Template
Write the protocol before you screen a single study
You are planning a systematic review and need the method fixed and registered before the results can influence your decisions.
- Typical length
- Usually 3,000–6,000 words. Registration records are shorter because they use fixed fields, while a protocol submitted to a journal carries the full narrative.
- Usual tense
- Future tense throughout — the protocol describes work not yet done. Past tense appears only for scoping searches already completed.
- First person
- Common in the plural (“we will search”), though many protocol templates prefer impersonal constructions. Follow the register or journal template.
- Reporting standard
- PRISMA-P sets out the items a protocol should report; the completed review is then reported against the PRISMA statement. Both checklists are freely available.
- Registration
- PROSPERO registers health-related reviews and is the most widely used option; other fields use OSF Registries or discipline-specific registers. Register before screening begins.
- Check with your team
- Whether an information specialist will build the search, who screens in duplicate, and which appraisal tool matches the study designs you expect.
THE SKELETON
Section by section
Each block gives you the heading, what it has to accomplish, and the questions that fill it.
1
Administrative Information
Title, authors and their roles, contact details, funding, competing interests, and the registration record. The bureaucratic front page that makes the protocol citable.
- Does the title identify the report as a protocol for a systematic review?
- Which register will you use, and have you started the record?
- Who is guarantor, and what is each author contributing?
- Which funder supports the review, and does it have any role in the process?
2 · about 15% of the protocol
Background and Rationale
Why the review is needed: what is known, what is contested, and whether an existing review already answers the question or has gone out of date.
- What does existing evidence suggest, and why is it insufficient?
- Have you searched the registers for a review already underway on this question?
- What decision would a completed review inform, and for whom?
3
Objectives and Review Question
The question stated in a structured form — PICO, PICo, SPIDER, or the framework your field uses — so that eligibility criteria follow mechanically from it.
- What are the population, intervention or exposure, comparator, and outcomes?
- Which outcome is primary, and which are secondary?
- Are there planned subgroup questions, and are they specified now rather than later?
4 · about 15% of the protocol
Eligibility Criteria
Exactly what makes a study includable: design, population, intervention, comparator, outcomes, setting, time period, language, and publication status.
- Which study designs are eligible, and which are excluded?
- Are you restricting by language or date, and can you justify it?
- Will you include preprints, theses, conference abstracts, or unpublished data?
- How will you treat studies that report the outcome only partially?
5 · about 15% of the protocol
Information Sources and Search Strategy
Every database, register, and supplementary source, with the full search string for at least one database and the date the search was last run.
- Which databases, and over which coverage dates?
- What is the full Boolean string for your primary database, including controlled vocabulary terms?
- Will you hand-search key journals, check reference lists, or contact authors?
- Has an information specialist reviewed the strategy?
6 · about 15% of the protocol
Study Selection and Data Extraction
How records move from search results to included studies, and what is pulled from each one. Independent duplicate work at each stage is the standard expectation.
- How many reviewers screen titles, abstracts, and full texts, and how are disagreements resolved?
- Which software will manage deduplication and screening?
- Which fields will be extracted, and has the form been piloted?
- How will you handle multiple reports of the same study?
7
Risk of Bias and Certainty Assessment
The tools used to appraise included studies and to rate confidence in the body of evidence, plus who applies them and how disagreements are settled.
- Which appraisal tool fits your included designs?
- Will assessment be done independently in duplicate?
- How will risk of bias feed into the synthesis — narrative comment, sensitivity analysis, or exclusion?
- Will you rate certainty of evidence, and with which approach?
8 · about 15% of the protocol
Synthesis Plan
What you will do with the extracted data, including the conditions under which you would pool results and the conditions under which you would not.
- Is meta-analysis planned, and what degree of heterogeneity would rule it out?
- Which effect measure and which model?
- How will heterogeneity be assessed and explored?
- What is the narrative synthesis plan if pooling is inappropriate?
- How will reporting bias across studies be investigated?
9
Timeline, Dissemination, and Amendments
Project dates, where results will be published, and the commitment to log any protocol change with its date and reason.
- What are the expected dates for search, screening, extraction, and write-up?
- Where will the completed review be published or deposited?
- How will amendments be recorded in the registration record?
SENTENCE STARTERS
Phrasing you can adapt
Lift these and change the specifics. Academic phrasing is a shared convention, not something each writer has to reinvent.
This review will examine whether … compared with … affects … in …
The primary objective is to synthesize evidence on …
Secondary objectives are to determine whether the effect varies by …
The review question is framed using the PICO structure as follows.
MEDLINE, Embase, and … will be searched from inception to …
The search will combine controlled vocabulary terms with free-text synonyms for …
Reference lists of included studies will be screened for additional records.
No language restriction will be applied; non-English records will be translated where necessary.
Where studies are sufficiently homogeneous, results will be pooled using a random-effects model.
Heterogeneity will be assessed using …, with substantial heterogeneity prompting …
If pooling is inappropriate, findings will be synthesized narratively and presented in structured tables.
Sensitivity analyses will exclude studies at high risk of bias.
Questions about the systematic review protocol
Why register a protocol at all?
It fixes your decisions before you see the results, which is what stops a review from drifting toward the conclusion its authors already held. It also prevents duplicated effort — other teams can see the question is being addressed.
What if I need to change the method after registering?
Change it, and record the amendment with its date and reason in the registration record and the final report. Undisclosed deviation is the problem; documented deviation is normal practice.
How is this different from a scoping review protocol?
A scoping review maps what exists on a broad topic and does not usually appraise study quality or pool results, so its protocol has broader eligibility and no meta-analysis plan. A systematic review answers a narrow question and commits to appraisal and synthesis.
Can one person do a systematic review?
You can, but screening and extraction are expected to be done independently in duplicate, and a solo review will be judged accordingly. If you have no second reviewer, say so in the protocol and describe what you did instead.
How long does a systematic review take?
Commonly many months from protocol to submission, with screening the largest single cost. Build the timeline around the number of records you expect the search to return rather than around your deadline.
Related templates
Literature Review Outline
You are writing a standalone review — for a course, a thesis chapter, or a review journal — and you have far more sources than shape.
Methodology Section Template
You ran a study and now have to describe it so precisely that a reviewer can evaluate it and a stranger could repeat it.
Dissertation Proposal
You are asking a committee to approve a study you have not run yet, and the document has to convince them it is worth doing and that you can do it.
Annotated Bibliography
You have been asked to show not only what you have read but what you made of it, usually before a longer piece of writing is due.
An outline is a start
Plan mode is the rest
Copyeditor’s Plan mode turns a structure like this into sections you produce one at a time, with citations landing in the bibliography as you go.
