Revising ICH E3 and E6
ICH E3, the guideline for the clinical study report, and ICH E6, which covers the protocol, conflict in places, and practice moved on in the two decades after they were issued. Both therefore came up for review, which a working group began in 2014. A further question is whether the report will get a structured template of its own, as the protocol did with ICH M11.
Gaps between and within the two guidelines
ICH E3, which gives the authoring guidance for clinical study reports, and ICH E6, whose Section 6 describes the topics for inclusion in a protocol, were developed simultaneously and issued in 1995 and 1996 respectively. E6 and E3 inform a writer how the account of a trial's conduct and the account of its output should be aligned to tell the clearest story of the trial. Although the two were developed simultaneously, in 2014 certain sections of them conflicted, and some parts of E3 were ambiguous. As a result, the information needed to report a clinical trial might not be adequately captured at the beginning of the study, and the guidelines were often interpreted when the results were reported. The ambiguity in E3 also leads to varying interpretations and, ultimately, to different ways of reporting the data. For example, ambiguity in E3's guidance on displays of adverse events can lead to reports that summarise only the numbers of patients experiencing the events without detailing the events themselves. Such reports confound the identification of patterns in event frequencies and compromise the description of individual laboratory abnormalities in the context of adverse events.
E3 and E6 were defined some thirty years before ICH M11 set out the protocol, so what the report should look like was known before what had to go into it, and the two older guidelines often led to some mismatch between what the report requires and what the protocol captured.
In 2014 the protocol guidance in E6 was minimal and open to interpretation. Its protocol sections were much more skeletal than the reporting guidance, which left more room for de novo interpretation. The brevity of E6 left room for more detailed guidance on interpreting it, which would better support both the preparation of protocols and the study reporting that follows. Detailed protocol guidance needed to be developed to improve reporting, optimise reproducibility, enhance transparency and protect participants.
Gaps the clinical study report inherits
Study objectives are often not clearly linked to endpoints, which raises issues of reproducibility and also confounds statistical analysis planning and reporting. In 2014 nothing required the rationale of the study design to be documented, although it is best captured while the protocol is being developed. Protocols are usually developed rapidly, so the multidisciplinary team responsible inevitably focuses on producing a final protocol in the shortest possible time, and recording how and why a particular aspect of the design came about may be a low priority. A rationale left unrecorded can lead to reverse-engineering when the report is produced. A requirement to include the rationale for design elements in the protocol would increase reporting accuracy and enhance transparency.
The trial design section of the protocol in E6 (section 6.4) does not match the report's discussion of study design and choice of control groups in E3 (section 9.2), which can complicate reporting when the protocol did not adequately describe the rationale behind the study design. Such deficits are best addressed by better instructional text in protocol templates.
Mapping either the National Institutes of Health–Food and Drug Administration (NIH-FDA) clinical trial protocol template or the TransCelerate clinical protocol template to E3 leaves the same gaps for the report writer: endpoints is not a heading in E3; end of study is not clearly defined in E3, although its timing, the withdrawal of study drug and the completion of the end-of-study visit, is addressed in sections 9.4.1 and 9.5.1; and the information guiding the definition of screen failures is in sections 9.3.1 and 9.3.2, while the reasons screen failures were identified are in section 10.1. Mapping the NIH-FDA template also shows that E3 refers to no specific strategies for recruitment and retention, although section 9.1 may address similar aspects; that the prevention of loss to follow-up is covered in section 9.6, while the number lost to follow-up is presented in section 10.1; and that unanticipated problems map directly to no section of E3, although a thoroughly constructed flow chart in section 9.5.1 may address some of them. In the TransCelerate template, tests such as the electrocardiogram map to E3 section 9.5 where they define efficacy and to section 12.5 where they define safety.
Changes since the guidelines were issued
The needs E3 was written for have changed considerably since its publication in November 1995. At that time few people outside academia or large companies were using the internet, and for those who were, the browser of choice was Mosaic. WordPerfect was holding its own against Microsoft Word, PDF was a proprietary format that did not support hyperlinks, and XML, the backbone of all electronic submissions, did not exist.
