How Health Canada and the FDA release clinical information
Health Canada and the US Food and Drug Administration each release a sponsor's clinical documents in their own way, beside EMA Policy 0070. Health Canada's Public Release of Clinical Information matches the European policy in much of its scope, anonymisation and redaction and departs from it in places. The FDA piloted the publication of parts of clinical study reports rather than publishing full reports, and the three regulators held completely different positions on disclosure.
Health Canada's Public Release of Clinical Information
Regulators around the world have introduced measures to promote transparency, and two of them, the EMA and Health Canada, have gone further by mandating the anonymisation and publication of clinical study reports.
Health Canada's draft guidance on the public release of clinical information, supporting proposed changes to the Food and Drug Regulations and the Medical Device Regulations, was open for public consultation until 25 June 2018.
Health Canada concluded the consultation and issued a final regulation with final guidance on 12 March 2019. The guidance is broadly aligned with the EMA's Policy 0070 guidance, with the intention of streamlining sponsor effort. Health Canada's Public Release of Clinical Information, introduced in 2019, mirrors the EMA's requirements, with disclosure of clinical study reports required for market authorisation. It makes anonymised clinical data from drug submissions and medical device applications available to the public, with mandatory disclosure for new marketing approval submissions whose final regulatory decision was made after March 2019, and disclosure on request for those decided before.
What Health Canada releases
For drugs, Health Canada releases the clinical overview (Module 2.5), the clinical summaries (Modules 2.7.1 to 2.7.4) and the clinical study reports, a report being taken as a single document with its protocol, sample case report forms, investigator-related information, information on the test drugs or investigational products, technical statistical documentation, related publications, patient data listings, and technical statistical details such as derivations, computations, analyses and computer output. For devices, it releases summaries and detailed information on all clinical studies and investigational testing that provided evidence of safety. Individual patient records are not publicly released with the other clinical information.
Health Canada's publication process and timelines
The publication process has five phases: initiation, submission, review, finalisation and publication. Health Canada targets uploading the redacted and anonymised package within 120 calendar days of process initiation. A positive regulatory decision triggers the publication process; a negative decision triggers publication 31 calendar days after the date of the notice, which may be halted if a letter of intent for reconsideration is submitted. Day by day, the process runs as a process initiation meeting with Health Canada on days 1 to 20, the applicant's submission of the redaction proposal package on days 21 to 60, Health Canada's review of that package on days 61 to 90, the applicant's response to Health Canada's comments on the anonymisation report and confidential business information on days 91 to 115, and the submission and publication of the final redacted document package on days 116 to 120. There are no financial penalties for missing the timelines.
Where Health Canada's programme departs from Policy 0070
Health Canada's programme differs from Policy 0070 in several ways:
- It applies to device trials as well as to drug and biologic trials. It publishes the packages for device applications, which the EU programme does not.
- In addition to proactive submissions for market authorisation, it publishes historical studies in response to access-to-information requests from the public.
- It has not announced plans to enforce disclosure of individual participant-level data, whereas the EMA has indicated its intent for a second phase encompassing participant-level data.
Accounts differ on which documents are in scope. One names abnormal laboratory value listings as a minor variance, in scope for Health Canada's programme but not for Policy 0070. The other gives both the documents in scope and the need to redact personal and confidential business information as the same as under Policy 0070.
Anonymisation and redaction at Health Canada
Health Canada's guidance asks manufacturers to anonymise the clinical information with a risk-based statistical anonymisation process closely aligned to the EMA's guidance, and, like the EMA, recommends a threshold of 0.09.
Health Canada has been explicit in its preference for a quantitative approach to measuring re-identification risk, but during the early period of adoption it accepted submissions in which a non-analytical or qualitative approach was taken, as well as those in which redaction was applied. There is a strong indication that it is encouraging a move away from reliance on these methods: certain heavily redacted submissions were published with a notice that their extensive redactions to patient information do not conform to Health Canada's guidance, which encourages manufacturers to retain the analytical value of information by using other transformation methods, such as generalisation or randomisation, and to apply them to the specific information that risks re-identifying an individual rather than to redact broad sections of information. The notice states that Health Canada encourages manufacturers to anonymise personal information according to the principles of its guidance document on the public release of clinical information, that it will continue to explore ways to help ensure all publications include anonymised personal clinical information, and that inquiries for access to the redacted information may be submitted to its Information Science and Openness Division.
