SPIRIT 2025 · Trial Protocols

The SPIRIT 2025 checklist, item by item.

SPIRIT is the reporting standard for randomised trial protocols — the document an ethics committee, a funder, and eventually the public will judge your trial by before a single participant is enrolled. The 2025 update sets out 34 minimum items plus a schedule diagram. This is a complete walkthrough of what each item requires — and where protocols most often fall short.

Last updated: August 29, 2026
Check Your ProtocolLearn About AI Peer Review
The guideline

What SPIRIT is.

SPIRIT — Standard Protocol Items: Recommendations for Interventional Trials — defines the minimum a randomised trial protocol must contain. Where CONSORT checks the report of a finished trial, SPIRIT checks the plan: the document an ethics committee approves, a sponsor's quality system audits against, and a registry makes public. Protocols fail in predictable ways — primary outcomes named without a measurement variable or time point, sample sizes asserted without assumptions, allocation concealment described in one vague sentence, data-sharing intentions that name no repository. SPIRIT exists so that none of those gaps survives to enrolment.

The original SPIRIT 2013 statement carried 33 items. SPIRIT 2025(Chan et al., BMJ 2025;389:e081477 — published simultaneously in JAMA, The Lancet, Nature Medicine, and PLOS Medicine) updates it to 34 items plus a schematic diagram of the schedule of enrolment, interventions, and assessments. An Explanation & Elaboration document (BMJ 2025;389:e081660) gives the rationale and good-reporting examples for every item, and the guideline is maintained alongside CONSORT at consort-spirit.org and listed by the EQUATOR Network. The two 2025 updates were deliberately harmonised: a protocol written against SPIRIT maps item-for-item onto the final report written against CONSORT.

SPIRIT applies to protocols of randomised trials of any design. If the trial is finished and you are reporting results, you need CONSORT 2025. If your protocol is for a systematic review rather than a trial, the standard is PRISMA-P — see the reporting-guidelines hub for the full map.

The checklist

SPIRIT 2025, item by item.

