How to register your protocol.

Identify the document as a protocol and state the review design. Include the principal topic and population or research context, and identify an update of an earlier review when relevant. Mention meta-analysis or network meta-analysis when it is planned, while acknowledging in the methods that synthesis depends on the available evidence. Use sentence case and avoid claims about results before the review is conducted.

Provide the author’s full name, an actively monitored email address and institutional affiliation. Add the correct ORCID iD where available. This person coordinates communication about the record and its updates. Confirm that all listed authors have reviewed the submission and that their names, affiliations and contact details are accurate.

Identify financial and nonfinancial support, including grants, institutional support, supplied software or services, and relevant award numbers. If there is no specific funding, state this. Describe the sponsor’s role, or absence of a role, in designing, conducting, interpreting and disseminating the study. Funding information does not replace the separate competing-interests declaration.

Choose the actual stage reached when submitting, such as preliminary searches, formal screening, data extraction, risk-of-bias assessment, analysis or completion without publication. Add dates and explanatory details when necessary. Disclose delayed registration and which activities or decisions preceded it. This field records timing at submission; Item 33 describes the project’s current overall status.

Identify the institution or organisation responsible for the work, with department where relevant, city and country. Use the official institutional name. Provide additional affiliations in the relevant author fields or other-information section when the form permits. Do not substitute a postal address for an affiliation or assume that every author has the same institution.

Declare relevant financial and nonfinancial interests for the authors, including relationships with products, organisations or studies being evaluated. Explain how potential conflicts will be managed, such as assigning assessment of an author’s own study to other reviewers. If no relevant conflicts exist, state this explicitly. A funding declaration alone is insufficient.

If providing this optional contact detail, include the international country and area codes. Supply only an appropriate professional contact number. Check the form’s privacy information before entering personal contact details; do not assume that information submitted to a registration service is necessarily confidential.

State a focused, answerable objective. Use a framework suited to the design: PICO for many intervention questions, an exposure-based formulation where appropriate, or population, concept and context for scoping reviews. Specify the primary question and any secondary questions. Assess feasibility without changing the question merely to secure a favourable result or a larger pool of studies.

Explain the evidence gap and the contribution expected from the review. Identify relevant completed or ongoing reviews and explain any justified overlap, update or methodological difference. A small or empty evidence base can itself be informative; avoid altering the scope solely to prevent an empty review. Cite the key sources supporting the rationale.

Describe the condition, phenomenon, methodological problem or subject area. Include definitions needed to interpret the eligibility criteria. The description may concern health, sport, education or another eligible topic; a disease label is not necessary for every design. Define specialised terminology and distinguish the topic from the specific review objective.

Provide reproducible planned search strings, ideally for every database and at least a complete draft for one principal database. Identify the database and interface separately, such as MEDLINE via PubMed. Report controlled vocabulary, free-text terms, operators, limits and planned search dates or coverage. Describe updating, supplementary searching and search peer review where planned. Retain the executed strategies and dates for the final report. [12]

Specify the population, setting and relevant characteristics, including how mixed populations will be handled. Define eligible subgroups and justify restrictions. In reviews where the units are studies, documents, institutions or other research objects, describe those units explicitly. Use criteria precise enough for different reviewers to apply consistently.

Describe the intervention, exposure, index test, prognostic factor or other focal element as appropriate. Include eligible variants, timing, intensity and relevant co-interventions. If the review does not evaluate an intervention, explain the relevant concept rather than forcing an intervention framework. State “Not applicable” with a reason where appropriate.

Define the eligible comparator or reference condition and any acceptable variations. For diagnostic reviews, describe the reference standard where relevant. For network meta-analysis, explain how treatments will be grouped into nodes. When the review question has no comparator, state this and explain why.

Name the eligible study designs and relate their inclusion to the question. Randomized trials are often appropriate for intervention effects, while observational and other designs may address different questions. Do not use a universal evidence hierarchy to exclude designs needed for the objective. State how different designs will be analysed or kept separate.

Combine the question framework, designs, setting, dates, language and publication-status criteria into an operational definition. Justify restrictions. Explain how multiple reports of one study, overlapping samples and uncertain eligibility will be handled. Prespecify decisions and document later changes with their timing and rationale; do not silently alter criteria after examining study findings.

