CORE Reference for clinical study reports
CORE Reference is an open-access user manual for authoring the clinical study report of an interventional study, built on ICH E3 and written because E3 had not kept pace with modern study designs and the public disclosure of reports. It offers content suggestions rather than a template, and it differs from the TransCelerate template and other resources in what it clarifies. Its preface sets out its assumptions and rules, and its shading and comment boxes show where its text comes from and why. It settles terminology, flags information sensitive to disclosure, lists the guidance it drew on as that guidance stood in May 2016, and carries E3's annexes.
What the resource consists of
CORE Reference is an open-access resource for authoring the clinical study reports of interventional studies, held to be complete and authoritative. It is held to be complete because the components an author needs to produce a fit-for-purpose report were thought through carefully, and authoritative because it follows ICH guidances together with EU and US regional guidances and a number of important stakeholders were involved in developing it. It is based on ICH E3, and it contains both the guidance text of E3 and the guidance text derived from the 2012 questions and answers on E3.
CORE Reference comprises a text document, a mapping document and an explanation and elaboration paper. The text document consists of a preface followed by the resource itself. The text document is provided as a PDF and the mapping document as a spreadsheet, and together the two constitute the user manual. The explanation and elaboration paper, published in a BioMed Central journal in 2016, accompanies CORE Reference.
The mapping document, or mapping tool, shows how the sections of CORE Reference map onto E3's sectional structure. It is a four-page overview of how finely each main section of the manual is divided, set against the sections of the ICH E3 guideline. Users can download it to keep as a sectional reference, while the manual itself describes the content of every section and subsection in full. The manual can be downloaded with one click, and the mapping tool has been downloadable since it was published.
Why a manual was needed on top of E3
Any guidance reflects requirements at a static point in time. The clinical study report has its origins in the 1995 ICH E3 guidance on the structure and content of reports and in the 2012 questions and answers that supplement it, but modern study designs often integrate pharmacokinetic, pharmacodynamic, pharmacoeconomic and pharmacogenomic elements with a safety and efficacy backbone. To meet complex clinical-regulatory requirements, such studies need a fit-for-purpose reporting framework that may differ substantially from that of the more straightforward efficacy and safety studies which E3 set out to support. An ever-growing substructure of regulatory guidance also carries further content requirements that must be worked into a report, and CORE Reference integrates them. Even the most experienced and laterally thinking author must be extraordinarily diligent and well informed to keep pace.
Since 1995 there have been incomplete attempts, regional and global, to clarify the reporting guidance for clinical study reports, but the original ICH E3 guidance has not been formally revised and reissued. This left individual writers to interpret the guidances subjectively. So the ambiguities in E3 and its 2012 questions and answers needed to be addressed and reconciled with the EU and US regional report guidances, and practical suggestions presented for developing reports that incorporate clarifications of the numerous guidances. Any such effort would essentially take the form of a user manual for report authors.
Since the 1995 E3 guidance and its 2012 questions and answers were published, awareness of the importance of disclosing clinical study results has grown, and the use of the report as a key source document for meeting emerging obligations has led to a re-examination of how the ICH guidelines are applied in preparing reports. A changing regulatory environment and modern drug development create emerging reporting challenges.
Public disclosure of reports, mandated in the European Union, has profound effects on the way reports must be written. The effects on the report are multiple and complex, and lessons would be learned as reports were disclosed in increasing numbers. Regulators around the world were following the European Medicines Agency's lead on public disclosure of clinical data. As a result of all this, report authors faced increasing challenges in creating reports that support heterogeneous study designs while covering all the important and emerging content requirements, among them current and future public disclosure requirements. In 2018, policies and guidelines in development beyond European borders were expected to further mandate carefully authored reports for sharing in a global public arena.
Submitting and then publishing clinical study results is a two-step process: a submission-ready report, which may contain sensitive data, is produced first, and those data must be removed after submission to produce the final disclosure-ready report. The primary use report, which the European Medicines Agency calls the scientific review version, is the technical document for regulatory review and comprises the full report text and all its appendices. The secondary use report, the agency's redacted clinical report, is for public disclosure: it comprises the redacted report text and selected appendices, with the sensitive information of the primary use report redacted.
While it was still in its final stages of review, the manual was expected to help clear up some ambiguities in the existing guidelines, to describe current best practice, and to help foster the responsible sharing of clinical trial data by suggesting safeguards to protect patient confidentiality and proprietary company information.
How far the manual binds the report
E3 is a guidance and not a template, and CORE Reference is not one either. CORE Reference cannot be a template, because there are so many study designs. It offers content suggestions, and how that content is arranged to fit the study design is the author's decision, though the author has to present what the regulators want while being careful to cover the content they require. CORE Reference also offers pragmatic structural solutions for reporting a wider range of the study designs common in modern drug development, beyond the typical mainstays of safety and efficacy.