From the clinical perspective, the report and the protocol that the two guidelines address are the core source documents of developing, registering and communicating about new medicines, and that process advanced in many ways over the two decades after E3 and E6 were issued. Many new considerations have arisen since the guidelines were promulgated and are being integrated into the calculus of new drug development, among them disclosure and transparency, structured risk–benefit analyses, electronic data capture and electronic filing for marketing approval.
Since E3 and E6 were developed, unwieldy paper-based systems have given way to multiple electronic systems, and information is accessible electronically and shared through professional and social media platforms. How the resulting insights were then interpreted and disseminated fell outside the remit of statutory regulation as it stood in 2014. Clinical study data must be summarised with absolute clarity, at a level that supports informed interpretation and minimises aberrant claims or criticisms. Forthcoming transparency regulations meant that incorrect interpretation of E3 had to come to an end, and that the guideline had to include clear, directive language devoid of ambiguity.
Updates to the guidance by 2014
By 2014 no formal revision and reissue of the original ICH guidance for protocols or reports had occurred. The attempts made since 1995 to clarify the reporting guidance were isolated and incomplete: regionally, through the European Medicines Agency's 2004 guidance on adapting appendices for reports included in marketing authorisation applications, and globally, through ICH's 2012 supplementary questions and answers document. A consolidated presentation of updated report authoring requirements was also published in 2014.
Annex IV Section A of the 2014 EU clinical trials legislation No. 536 lists the items to be included in publicly posted results summaries, and integrating those items into the guidance for the report synopsis could create efficiencies.
The working group's review
Ambiguity in the guidelines, regulatory and technical developments over the 20 years after them, initiatives to make clinical trial data more transparent and public disclosure requirements together made a review and possible revision of both guidelines necessary. E6 and E3 are inextricably linked, so a review of one requires a review of the other.
At an EMWA meeting in Budapest in May 2014, the question arose whether those who prepare the report and the protocol could be given written guidance that reflects current practices and anticipates future developments as far as possible. That month EMWA assembled a group of experts, the Budapest Working Group, and began a two-year collaboration with a variety of stakeholders to review E6 and E3. The group's recommendations were meant to address the needs of the widest possible community, incorporate developments since the guidelines were first issued and facilitate responsible sharing of clinical trial data.
A de novo review of E3 by current end users was to provide recommendations that minimise ambiguity. The end product was not to be a complete rewrite of E3, because the guideline generally suits its intended purpose. E3 links to many other guidance documents, E6 among them, and to other industry processes and procedures, so the recommendations had to anticipate a possible domino effect. An oversight review of the de novo recommendations was to ensure that broader issues with collateral impact were handled appropriately. Stakeholder evaluation and support of the combined reviews were then to ensure that the final recommendations address the needs of the widest possible community. Because protocol guidance must address some protocol components that ultimately feed into E3, a project to develop recommended protocol guidance was to be undertaken as well.
By the EMWA meeting in Florence, the forensic review of E3 had made significant progress: oversight evaluation of E3 was nearly complete, stakeholder introduction packs had been distributed, and de novo review of the protocol guidance had begun. The group's reviews were expected to be completed in January 2015 and stakeholder reviews to begin in March. Extensive work had been needed on the report guidance, as the group assessed how best to communicate the many subtleties and nuances of the existing E3 guidance.
Whether the report will get an M11 of its own
In 2023 the CORE Reference team knew of no forthcoming ICH initiative to release a template for clinical study reports analogous to M11 for protocols, and no regulator had approached the team. M11 was then at step 2, still to be finalised, and one member of the team believed it would probably take on a good deal of what is in the TransCelerate resources. Those resources do not follow the ICH structure and move sections around a good deal, and the same member held that a report equivalent, if it came about, would be a long, drawn-out endeavour, with a great deal for the regulators and TransCelerate to discuss and agree.