What SPIRIT is and who maintains it
SPIRIT, the Standard Protocol Items: Recommendations for Interventional Trials, is a checklist that sets out the minimum content a clinical trial protocol should contain. It is maintained by the same international group responsible for CONSORT, and the two statements were updated together as CONSORT 2025 and SPIRIT 2025 so that the items a protocol promises line up with the items a results report later has to address (Chan et al., 2025).
A SPIRIT-compliant protocol covers administrative information (registration, funding, roles), background and rationale, the specific objectives and hypotheses, the full trial design including eligibility, interventions and outcomes, the planned sample size and statistical analysis, and the arrangements for data management, monitoring, ethics review, consent and dissemination. It is written before recruitment starts, not drafted retrospectively from what happened.
The checklist is deliberately organised to match the life of a trial rather than the structure of a later journal article: it moves from why the trial is needed, through exactly what will be done to whom and when, to how the results will be analysed and the study closed down. A protocol that follows this order is easier for an ethics committee or a supervisor to review, because each section answers the question a reader would naturally ask at that point.
Why it exists
A protocol written before a trial begins is the only record of what was planned, independent of what the results turn out to be. Without it, a reader of the final paper cannot tell whether the primary outcome, the sample size or the analysis method were decided in advance or chosen afterwards because they gave the most favourable result. This distinction between what was planned and what was reported is central to preregistration more broadly, which separates genuine prediction from the retrospective framing of a result as if it had been predicted (Nosek et al., 2018).
SPIRIT exists to make the protocol itself complete enough to serve this purpose. A protocol that only sketches the intervention and says outcomes will be "assessed at follow-up" gives almost no protection against this problem, because there is nothing precise to compare the final report against.
There is a second, more practical reason SPIRIT matters for a doctoral trial. Writing the full protocol forces decisions that are much harder to make honestly once recruitment has started and early results are starting to come in: exactly which outcome is primary, exactly how missing data will be handled, exactly what counts as a protocol deviation. Making these decisions in advance, under SPIRIT's headings, is itself a safeguard against the researcher's own hindsight bias, not only a document for other people to check.
Which studies it fits
SPIRIT is written for protocols of randomised, controlled, interventional trials, in the same way that CONSORT is written for their results. It applies at the point a trial is designed and registered, before any participant is recruited, and it applies regardless of the trial's size: a single-centre doctoral trial with sixty participants needs the same categories of content as a large multicentre trial, even though the detail within each category will be shorter.
- A doctoral randomised trial of an intervention, whether clinical, psychological or educational in a health setting, should have a SPIRIT-structured protocol registered before recruitment.
- A statistical analysis plan, usually finalised slightly later than the protocol but still before the data are unblinded, sets out the detail SPIRIT leaves at a higher level, such as covariate adjustment and handling of missing data (Gamble et al., 2017).
- Sharing the protocol and, where possible, the underlying data and code once the trial reports its results follows the FAIR principles of findable, accessible, interoperable and reusable research outputs (Wilkinson et al., 2016).
The sample size and analysis sections of the protocol should already meet the basic statistical reporting expectations that any quantitative study is held to, such as those in the SAMPL guidelines (Lang & Altman, 2015), so that nothing essential has to be added for the first time once the trial is reported.
Where SPIRIT does not apply
SPIRIT is for trial protocols specifically, and should not be used as a general template for any research plan. A protocol for an observational cohort or case-control study has different concerns, chiefly the measurement and control of confounding, and should instead be planned with reference to the reporting items in STROBE, bearing in mind that STROBE itself is a reporting, not a protocol, checklist (von Elm et al., 2007). Protocols for systematic reviews use PRISMA-P, and protocols for diagnostic accuracy or qualitative studies have their own separate standards; none of these is a SPIRIT extension.
SPIRIT also stops at the point the trial is registered and under way. It does not tell you how to report the results once the trial finishes; that is CONSORT's role, and treating SPIRIT items as if they were sufficient for the final report will leave several CONSORT-specific items, such as the participant flow diagram and harms reporting, unaddressed.