The 34 items run in the order of the official checklist — administrative information, the new open-science section, introduction, methods, monitoring, and ethics. Twelve items carry lettered sub-items; confirm each sub-item, not just the heading. These are the topics each item asks you to cover — what is required, and why reviewers care.
Administrative information · items 1–3
01
Title and structured summary
Sub-items 1a–1b. The title must identify the trial design, population, interventions, and the fact that this is a protocol; the structured summary carries the WHO Trial Registration Data Set items. This is the material registries index and reviewers triage by — a title that hides the design costs the protocol its findability.
02
Protocol version
A version identifier and date. Trials run for years and protocols get amended; without version control nobody can say which version the ethics committee approved, which version the statistician analysed against, or what changed in between.
03
Roles and responsibilities
Sub-items 3a–3d: protocol contributors with affiliations and roles; the sponsor's name and contact information; the sponsor's and funders' role in design, conduct, analysis, and the authority to report; and the composition and responsibilities of the committees — steering, endpoint adjudication, data management. The reader must be able to see who controls the data and the decision to publish.
Open science · items 4–8 — new section in 2025
04
Trial registration
Registry name, identifying number with URL, and — new in 2025 — the date of registration. Prospective registration is what lets readers later check the report against the plan; the date is what proves the registration was prospective.
05
Protocol and statistical analysis plan access
Where the full protocol and the statistical analysis plan can be accessed. A published report citing an inaccessible protocol cannot be verified against it; this item makes the plan itself a public artifact.
06
Data sharing
Plans for making the de-identified participant data, data dictionary, statistical code, and other materials accessible. SPIRIT 2025 asks for a plan — what, where, when, under what access conditions — not a sentiment. This is one of the items that moved data sharing from the report's afterthought to the protocol's commitments.
07
Funding and conflicts of interest
Sub-items 7a–7b: sources of funding and other support, and the financial and other conflicts of interest of the principal investigators and — extended in 2025 — the steering committee. Declared at the protocol stage, before results exist to be spun.
08
Dissemination policy
Plans for communicating results to participants, healthcare professionals, the public, and other relevant groups. The 2025 revision folds authorship and professional-writer policies in here: who will write the report, and who decides it gets published, are protocol-level facts.
Introduction · items 9–10
09
Background and rationale
Sub-items 9a–9b: the scientific background with a summary of relevant studies examining benefits and harms — both, a deliberate emphasis of the 2025 revision — and the rationale for the choice of comparator. A trial justified only by its intervention's promise, with the comparator unexamined, is a design decision without a defence.
10
Objectives
Specific objectives relating to benefits and harms. Objectives phrased as marketing ("to demonstrate the efficacy of…") pre-commit the trial to one answer; objectives phrased as questions about benefit and harm are what the outcomes in item 16 must then operationalise.
Methods — involvement and design · items 11–12
11
Patient and public involvement
New in 2025: details or plans for involving patients and the public in the design, conduct, and reporting of the trial. If there was none, that is stated too — the item makes involvement a reportable fact rather than an assumed virtue.
12
Trial design
The type of design (parallel group, crossover, factorial), the allocation ratio, and the framework — superiority, equivalence, non-inferiority, or exploratory. The framework determines what the sample size in item 19 and the analysis in item 27 must be built to answer.
Methods — participants, interventions, outcomes · items 13–20
13
Trial setting
The settings and locations where the trial will run — countries, sites, care level. Setting is generalisability: a reader cannot judge whom the results will apply to without knowing where participants come from.
14
Eligibility criteria
Sub-items 14a–14b: inclusion and exclusion criteria for participants, and — where applicable — eligibility criteria for the sites and the individuals delivering the interventions (surgeons, therapists, centres). The second half is routinely forgotten in protocols for skill-dependent interventions.
15
Intervention and comparator
Sub-items 15a–15d: each intervention described in enough detail to replicate — how, when, by whom, with access to any manual (an expectation sharpened by TIDieR); the criteria for discontinuing or modifying the allocated intervention; adherence strategies and how adherence will be monitored; and the concomitant care that is permitted or prohibited.
16
Outcomes
Primary and secondary outcomes, each with its specific measurement variable, analysis metric (change from baseline, endpoint value, time to event), method of aggregation (mean, proportion), and time point. Five components per outcome. This is the most consequential item in the checklist: an outcome missing any of them cannot be held to account at analysis time.
17
Harms