Neither E3 nor the 2012 questions and answers mandates a template order for presenting design elements, so the structure can suit the individual study. Without a common approach, however, building a framework study by study inevitably results in wide variability between report structures. The manual suggests a focused structure and content that address the current guidance documents.
The manual presents a unified approach to writing clinical study reports, and one of its developers holds that the more common ground there is, the better for all concerned. Another, an advocate of standards that guide interpretation, rationale and best practice in complying with the expectations of regulatory authorities worldwide, holds that the manual serves that purpose as an accessible, clear resource for those who prepare clinical study reports. The manual facilitates the authoring of a content-driven report that is as disclosure-ready as possible, and it has the potential to increase the quality of final reports and to enhance consistency within and between sponsors.
The manual follows E3 because it had to align with the regulators' actual guidance. Portions of it can still serve other studies, and writers have reported applying them to device, animal and observational studies.
The manual keeps E3's level-one heading hierarchy, so that it maps easily onto the original guidance. The working group changed no level-one or level-two heading numbers of E3, even where a section would make more sense elsewhere, as the reference section would at the end. One reason is that, from experience, the FDA holds granulated documents strictly to E3's section 16.2 numbering, so a report whose numbering departs from E3 risks rejection. The other is that guidances speak to one another: where another guidance refers to E3's section numbering or to an E3 level-one heading, keeping that heading undisturbed avoids conflict with that guidance and keeps its reference in context. Beyond those headings, the most appropriate structure remains the author's decision, and the manual offers some guidance drawn from its development team's experience.
What the manual offers beside the TransCelerate template and other resources
Clinical study reports are not necessarily destined for a regulatory submission dossier, because many drugs involved in clinical studies fail during development and their studies never progress to a product submission. A report written for each study must stand alone for review by regulators, sponsors, investigators, investors and other interested parties. A standalone report contains a full description of the study, including the source tables, figures, listings and all the appendices needed to understand the study's context, with minimal cross-references to external documents.
Large pharmaceutical companies, which house their reports within complex, closed document management systems, are not the only sponsors of clinical studies, and they do not own all products from inception through to licensing. Sponsors also include biotech developers, investigators and charities, whose less elegant and agile systems may inadequately support content reuse and document linkage, both within the closed system and externally, for example when a product changes hands. In preparing full submission dossiers, TransCelerate's protocol and report templates, their content-reuse solutions and the cross-talk among documents within the submission are useful to the sponsor and the regulator. When reports need to stand alone, as at some point in a product's development they all must, extensive content reuse and hyperlinking cannot always adequately serve the needs of their mixed audience. The TransCelerate report template in particular may not serve sponsors outside large pharmaceutical companies well, including even the larger contract research organisations that need flexibility in their reporting template to serve a wide range of client types.
CORE Reference predates the TransCelerate report template by two and a half years. Besides supporting the need for standalone, publicly disclosable reports, it is the only resource that incorporates clarifications of particular regulatory guidance or legislation applied to granular report content requirements. It also clarifies how to interpret wording in ICH and regional guidance that is difficult to understand or ambiguous, guiding writers to informed choices that produce a report fit for reporting their study.
CORE Reference also stands apart from other open-access resources, including guidance and templates, in its clarifications that aid the interpretation and understanding of reporting and public disclosure requirements.
The preface
The preface should not be skipped: it sets out the assumptions the working group made and the rules it set, and the regulatory and public disclosure references cited in it show how the manual was built. One account gives the preface as the first twenty pages of the manual. Another has it end on page twenty-one.
The preface opens by saying what CORE Reference is: a user manual that helps medical writers navigate the relevant guidelines as they create reports, and that clarifies the areas of ICH E3 and the 2012 questions and answers that conflict or are difficult to understand. On why the manual is needed, it explains that the report should be considered a single document with two uses, each with its own purpose and audience. It also clarifies the manual's intended use and the underlying principles that inform how the resource is used.
Terminology and general clarifications
Unless the manual states otherwise, its terminology follows the CDISC glossary, and pages six to ten present the key terms it uses together with the reasons for choosing each one. The choices recorded there include study or trial; patient, subject or participant; and investigational product, study medication, study treatment or study drug. The debate over the last of these runs for a further page.
Among subject, patient and participant, the author can choose any word, provided it is used consistently. The manual settled on "subject" and explains why, while TransCelerate uses "participant". Because the resource is globally applicable and some regions may have specific requirements for the language used, the manual gives options rather than a ruling.
Alongside the key terms, the manual defines the public disclosure terms: protected personal data, personal protected information, individual personal data and commercially confidential information, and anonymisation, masking and redaction.
