Presenting the results for a regulatory reviewer
A clinical study report presents its results so that a regulatory reviewer can find, read and re-examine them. The report is easy to navigate for a reviewer comparing applications, it summarises its own data completely enough for the dossier's summary documents to simplify, it carries its data in tables, graphs and short sentences rather than continuous prose, and it leaves the data in a form the reviewer can inspect and reanalyse.
Navigability for a reviewer comparing applications
A report fit for purpose fulfils the reporting requirements and supports the work of the regulatory reviewer, a time-poor professional who often reviews and compares data across different marketing authorisation applications. The reviewer will appreciate clarity, consistency and brevity, and must be able to find information with ease. The report's author is key to making that possible. Examples of what helps are the report version, draft or final, and its date displayed clearly on the title page and on every page; a data cut date for long-running studies with interim reporting of data, which supports pharmacovigilance; selected tabular information in the narratives; and appropriate data included in the text rather than cross-referred to appended data. Reports should focus on their customer, the regulatory authority, should be reviewer-friendly, and should be written in clear English, because English might not be the reviewers' native language.
Two developments in company process can work against the reviewer. Some larger sponsors and contract research organisations have shifted away from purpose-built in-text summary tables, towards hyperlinking to end-of-text summary tables only or importing unedited statistical outputs directly into the report body. These changes are no doubt intended to streamline statistical and reporting outputs, but they may harm the comprehensibility of the intended message and, in practice, navigation within the integrated document. And as changes in company procedures make templates evolve, the reports within one application may be structured differently, which could reasonably be expected to affect the work of reviewers both within and between dossiers. Restructuring a report for its own sake can likewise lead to differing structures within an application, which may be unhelpful for the reviewer.
Completeness for the summary documents
A report must summarise its own within-study data completely, so that the dossier's summary documents can later simplify them. Regulatory reviewers may be most interested in the summary and overview documents derived from the reports, but the dossier has to rest on well-prepared individual reports. Detailed data from these building-block reports are typically abstracted and repurposed for the submission summary documents: the marketing authorisation application in Europe, the New Drug Application in Japan and the USA, and the US biologic and device equivalents. A widespread lack of understanding of this point can complicate writing the report.
Adverse events are one example. Although adverse effects are of ultimate regulatory interest in a submission summary document, the actual numbers of patients experiencing events have first to be summarised in the report. Ambiguity in ICH E3's guidance on displays of adverse events can produce reports that give only the numbers of patients experiencing events without detailing the events themselves. That confounds the identification of patterns in event frequencies and compromises the description of individual laboratory abnormalities in the context of adverse events.
Tables, graphs and short sentences
Regulatory reviewers appreciate clearly written and well-presented documents. Clear presentation helps them understand the data and may ultimately streamline the review process. Getting safe and effective medicines to market is in the interest of all parties, and approval should not be slowed by suboptimal data presentation. Optimisations may include tabulating selected data now presented as narrative, and using graphs more than summary tables to illustrate trends.
Paragraphing can often be abandoned in the efficacy results section of a report, except for the introductory remarks. The key information in methods and results is often carried by detailed section headers, flowcharts, bulleted lists, tabular lists, tables and figures rather than written out, and those elements are often better than text, so single-sentence or very short paragraphs that link and complement them are usually enough. For example, where each variable has its own heading, such as overall survival, progression-free survival, time to treatment failure and duration of response, a topic sentence describing it only tells the reader what the heading already says. The few sentences that carry the messages of an in-text table can be grouped, bulleted or left standing separately. The same is often true of the safety section. Separate sentences also make it much easier to prepare the summaries and the synopsis by cutting and pasting.
What a table in a clinical study report says should be immediately apparent, so a reader is not left working out what a given figure means. Piles of material the reviewer has to guess at are the problem, and the remedy is to lay out the main points. Merely referring the reader to the end-of-text tables does not make a document easier to review. A medical writer supports the reviewer by telling them what the data mean and putting the data in context.
The key messages of a report must be clear and unambiguous. In a nutshell, they should support the efficacy and safety of the drug under development and clearly show how the compound will affect the target indication. They should also be reviewer-friendly, consistent and convincing, and the report should be unbiased and objective and avoid superlatives. Many reports are excessively long, complex, difficult to read and repetitive, and particular effort should go into their summary sections. Crucial information lost in a long, complicated document has the same impact as a document that is too brief and leaves out essential information.
Data the reviewer can inspect and reanalyse
Regulatory authorities may want to reanalyse certain data or investigate a specific subset of patients, so data listings should be in a format that is easy for the reviewer to use, because going through such documents can be excessively cumbersome. Safety data are usefully shown both against normal ranges and as the change from baseline. A useful attitude for the writer is to imagine being the reviewer. Reports should be consistent across the project in format, structure and content.
ICH E3 recommends that the report and its appendices carry enough individual patient data, including demographic and baseline data, and enough detail of the analytical methods for an authority to replicate the critical analyses. Every analysis, table and figure should identify, in the text or as part of the table, the set of patients it was generated from. In any table, figure or data listing, estimated or derived values should be identified conspicuously, with a detailed explanation of how they were derived and what was assumed.