Item 3: Where the trial protocol and statistical analysis plan can be accessed

Example

“The full trial protocol and the Statistical Analysis Plan can be accessed in the Supplementary Material” [74]. This article and supplementary material are open access.

​

Explanation

A protocol for the complete trial (rather than a protocol of a specific procedure within a trial, such as for the intervention) is important because it prespecifies the methods of the randomised trial, for example, the primary outcome (item 14). Having a protocol provides important context to interpret a trial, implement its findings, and facilitate replication and appraisal of risk of bias. It can also help to restrict the likelihood of undeclared post hoc changes to the trial methods and selective outcome reporting (item 10) [75-77]. Elements that are important to address in the protocol for a randomised trial are described in the SPIRIT 2025 statement [78].

A protocol may either include the full statistical analysis plan or may include a section outlining the main principles while referencing and reporting the full statistical analysis plan as a separate, more detailed document. The statistical analysis plan typically includes details about several aspects of the clinical trial, such as the data analysis plan for the primary outcome(s) and all secondary outcomes. Details about what to include in a statistical analysis plan can be found elsewhere [79]. 

The protocol should be signed off by the trial steering committee before the allocation of any participants and data collection. Similarly, the full statistical analysis plan should be signed off by the trial steering committee before the dataset is closed for analysis. This allows transparent documentation of any subsequent changes to either document. In many trials, changes to the protocol and statistical analysis plan may happen after the trial onset for legitimate reasons (eg, in response to challenges that were not anticipated or new evidence). In these cases, each iteration of the protocol and statistical analysis plan should record the changes along with their rationale and timing.

There are several options for authors to consider to ensure their trial protocol, and full statistical analysis plan where applicable, are accessible to interested readers. The protocol and full statistical analysis plan (and their various iterations) can be stored in a repository, such as the Open Science Framework, which is free to use and access for all readers. Openness and accessibility are core elements of open science (see Open science section). Trial protocols and statistical analysis plans can also be published in journals such as Trials and BMJ Open. Open access publication would ensure that any reader, including patients and the public, can access the document. Trial registration (item 2) will also ensure that a minimum set of trial protocol details are available as part of the trial’s registration (https://www.who.int/clinical-trials-registry-platform), but often a registration record leaves large ambiguity about key protocol issues and statistical analyses. The more detailed trial protocol can also be posted on most clinical trial registries.

Ideally, the authors should give access to the protocol signed off by the trial steering committee before the allocation of any participants and data collection with any subsequent changes with their rationale and timing.

When submitting a completed trial report, trial authors can include their protocol as a supplemental document or provide a URL to its location. Such documentation can facilitate peer review and help identify reporting biases.