Why it matters
A protocol is the yardstick against which selective reporting, outcome switching and undisclosed post-hoc analyses are detected. Trial registries and journal editors increasingly compare a published trial's results against its registered protocol, and discrepancies between what was planned and what was reported are themselves now a recognised marker of risk of bias that systematic reviewers are trained to check for. A complete, SPIRIT-structured protocol is what makes that comparison possible in the first place; a vague one gives reviewers nothing precise to compare against, which looks protective but is not.
How it is applied in practice
In practice, SPIRIT is used alongside, not instead of, registration on a public trial registry. Researchers draft the protocol to the SPIRIT checklist, have it reviewed by their ethics committee and, for many trials, a data monitoring body, then register the trial before the first participant is enrolled. The statistical analysis plan is often finalised as a separate, more detailed document shortly before unblinding, building on the outcomes and analysis approach the protocol already committed to (Gamble et al., 2017).
For a doctoral trial, this sequence matters for the viva as much as for publication. An examiner who can see that your primary outcome and analysis were fixed in the registered protocol, and that any change is explained and dated, has far less reason to suspect that your results chapter was shaped around whichever analysis looked best.
It is also common, and acceptable, for a trial to amend its protocol once it is under way, for reasons such as slow recruitment or a safety signal. SPIRIT does not expect a protocol to be frozen forever; it expects every amendment to be dated, explained, approved by the relevant ethics committee and reflected in the registry entry, so that the history of what changed, and when, remains visible rather than disappearing into an undocumented final version.
An illustrative example
The passage below is a hypothetical illustration, not a quotation from a real protocol.
| Protocol element | Without SPIRIT | With SPIRIT |
|---|---|---|
| Outcomes | "Outcomes will be measured at follow-up." | "The primary outcome is change in the 10-item anxiety scale from baseline to 12 weeks; secondary outcomes are listed with their measurement instruments and time points, and the primary outcome is distinguished from all others." |
| Sample size | Not stated, or stated without justification | "128 participants (64 per arm) give 90% power to detect a between-group difference of 3 points (SD 6), two-sided alpha 0.05, allowing 15% attrition." |
| Analysis plan | "Data will be analysed appropriately." | "The primary analysis is an ANCOVA adjusted for baseline score, by intention to treat; missing data will be handled by multiple imputation under missing at random, with complete-case analysis as a sensitivity check." |
| Changes | Not addressed | "Any amendment to the protocol, including changes to outcomes or analysis, will be dated, explained and submitted to the registry and ethics committee before implementation." |
What to check before you register
- Is the trial registered, or ready to register, on a recognised public registry before recruitment starts?
- Does the protocol name one primary outcome, distinct from secondary outcomes, with its measurement instrument and timing?
- Is the sample size justified with a stated effect size, variability estimate, power and significance level?
- Does the statistical analysis plan state how the primary outcome will be analysed, and how missing data and multiplicity will be handled (Gamble et al., 2017)?
- Is there a stated process for recording and dating any protocol amendment?
- Official checklist and source: CONSORT and SPIRIT (consort-spirit.org) .
Key takeaways
- SPIRIT 2025 sets the minimum content of a trial protocol, written and updated alongside CONSORT 2025 (Chan et al., 2025).
- It matters because a protocol is the record that separates planned analysis from decisions made after seeing the data (Nosek et al., 2018).
- It fits randomised interventional trials; observational, systematic review and diagnostic protocols use different checklists (von Elm et al., 2007).
- A SPIRIT-structured protocol and a detailed statistical analysis plan together make later CONSORT reporting far easier to write honestly (Gamble et al., 2017).
- Sharing the registered protocol and, where feasible, data and code follows the FAIR principles for reusable research (Wilkinson et al., 2016).
- The protocol's statistical sections should already meet basic reporting expectations such as SAMPL's (Lang & Altman, 2015).
