Drafting the report from the precursor documents and the outputs
A clinical study report is drafted in two halves. Much of the methodological front end is summarised and adapted from precursor documents such as the investigator's brochure, the protocol and the statistical analysis plan. Writing the results is driven by the availability of the final tables, figures and listings. The results are presented factually, with interpretation left to the discussion, and only the results of preplanned analyses are reportable in the report itself.
The front end and its precursor documents
The text of a clinical study report divides into a predominantly methodological front end and a back end carrying the results, discussion and conclusions. Besides a stand-alone synopsis, the front end should contain the ethical information, the trial's administrative structure, the introduction, the objectives, the investigational plan, the statistical methodology, and the changes to the study or planned analyses. Much of the front end is summarised from precursor documents such as the investigator's brochure, the protocol and the statistical analysis plan.
Multiple authors, inevitably with different perspectives and standards, may have contributed to the precursor documents. Text from them should not simply be lifted but adapted, repurposed and written in the past tense where appropriate. Existing text may be better presented as a table where that adds clarity and aids comprehension.
Cross-referencing sections of precursor documents avoids unnecessary repetition in the marketing authorisation application, but care must be taken to cross-reference only accurate original material. Where accuracy is in question or the text is open to interpretation, better practice is to include abstracted, unambiguous information directly in the report. Whether to summarise precursor material in the report or only hyperlink to the original documents has been debated in industry, and as of 2014 guidance required the relevant front-end text sections to be populated.
The results
Writing the back end is driven by the availability of the final tables, figures and listings. In-text summary tables are commonly populated with selected relevant data from them. The aim is to distil the voluminous output data into easily comprehensible results messages while maintaining absolute transparency. Done analytically and with elegance, the distilling should help identify the signals in the noise. The presentation of results must be factual, and interpretation is not required except in the discussion.
Post-hoc and supportive analyses
Only data from preplanned analyses are reportable in the report itself, so the appropriate place for a post-hoc analysis undertaken by the sponsor is an ad hoc appendix. The rationale for an appended post-hoc analysis must be included in the report's section on changes in the conduct of the study or planned analyses (Section 9.8), and the analysis can be reported further in a journal article. Supportive analyses, for example to aid interpretation of the results, may also be appended, whether planned or post-hoc with an explanation in Section 9.8.