List the databases, registries, websites, repositories and other sources to be searched. Include citation searching, reference-list checks, grey literature and contact with investigators when planned. State coverage and planned update searches. Choose sources appropriate to the subject rather than reproducing a standard list that omits specialist literature. [12]

Define each primary outcome or principal variable, its measurement, eligible time points and any hierarchy for selecting among instruments or analyses. For intervention reviews, consider benefits and harms. Justify surrogate outcomes and acknowledge their limitations. For scoping or methodological projects, define the main concepts, characteristics or indicators to be charted or evaluated instead of imposing clinical outcomes.

Prespecify secondary outcomes or additional variables and explain their relevance. Define measurement and time points where applicable. Clarify how multiple outcomes or analyses will be prioritised and reported. Secondary outcomes should not be restricted to those that support a preferred interpretation of the primary findings.

Describe deduplication, screening, data extraction or charting, piloting and disagreement resolution. State the number of reviewers at each stage and whether decisions are independent or checked. Plan how multiple reports will be linked to one study, missing information sought, and full-text exclusions recorded. Identify software and versions, data storage and planned sharing. If automation or AI is used, describe its role and human verification as set out in Section 7.

Distinguish assessment of bias in individual study results from assessment of certainty across a body of evidence. Select design-appropriate tools, name their versions and describe assessors, independence and disagreement resolution. Explain how judgments will affect the synthesis. Where appropriate, specify a certainty framework such as GRADE by outcome. GRADE does not replace study-level risk-of-bias assessment, and it is not a universal requirement for every scoping or methodological study. [13,14]

Explain when studies will be combined, the planned effect measures and statistical model, and how differences between studies will be assessed. Prespecify handling of missing data, dependent estimates, multi-arm studies and conversions when relevant. Do not choose the model solely from a heterogeneity significance test. Describe the synthesis when meta-analysis is inappropriate. For network meta-analysis, address transitivity, inconsistency and the interpretation of rankings. For qualitative or scoping work, specify coding, charting and synthesis methods. [13]

Prespecify a limited set of plausible effect modifiers, the rationale for each and the planned comparison method. Where relevant, describe interaction tests and any meta-regression. Consider whether sufficient data are likely to be available. Label analyses introduced after the protocol as exploratory and avoid interpreting a difference between significant and nonsignificant subgroup results as proof of an interaction.

Describe analyses that will test the effect of key decisions or assumptions, such as excluding studies at high risk of bias, using alternative models or changing assumptions about missing data. Distinguish these robustness checks from subgroup analyses. Identify which analyses were prespecified and explain any additional checks performed during the review.

State whether searches and study inclusion will have language restrictions and justify any restrictions. Describe translation arrangements where relevant. The language of an eligible study is separate from the platform’s requirement to submit the protocol in English. Retain original-language search expressions when needed to make database searches reproducible.

List the countries in which the review team is conducting the project. These describe the research team, not necessarily the countries of the studies to be included. Put geographical restrictions on eligible evidence in the eligibility criteria instead. Keep this information consistent with the authors’ affiliations.

Use this field for details that do not fit elsewhere, such as related registrations, links to supplementary methods, repositories, important dates or explanations of delayed registration. Include references in a consistent citation style. Identify related or earlier protocols and explain their relationship. Provide links to accessible supporting material without exposing confidential information.

Choose terms that describe the topic, population or context, study design and principal methods. The existing INPLASY guidance recommends four to eight keywords or phrases separated by semicolons. Use terms that readers are likely to search and avoid irrelevant terms added only to increase visibility.

Describe how findings will be shared, such as through a journal article, preprint, conference, repository or accessible summary. Include plans to share data, extraction forms and analysis code where appropriate. Explain how the final publication will be linked to the registered protocol and how protocol deviations will be reported.

Enter each author’s full name, email where requested, ORCID iD where available and contribution. Use consistent spelling and confirm authorship and contribution statements with the team. Contributions can be described using clear roles, including design, searching, screening, extraction, analysis and writing. Declare relevant use of AI as a tool; responsibility for the content remains with the human authors.

List all authors in the agreed order and follow the form’s surname-and-initials instructions. Check compound surnames, particles, accents and culturally appropriate name structure with the individual authors. Do not infer name components from nationality. Ensure that the abbreviated list corresponds exactly to the full names provided in Item 30.

Provide the email address requested for invoices and payment confirmation. Keep it accurate and monitor it after submission. Distinguish billing communication from the corresponding-author contact published with the record. Keep the submission receipt and check spam or junk folders if an expected message is missing.

Choose the status that accurately describes the project. After registration, report when it is discontinued, completed without publication, or completed and published using the free status-update form. Supply the review or preprint URL when available and clearly identify a preprint as such. A change of status does not replace an amendment describing substantive changes to the protocol. [6]