Posting the protocol and statistical analysis plan on ClinicalTrials.gov
Under the US final rule, an applicable clinical trial whose results must be posted on ClinicalTrials.gov must also have its protocol and statistical analysis plan posted there. The requirement raises the questions of which trials it reaches, when the documents are due, what may be redacted and who decides, and how the documents are prepared for submission.
Which trials and which documents
Under the US final rule on Clinical Trials Registration and Results Information Submission, the responsible party for every applicable clinical trial whose results must be submitted, other than a pediatric postmarket surveillance of a device product that is not a clinical trial, must provide a copy of the protocol, and of the statistical analysis plan if the protocol does not include it. The copy must include all amendments approved by a human subjects protection review board, where applicable, before the time of submission that apply to all clinical trial facility locations. For both documents, at least the most recent version, after the latest global amendment, needs to be posted. The requirement affects all applicable clinical trials with a primary completion date on or after 18 January 2017.
Rationale and impact
The proposed rule had not required publication of the full protocol and statistical analysis plan, but it invited comments on the benefits and burdens of such a potential requirement. After assessing the comments, the US Department of Health and Human Services concluded that the benefits of making the documents public would clearly outweigh the burdens on responsible parties. The main advantages given are that the documents improve transparency and the quality of reporting, are necessary for a full understanding of a study's results and their replication, safeguard against reporting bias, facilitate meta-analyses, improve the design of future studies, reduce unnecessary duplication of studies, promote standardisation of protocol elements, and avoid multiple individual requests for the documents.
Following the rule's implementation, ClinicalTrials.gov had by 2019 become the most comprehensive source of original protocols and statistical analysis plans for recent studies. On 3 March 2019 it held more than 3,500 records of interventional studies with a protocol or plan publicly available, and more than 93% of these studies had a primary completion date on or after 18 January 2017, the rule's effective date, which shows the large impact the rule had already had. Many protocols and plans, once considered confidential, were by then also public on other platforms: the EMA's Policy 0070 Clinical Data website, the websites of some journals that follow the Recommendations of the International Committee of Medical Journal Editors, and the clinical trial websites of a number of sponsors.
When the documents are due
By default, the protocol and statistical analysis plan must be made public at the same time as the results, within twelve months of the primary completion date. In certain cases, results posting, and with it the publication of the documents, may be delayed for up to two years: if the product was not yet initially approved by the FDA when the trial reached its primary completion date, or if a new use of the product, such as a new indication, has been filed with the FDA or is planned to be filed within one year. In exceptional cases, an extension of the deadline can be requested for good cause. Because the protocol and statistical analysis plan are now part of the mandatory registration and results information, a responsible party that fails to submit them is exposed to the civil money penalties the FDA can seek for that failure.
Redaction
The rule states what may be redacted, but in 2019 guidance on the extent and format of redaction was scarce at best. Before submission, the responsible party may redact names, addresses and other personally identifiable information, and any trade secret or confidential commercial information as the Freedom of Information Act and the Trade Secrets Act define those terms, contained in the protocol or statistical analysis plan, unless that information is otherwise required to be submitted. The responsible party, not the FDA, decides on the redactions and makes them. Essential details necessary to understand the results must not be redacted. Personally identifiable information about individual participants is not expected in the documents, but if present it should be redacted. The FDA reserves the right to issue more specific guidance on redaction later and to challenge a responsible party whose redactions appear inappropriate.
A cursory review, reported in 2019, of a few randomly selected studies conducted by 20 mid-sized and large biopharmaceutical companies found the extent and format of redactions quite variable. Some documents had no or almost no redactions, while others had full paragraphs or occasionally full sections redacted. Some redactions followed the Policy 0070 style, and others were simple black bars without overlay text. Overall, some common principles emerged: redaction of the names and addresses of certain sponsor and vendor personnel, and a tendency to redact exploratory endpoints and related analysis methods.
Before redacting, a responsible party has many questions to answer. The decisions are company-specific and affect, for example, the consistency of redactions across public platforms and the effort needed to prepare redacted documents. The questions include these:
- Should personally identifiable information be redacted the same way as for other transparency channels, such as EMA Policy 0070?
- How much should be redacted as commercially confidential? Product development is usually at an earlier stage when documents are published on ClinicalTrials.gov than when they are published under Policy 0070, so more information may need to be considered commercially confidential.
- Should copyrighted content, such as questionnaires or scales, be redacted? Unlike the Policy 0070 Clinical Data website, ClinicalTrials.gov needs no login or acceptance of terms of use to view or download documents, so the responsible party has no control over what a user does with them.
- Which style and format of redaction should be applied?
- Should redacted documents be replaced by later versions with fewer redactions once those become available on other platforms? The rule does not require protocols and plans to be updated except for a protocol amendment.
Further questions, such as when to prepare the redacted documents, which functions to involve, how to decide on redactions and which software to use, are largely independent of ClinicalTrials.gov.
Cover page and file format
The cover page of each document must state the NCT number, the ClinicalTrials.gov identifier, if it is available, together with the official study title and the date of the document. A protocol is normally finalised before the NCT number is assigned, so the original protocol typically lacks it, and an extra cover page may be added or the number inserted on the title page of the redacted version. The documents must be submitted in a common electronic document format, which the Protocol Registration and Results System specifies as PDF/A.