How harms are defined and how they will be assessed — systematically (solicited, instrument-based) or non-systematically (spontaneous report). The 2025 revision brings the CONSORT Harms emphasis into the protocol: a trial that only plans to measure benefit has planned half a trial.
18
Participant timeline
The time schedule of enrolment, interventions, and assessments — the signature SPIRIT item, with a schematic diagram strongly recommended. The schedule figure is the one page of the protocol every coordinator, monitor, and reviewer actually keeps open.
19
Sample size
How the sample size was determined, including all assumptions supporting the calculation — effect size, variability, alpha, power, allocation ratio, attrition inflation, and where each number came from. A stated N without auditable assumptions is a guess wearing a calculation's clothes.
20
Recruitment
The strategies for achieving adequate enrolment — sources, settings, and any planned facilitation. Under-recruitment is the most common way trials fail; SPIRIT asks for the plan before the failure, not the excuse after it.
Methods — assignment of interventions · items 21–24
21
Randomisation: sequence generation
Sub-items 21a–21b: who generates the allocation sequence and by what method; the type of randomisation (simple or restricted), any stratification factors, and — to protect concealment — block details held in a separate secure document rather than the protocol itself.
22
Allocation concealment mechanism
The mechanism used to implement the sequence — central computer or telephone randomisation, sequentially numbered opaque sealed containers — and the steps taken to conceal the sequence until interventions are assigned. Concealment is what stands between randomisation and selection bias; it must be operational, not aspirational.
23
Implementation
Who enrols participants, who assigns them to interventions, and what access those people have to the allocation sequence. The division of labour is the point: the person enrolling should not be the person who can foresee the next allocation.
24
Blinding
Sub-items 24a–24c: who will be blinded after assignment (participants, care providers, outcome assessors, analysts); how blinding will be achieved, including the similarity of the interventions; and the circumstances and procedure for unblinding. "Double-blind" without naming who is blinded and how is a label, not a method.
Methods — data collection, management, analysis · items 25–27
25
Data collection methods
Sub-items 25a–25b: the plans for assessing and collecting data, the processes promoting data quality, descriptions of instruments with their known reliability and validity, and where the data collection forms can be found; plus retention and follow-up plans — including collecting outcome data from participants who discontinue the intervention, who remain in the trial.
26
Data management
Plans for data entry, coding, security, and storage, with reference to where the detailed procedures live. Includes who will have access to the trial dataset — merged in 2025 from what was a standalone item.
27
Statistical methods
Sub-items 27a–27d: the methods for comparing groups on primary and secondary outcomes and harms; the definition of the analysis populations and how participants are assigned to them; how missing data will be handled; and any additional analyses — subgroup and sensitivity — specified in advance, which is the only time a subgroup analysis can be specified credibly.
Methods — monitoring · items 28–29
28
Data monitoring committee
Sub-items 28a–28b: the committee's composition, role, reporting structure, and independence — or an explanation of why no DMC is needed; and the plans for interim analyses and stopping guidelines, including who can see interim results and who decides to stop.
29
Trial monitoring
New in 2025, replacing the old auditing item: the frequency and procedures for monitoring trial conduct — or a statement that there is none and why. Monitoring is where protocol promises meet trial reality; the protocol says how that meeting is scheduled.
Ethics and other considerations · items 30–34
30
Research ethics approval
The plan for seeking research ethics committee / institutional review board approval. The protocol is the document the committee approves — this item closes the loop by stating which committees, and when.
31
Protocol amendments
Plans for communicating important modifications — to eligibility, outcomes, analyses — to the relevant parties: investigators, ethics committees, registries, regulators. An amendment that changes the primary outcome and reaches the registry late is indistinguishable from outcome switching.
32
Consent or assent
Sub-items 32a–32b: who will obtain informed consent or assent and how; and, where applicable, the provisions for ancillary studies that will use participant data or biological specimens — consent for the trial does not automatically cover the biobank.
33
Confidentiality
How personal information will be collected, shared, and maintained to protect confidentiality — before, during, and after the trial. This is the operational counterpart of the data-sharing plan in item 6: sharing de-identified data responsibly presumes confidentiality was engineered from the start.
34
Ancillary and post-trial care
Provisions for ancillary and post-trial care, and for compensation for those harmed by trial participation. What happens to participants when the trial ends — continued access, transition of care, harm compensation — is a protocol commitment, not a discretionary courtesy.
The contrast

