Lean authoring: cutting a regulatory document to its key messages
Lean authoring, also called focused authoring and lean writing, removes unnecessary words and content from a regulatory document so that its key messages come through. It works by putting a few questions to every passage, and its effect shows in single sentences and in whole clinical study reports. Cutting too far leaves too little, and adoption depends on bringing the team and the organisation along.
What lean authoring is for
Lean authoring is a way of writing that removes unnecessary words and content to focus documents on key messages. It has also been called focused authoring. It preserves the valuable real estate in a document and the attention of its reader by letting the key messages come through. Some sections may even become longer, but all the fluff is removed.
Authors are routinely faced with expedited timelines to write documents about highly complex studies with overwhelming amounts of data, and reviewers are then presented with complex content whose messages are often hidden among unnecessary text and numbers. Lean authoring is important in medical writing for large documents, data-driven content and short authoring timelines. Its practitioners at Merck hold that medical writers must use lean authoring principles as standard practice to keep the important information in their documents clear.
Successful incorporation of lean authoring principles leads to streamlined documents that are easy to read, straightforward to interpret and improve the overall reader experience. The benefits claimed for it are easily identifiable key messages, reduced writing, review and quality-control time, and increased quality through reduced redundancy. In one practitioner's view, unnecessary words create work for readers, and the medical writer's job is essentially to take the work out of reading. Focused authoring takes to heart the line attributed to Dr Seuss that "the writer who breeds more words than he needs is making a chore for the reader who reads".
Lean authoring is not only about cutting content but also about improving the reader experience by focusing documents on the key messages and the most relevant content. Merck & Co. reported using it to reduce document length by 50% to 75%. Since the company implemented it, feedback from its document reviewers had been overwhelmingly positive, and with less time needed to read the documents and key messages easier to understand, many had called them "enjoyable to read".
Writers reporting safety data claim these benefits for lean writing:
- It highlights key messages clearly and quickly, focusing the reviewers' attention on what really matters.
- It provides clarity, forcing decisions on what is required to tell the story and reducing unnecessary or irrelevant detail.
- It improves accuracy, as a change needs to be made in only one place, so there is less chance of introducing consistency errors.
- It reduces complexity, as cross-referring avoids constantly aligning changes in each document across all the other documents being prepared in parallel; this considerably simplifies project management and saves time, which ultimately reduces the cost of preparing the submission.
- It supports global submissions, reducing the amount of text that must be localised to different regions and so facilitating parallel submissions; if desired, region-specific information can be included as an appendix instead of in the main document and could be replaced at the publishing stage to keep the document compliant with regional regulatory considerations.
- It lowers the redaction burden, making it easier to comply with clinical data disclosure initiatives worldwide.
The questions put to every passage
Three questions are put to the text to focus a document on its key messages. In the version Merck writers published in March 2021, the questions are whether the text is needed, whether the message is clear and whether the presentation is effective. In the version presented in November 2018, they are whether the text is needed, whether it is clear and whether it is accurate. That version puts two steps before the questions: assessing and understanding the audience, and explaining the approach to get buy-in from collaborators.
The audience
Understanding the needs, priorities, background and experience of the audience, and what it hopes to get from the document, helps drive what text is needed. Once the writer knows whether the document is for regulatory reviewers or health care providers, for example, it can be tailored to their needs. Whoever the readers are, they will be busy, so the writer should work hard to gain and keep their attention.
Is the text needed?
In a document for regulatory reviewers, unnecessary text can be eliminated by tailoring the introduction to the audience, avoiding descriptions of methods in the results and not repeating data presented in tables. Introductions for regulatory authorities, which are therapeutic-area experts, can be minimal, whereas introductions for academic audiences may require more therapeutic-area context. There are always team members who want to add more information to the introduction, and they have to be reminded to focus on the key messages a regulatory reviewer really needs to know. Each message should be said once, by using cross-references, keeping content in the section where it belongs, and relying on tables to present data while avoiding numbers in text. A team member who fears that reviewers will not read the methods section may ask for the methods to be summarised in the results; the writer should not agree, and should cross-reference from the methods to the results instead. Most data should stay out of the text, because a table communicates results better than words ever could. Key values and other numbers are sometimes needed in the text, and they can be placed in parentheses, or a smaller version of a large results table can be added to the text to highlight the key results.
Is the message clear?
Clear text needs consistency, a considered format and concision. Every word in a sentence should be needed, either serving a grammatical purpose or adding information that moves the document forward, in contrast to the style usually seen in clinical study reports and even in manuscripts. A point that can be said in nine words should not take 15, as simpler language leads to easier comprehension.
The best documents communicate information in "one voice", using consistent terms and parallel structure throughout. Consistent terminology and a consistent order of presentation let readers focus on the content and move more quickly through the messages.
Bullet points, tables, figures and other visual cues can present information instead of lengthy sentences or paragraphs; visually organised information is easier to understand, read and interpret.
Text can be made more concise by asking whether each sentence can be read out loud in one breath: whether its semicolon is the best choice, whether it runs to 2 lines or fewer and whether each word adds value. A no to any of these means the sentence is too long and should be rewritten.
Is the presentation effective, and is the text accurate?
Accuracy in the 2018 version does not necessarily mean verification against the source document; it means that the words used accurately reflect what the writer means. Accuracy can be achieved first by establishing an order and keeping to it: if an active treatment group is discussed before a placebo group, the active treatment group always comes first. Comparison language should then be consistent: a report can give the number of subjects who survived or the number who died, but not both unless there is a reason.