The manual's general clarifications also cover language and links. The choice between UK and US English should not matter, provided it is consistent within a document and across the documents in the submission package. The manual does not support links from the report to the appendices where the linked information is necessary for understanding the results in context; such text should be placed in the body of the report instead.
Shading and comment boxes
E3 and the 2012 questions and answers do not have to be kept open alongside the manual, because everything is in one document, where shading and clarifications show where all the guidance text comes from. Shaded text is ICH E3 text, text the working group added is not highlighted, and text from the 2012 questions and answers is shaded, italicised and boxed. The key to all three appears at the foot of every page, in case a reader forgets it part-way through the document.
Comment boxes down the right-hand side of each page carry the working group's rationale for its proposed changes, with any applicable references, and clicking a reference leads directly to it. Every piece of E3 text is included as original wording, included with modified wording and the modification explained, or omitted with the omission shown and its reason given. Every piece of the 2012 questions-and-answers text is included and explained. Where alternative presentations of the same information would work equally well, they are shown, with a rationale that lets the author choose for the study in hand. Throughout the manual, value-add comments guide the author on what the EU Clinical Trials Regulation additionally requires for EU studies.
Disclosure readiness and the flags for sensitive information
The manual is a resource for authoring reports fit for the public disclosure environment. It proposes that the report be as disclosure-ready as possible from the outset, to safeguard against inadvertently identifying participants or commercially confidential information, to assure optimally timed public disclosure of results, and to be as cost-efficient as possible. Smart authoring of the regulatory review version, incorporating proactive anonymisation of sensitive data, keeps data utility high while minimising the redaction needed in the public disclosure version. The manual offers practical suggestions for developing reports that need minimum redaction and modification before public disclosure. It also gives insights and suggestions on anonymisation techniques and approaches that minimise the redaction required in the publicly disclosed report.
Guidance on public disclosure is integrated into the manual through discrete colour-coded comments prompting the author to consider both the primary and the secondary use report. These comments should help authors make informed choices as they navigate the evolving and complex area of redacting sensitive information before public disclosure.
The manual makes its content suggestions for the primary use report. Awareness comments highlight possible protected personal data in blue text and possible commercially confidential information in red, marking the portions of the primary use report that should be considered for their impact on the secondary use report and so for redaction in it. In one example from the body of the manual, the names of individuals other than a principal investigator, and telephone numbers, are marked in blue as content that could be considered protected personal data, with references included where applicable.
In Section 10, on study subjects, a comment on protected personal data links back to the preface's explanation of the risk of de-anonymisation from aggregated data. Generally, any listing information that has to appear in the primary use report and could de-anonymise participants will require redaction in the secondary use report.
The manual highlights the hotspots in a report where sensitive information, such as personal information or commercially confidential information, may need safeguarding in line with EMA Policy 0070, and it links to the relevant guidance, which makes it an all-inclusive resource. CORE Reference is the only resource that shows in granular detail where either kind of information may occur in a clinical study report. It is also the only known freely available resource that pinpoints the sections of an ICH E3-compliant report potentially affected by public disclosure considerations. It identifies each point in such a report where anonymisation considerations should apply.
The reference list and its dates
From page ten to the end of the preface, the manual details the guidance documents used to develop it. These references reflect guidance as it stood in May 2016, and guidance has since been updated. They fall broadly into regulatory and public disclosure references.
The list marks which references are region-specific. The regulatory references include ICH E6, E3 and E9, the CDISC glossary and the EU Clinical Trials Regulation, the last flagged as EU-only. The study data tabulation model is flagged as US-specific, the CHMP note for guidance on the inclusion of appendices to clinical study reports as EU-specific, and the FDA guidance for industry on the submission of abbreviated reports and synopses as US-specific but used in multiple regions. Page thirteen holds a selection of data privacy references with the same regional marking, HIPAA as US-specific and the Health Canada Act as Canada-specific. The references describing and defining commercially confidential information are all EU-specific, among them the EMA's Policy 0043 and Policy 0070 with their accompanying external guidance.
The disclosure references cover clinical trial data sharing and include the obligation to post results to the US and EU trial registries. The EU Clinical Trials Regulation and its clinical trial information system were not in place when the manual was put together. The references on redaction include Opinion 05/2014 on anonymisation techniques, which the EMA also quotes in the Policy 0070 guidance. When the manual was published, its standards for de-identifying data came from the references on page eighteen of the preface, Opinion 05/2014 among them. Standards for de-identifying data are continuously evolving, and medical writers should keep up to date with them.
The annexes on statistical issues
The manual's annexes should not be overlooked, and for each one it gives its number and title in ICH E3 and its position in CORE Reference. Annex 4, for example, is invaluable for statisticians and programmers and supports ICH E9, so the manual should be shared with statistical colleagues. The material of Annex 4 appears in E3 but not in E9.