Adequate vs. inadequate reporting.

Three of the SPIRIT items most consistently under-specified in submitted protocols. Each shows the version that fails review next to the version that passes — with the reasoning an ethics committee, statistician, or reviewer would use.
Outcome specification — item 16
Inadequate
"The primary outcome will be improvement in knee pain. Secondary outcomes will include function and quality of life."
Adequate
"Primary outcome: change in the WOMAC pain subscale (0–100, higher = worse) from baseline to 12 weeks, analysed as the between-group difference in means (ANCOVA adjusted for baseline score). Secondary outcomes: WOMAC function subscale (same metric and time point); EQ-5D-5L index change at 12 and 26 weeks (mean difference); OMERACT-OARSI responder status at 12 weeks (proportion)."
Why it mattersSPIRIT item 16 requires each outcome to carry five components: the specific measurement variable, the analysis metric, the aggregation method, and the time point, for both primary and secondary outcomes. "Improvement in knee pain" fixes none of them — which instrument, changed from when to when, summarised how? An outcome left that loose at the protocol stage cannot be checked at the analysis stage, and any of a dozen analyses could later be presented as the planned one. The adequate version is auditable: anyone can compare the eventual report against it line by line.
Sample size assumptions — item 19
Inadequate
"We plan to recruit 200 patients (100 per arm), which the investigators consider sufficient to answer the study question based on experience with similar trials."
Adequate
"Sample size: detecting a between-group difference of 8 points on the WOMAC pain subscale (minimal clinically important difference; SD 20, from our pilot of 46 patients) with two-sided α = 0.05 and 90% power, allocation 1:1, requires 133 per arm. Inflating for 15% attrition (the rate in our pilot at 12 weeks) gives 157 per arm — 314 in total."
Why it mattersSPIRIT item 19 requires how the sample size was determined and all assumptions supporting the calculation. "The investigators consider it sufficient" contains no effect size, no variability estimate, no error rates, and no source for any of them — a reviewer cannot distinguish it from a number chosen for feasibility. The adequate version states every assumption and where each came from, which is precisely what makes it checkable: change any input and the required N changes visibly, rather than silently.
Data sharing — item 6
Inadequate
"Trial data will be made available upon reasonable request to the corresponding author."
Adequate
"De-identified individual participant data, the data dictionary, and the statistical analysis code will be deposited in Vivli within 12 months of the primary publication, available under managed access to researchers with an approved analysis proposal. The protocol and statistical analysis plan will be published as open-access supplements with the primary report."
Why it mattersItem 6 sits in SPIRIT 2025's new open-science section, and it asks for a plan: what will be shared, where, when, and under what conditions. "Available upon reasonable request" specifies none of these — no repository, no timeline, no access mechanism, and no definition of "reasonable" — so it commits the trial to nothing a reader could later hold it to. The adequate version names the repository, the timing, the access model, and covers the protocol and SAP required by item 5. Funders and journals increasingly read this item first among the open-science commitments.
The check

How PeerReviewAI evaluates SPIRIT compliance.

SPIRIT compliance is checked on every trial protocol submitted for AI Peer Review ($49) — included, no add-on required.
01
Auto-detected as a trial protocol
Upload your protocol and it is identified automatically: future-tense methods, planned analyses, and the absence of results distinguish a protocol (SPIRIT) from a completed trial report (CONSORT). You do not have to choose the checklist.
02
Item-by-item qualitative assessment
Every SPIRIT item is evaluated against your protocol text. The output is a qualitative judgment per item — adequate, incomplete, or missing — across the full checklist, including the lettered sub-items.
03
The load-bearing items get scrutiny
The review reads a protocol the way a methods reviewer does: are the outcomes fully specified with variable, metric, aggregation, and time point; are the sample-size assumptions stated and sourced; is allocation concealment operational; do the open-science items commit to anything checkable.
04
Included in Peer Review
SPIRIT compliance checking is part of Peer Review ($49) — no add-on required: the full per-item checklist table, plus a compliance audit against your target journal's author instructions. The $2.99 Journal Brief does not include checklist compliance.
Related guidelines

If your document is not a trial protocol.

CONSORT 2025
Randomized controlled trials
The other half of the pair: when the trial is finished, the report is written against CONSORT 2025 — deliberately harmonised with SPIRIT. Read the deep-dive on CONSORT.
Read more
PRISMA 2020
Systematic reviews
Protocols for systematic reviews follow PRISMA-P rather than SPIRIT; the completed review follows PRISMA 2020. Read the deep-dive on PRISMA.
Read more
Hub
All reporting guidelines
Not sure which guideline applies? The hub maps every study type to its checklist — CONSORT, STROBE, PRISMA, STARD, ARRIVE, and 15+ more.
Read more
SPIRIT · checked on every trial-protocol review

Check your protocol against SPIRIT 2025.

FAQ

Questions, answered.

Don't see yours? Email us — we read every one.