The reader is engaged by starting with the most important information and by making precise statements and precise, non-biased comparisons, while avoiding too many comparisons. Discussion of subgroup analyses should be limited when they are similar to what is published about the group as a whole.
Possible bias shows in opinion presented as a statement of fact, in absolutes such as always, never, certainly and totally, and in approximate terms used inconsistently. Approximate terms such as most should be used consistently, and cut-off points can be established for when to use terms such as most, more, similar and the majority.
Before-and-after revisions
Pairs of wordy sentences and clearer replacements show the method at sentence level.
- Over the course of the study, a total of 31 participants were randomised (21 participants were randomised in Treatment Group 1 and a total of 10 participants were randomised in Treatment Group 2) becomes A total of 31 participants were randomised (21 in Treatment Group 1 and 10 in Treatment Group 2).
- Table 10-4 displays the demographic information collected for the participants enrolled in the trial, which shows that the mean age of participants was similar in the PJ-123 treatment group and the Happimab treatment group becomes The mean age of participants was similar in the 2 treatment groups (Table 10-4).
- The safety population included all participants who were dispensed study medication in the study and were documented to have taken at least one dose of investigational treatment becomes All participants who received ≥1 dose of study medication were included in the safety population.
- The proportion of participants who reported serious adverse events in Treatment Group 1 was lower than the proportion of participants who reported serious adverse events in Treatment Group 2, as shown in Table 12-5 becomes The proportion of participants with serious adverse events was lower in Treatment Group 1 than in Treatment Group 2 (Table 12-5).
Cutting too far
Lean writing is beneficial, but consideration should be given to when to apply it: it avoids wordiness and redundancy, yet reporting too little information is unhelpful too. Overly lean early drafts may require more substantial revisions later; it is much easier for a team to decide that information is no longer needed than to notice that something that needs to be reported is missing.
A clinical study report cut to its key messages
Merck retooled the process and structure of its clinical study reports so that they both facilitate the job of the health authority reviewer and fully leverage the efforts and expertise of the reviewers and authors the company employs. The project was part of a movement within the company to simplify the documentation for clinical studies. To move from its dense reports to the simplified ones, a team of cross-therapeutic specialists collected cross-functional input, conducted benchmarking within the industry and critically analysed the template and procedures for the report.
The revised template and process used focused authoring principles, leveraged the electronic environment, addressed disclosure concerns and built on structured authoring initiatives. In the simplified template, section 9 uses CORE (Clarity and Openness in Reporting), sections 4 to 9 are streamlined and linked to the protocol, sections 10 to 12 are revised in a focused-author style, and section 13 is clarified.
The guiding principles of the simplified report included building on industry best practices, embracing focused authoring principles, leveraging the electronic environment and leveraging the core protocol template. They also included avoiding distracting redundancies and unnecessary descriptions, reusing reviewed and approved text from the protocol, keeping the protocol only one click away, and compliance with ICH E3, the FDA, the Code of Federal Regulations and the CORE Reference. The principles further included a concise, focused document that helps the reviewer see the key results messages, reduced author, reviewer and quality-control time, a reduced potential for error, building on existing initiatives to simplify documentation, minimal customisation, and applicability to most studies.
A simplified report is no longer a comprehensive document containing all study information but a clear and concise one that reports the results and key messages.
In one comparison, simplification reduced a report from 27 pages to 4. The original report had a lot of white space and 14 tables, where the simplified report was reformatted and had a single adverse event table. The original restated in text the information already in its tables, where the simplified report briefly summarised the key methods and analyses. The original had a combined section describing both the methods and the results, where the simplified report gave the key messages on efficacy and safety. The original described additional endpoints in detail, where the simplified report included only the primary and secondary endpoints. The original used paragraphs, and the simplified report bulleted lists.
Implementing such changes requires communication, training and reinforcement to achieve the best outcomes.
Bringing the team and the organisation along
Lean authoring may be new, and new ways of working may be daunting to writers, their teams or their organisations; some teams may be eager to adopt a writing strategy that Merck writers describe as becoming the industry standard. For more hesitant organisations, successful adoption depends on sponsorship and advocacy from senior leadership, socialisation to gain support from authors, teams and management, effective training followed by setting and reinforcing expectations across projects, and processing documents and templates as tools for successful implementation and execution.
The team needs to be just as aligned with the strategy as the writer. That takes a proactive approach: telling the team that the style may be different from what they are used to and unexpected, which in some cases creates a lot of anxiety, and helping them understand that even though it is different, it is better.
An email to reviewers can set these expectations in advance. A sample email tells reviewers that a focused authoring style was used throughout the document, that aspects of other styles they may be used to seeing in past clinical study reports or manuscripts are now handled differently, and that the approach helps the reader stay focused on key messages. Instead of methodology repeated in the results sections, lots of numbers from in-text tables repeated in the text, "roadmaps" describing what the current and subsequent sections will discuss, and descriptions of in-text tables, reviewers should expect very brief reminders with cross-references to the relevant sections, and only when absolutely necessary; text that describes the results and keeps to the key points without cluttering it with numbers; no roadmaps, as the navigation pane and table of contents show what is coming up; and results described directly, with the table cross-referenced at the end.