman.europa.eu 13 June 2014: 1 page.
10- Doshi P "EMA policy on transparency is "strikingly" similar to deal struck with drug company, say experts" BMJ 2014;348:g3852.
11- EMA "Management Board delays formal adoption of EMA publication of clinical trial data policy to October 2014" press release; 9 July 2014: 2 pages.
12- Sharma V "EMA dilutes ambitious data re-analysis plans" www.scripintelligence.com 12 September 2014: 2 pages.
13- Goldacre B "What the Tamiflu saga tells us about drug trials and big pharma" The Guardian 10 April 2014.
14- Jefferson T "EMA's release of regulatory data—trust but verify" blogs.bmj.com; 7 October 2014.
15- Lemmens T "EMA's Proposed Data Release Policy: Promoting Transparency or Expanding Pharma Control over Data?" Plos Blog 30 May 2014: 3 pages.
16- Cochrane AEMG, HAI Europe, ISDB, MiEF "EMA's policy on pharmacovigilance: access to qualitative data is needed, pharmacovigilance data are not "trade secrets"" Paris, 15 September 2014: 17 pages.> Pdf
17- Silverman E "EMA Remains Under Fire for its Policy on Disclosing Clinical Trial Data" blogs.wsj.com; 3 October 2014.
18- EMA "Questions and answers on the European Medicines Agency policy on publication of clinical data for medicinal products for human use (EMA/357536/2014 Corr1)" 2 October 2014: 10 pages.
19- Willmott G, MEP "Juncker should reverse decision on medicines shake up" 11 September 2014: 1 page.
20- Santos Quintano A. "EMA policy on clinical data publication a step ahead, but falls short of true potential" 15 October 2014 http://haistaffblog.blogspot.nl/2014/10/ema-policy-on-clinical-data-publication.html.

> Click here to download the joint statement (pdf, 689 Ko)