How Health Canada and the EMA redact, and what they agree to redact, is very similar.
Reusing packages approved under Policy 0070
Health Canada has indicated its intent to accept packages the EMA has already approved, to avoid duplicated effort. Packages approved by the EMA under Policy 0070 can later be used for Health Canada submissions. Redacted documents previously accepted under Policy 0070 may be submitted, and the EMA's specifications for redacted text may be used in finalised documents. The EMA was increasing its collaboration with Health Canada, and although the two have slightly different requirements in the anonymisation report for how redacted data are redacted and labelled, both types of redaction are accepted, so the redactions need not be changed to one agency's conventions for a package to be submitted to it.
Health Canada's Clinical Information Portal
Clinical information on drugs and medical devices is publicly available through Health Canada's Clinical Information Portal. The portal held clinical information on a few products from volunteer organisations in advance of the regulation's 12 March 2019 in-force date; drug submission data were to be posted afterwards as each review was completed, proactive disclosure of device data was to start in 2021 in line with the EMA's intended timetable, and drug and device data already available on request were to be publicly posted.
The portal is very similar to the EMA's platform, and no account is needed to use it. One user of both holds that its documents are better organised: it seems to collate them into phase 1, phase 2 or phase 3 studies, so they can be accessed much more efficiently, whereas on the EMA's platform the studies do not seem to be in any particular order.
The FDA's Clinical Data Summary Pilot
The FDA announced plans to publish clinical study reports in a pilot scheme to evaluate whether disclosing certain information included in them after approval of a new drug application improves public access to drug approval information. The plans indicated that the report body, the protocol and the statistical analysis plan would be shared, and that portions of the reports would be publicly posted once the pilot concluded.
In this Clinical Data Summary Pilot Programme, the FDA makes the redactions itself. The first published clinical study report, the pivotal study report for Erleada, left the names of sponsor staff, including the names of medical writers, unredacted, while redacting the details for local sponsors, selected formulation and other product-related information, sensitive information in patient narratives, site identifier, subject identifier and date of death from death listings, and site identifier, subject identifier and assessment date from abnormal laboratory and other similar listings.
On 26 June 2019 the FDA concluded the recruitment phase of the pilot, in which one sponsoring pharmaceutical company voluntarily participated. The FDA published a Federal Register notice, under Docket No. FDA-2019-N-2012 for New Drugs Regulatory Program Modernisation: Improving Approval Package Documentation and Communication, seeking feedback on the pilot through stakeholder questions on the potential benefits or risks, resource requirements and challenges of publicly releasing a limited number of sections from certain clinical study reports at the time of marketing approval, and the comment period closed on 26 August 2019. The same notice sought public comment on a new integrated template that the FDA will use to document its review of new drug applications and efficacy supplements. The pilot and the feedback the FDA sought on it opened a potentially new pathway for the public disclosure of clinical information, an alternative to the publication of clinical study reports and clinical summary documents seen until then from the EMA and Health Canada.
Where the three regulators stood
Activities around the disclosure of clinical documents were slow after September 2018, and the three main regulators in the ICH family contributing to the disclosure narrative held completely different positions. The EMA kept its clinical data publication activities on hold, with no indication of when they might resume or whether public disclosure would then apply retrospectively to documents. Health Canada was actively disclosing clinical documents, with guidance broadly similar to the EMA's but with its regulators discouraging redaction in favour of proactive authoring (qualitative anonymisation) and, ultimately, quantitative anonymisation methods. The FDA was considering its options. It had solicited opinion on how it might best support disclosure of clinical documents and announced the conclusion of its Clinical Data Summary Pilot in March 2020. It was not disclosing clinical documents but had identified a possible approach for disclosing study reports. The framework of that approach includes these principles:
- A centralised international library managed by an independent body would make information available to the public, rather than each regulatory authority having its own system.
- An on-demand system would be set up in which some documents, such as clinical summaries and an index of study reports, would be published automatically, and the public could request documents that sponsors would add to the library.
- Anonymisation and disclosure standards would apply, with the PHUSE standards particularly mentioned.
- Sponsor commitment to use the international library system would be voluntary.