Contents
On this page
- Scope and limits of the PICO specification
- The three levels at which PICO operates
- Defining the population without presupposing the answer
- Specification of the intervention and comparator
- Pre-specification and ranking of outcomes
- PICO variants and alternative element sets
- Translation of PICO into applicable screening criteria
- Placement and timing of the PICO specification
- Limitations of the PICO framework
- Key points
- Application to your study
- Frequently asked questions
Academy · Research Data Analytics & AI · PICO framework
Framing a Systematic Review Question with PICO
PICO (Population, Intervention, Comparison, Outcome) is not a title formula but a specification that must be written at three distinct levels: the review’s planned scope, the question that each synthesis answers, and what the included studies did. Most review protocols address only the first level, and the gap between the levels is a source of eligibility disputes, post hoc subgrouping and unexplained heterogeneity.
Population, Intervention, Comparison and Outcome are four elements that are quickly learned, yet the PICO framework remains the most common point at which a systematic review goes wrong. The difficulty is rarely that authors do not know the mnemonic; rather, they write a PICO loose enough to secure general agreement and discover at screening that it cannot decide a single borderline record. A usable PICO is one that two screeners working separately apply to the same hundred abstracts and reach nearly the same answer. This article is organized around that test.
PICO for a systematic review at a glance
- Elements
- Population, Intervention, Comparison, Outcome, often with S for study design28
- Written at
- Three levels: the review scope, each planned synthesis, and the included studies2
- The test
- Two screeners applying it separately to the same records reach nearly the same decisions
- Most often missing
- A named comparator, an outcome time point, and a default for records that are silent
- Recorded where
- A registered protocol, before screening begins6
- Reading time
- About 14 minutes, with 3 figures and 5 tables
Learning objectives
- Formulation of a review question at all three levels distinguished by the Cochrane Handbook, and the reasons the third level explains heterogeneity
- Translation of each PICO element into an eligibility criterion that a screener can apply without further clarification
- Explicit specification of the comparator, the element omitted most often and the one that most changes what the review can conclude
- Pre-specification and ranking of outcomes, including those expected to be poorly reported
- Selection between PICO and its variants on the basis of review type rather than habit
- Recognition of the questions that PICO cannot frame, and the structures to use instead
Common misconceptions
- “PICO is a way of writing the review title.” PICO is a specification for eligibility decisions. A PICO that cannot resolve a borderline abstract serves no functional purpose, however well it reads at the top of the protocol.
- “One PICO per review.” The Handbook distinguishes three levels at which PICO operates: the review PICO that sets the scope, the PICO of each individual synthesis, and the PICO of the studies actually included. When these levels are conflated, reviews pool studies that answer different questions.
- “The comparator is obvious, so it need not be stated.” Usual care, placebo, no treatment, waiting list and an active alternative represent four or five different questions with different answers. If the comparator is left unspecified, the included studies will define it, and will do so inconsistently.
- “A broader scope is safer, because it can always be narrowed later.” Narrowing the scope after the yield has been seen is a post hoc decision, and reviewers will interpret it as such. Breadth is a legitimate choice, but it must be declared in advance and justified on its own terms.
- “PICO fits any review.” Published typologies of review types extend well beyond the intervention review for which PICO was designed; one widely used analysis describes 14 distinct review types, each with its own methodology. Diagnostic accuracy, prognosis, qualitative and methods reviews each require a different set of elements.
Scope and limits of the PICO specification
PICO specifies four elements: the Population in which the question is asked, the Intervention whose effects are of interest, the Comparison against which that intervention is evaluated, and the Outcomes by which the comparison is judged 28. The current edition of the Cochrane Handbook for Systematic Reviews of Interventions, version 6.5 published in 2024, treats these as the elements that encapsulate a review question for an intervention review 2,3.
PICO does not determine any aspect of method. It does not specify how to search, who screens, which risk-of-bias tool applies or whether to pool. Nor does it, by itself, generate eligibility criteria; it generates the four headings under which eligibility criteria must be written. That translation step is the one most protocols omit, and it is where the substantive work lies. “Adults with type 2 diabetes” is a population; “adults aged 18 or over with a clinical diagnosis of type 2 diabetes recorded by the study investigators, excluding studies in which more than 20 per cent of participants had type 1 diabetes and the subgroups were not reported separately” is an eligibility criterion.
The practical test of a PICO is not whether it reads well but whether two people can apply it independently to the same set of records and agree. If they cannot, the disagreements are not a screening problem to be arbitrated later; they are a specification problem to be resolved before screening begins. Our page on title and abstract screening describes the calibration exercise that reveals such problems.
The three levels at which PICO operates
This distinction has the greatest effect on how a protocol is written, and it is frequently overlooked. The Handbook separates three levels 2,3.
The review PICO is the planned scope: the broadest set of populations, interventions, comparators and outcomes the review will admit. The eligibility criteria are constructed from it, and the search must cover it 13,26.
The PICO for each synthesis is narrower. It is the question that a particular analysis answers: one forest plot, one pooled estimate, one row of a summary of findings table. A review whose scope admits three drug classes and four comparators does not have a single question but a structured set of questions, and the protocol should state which analyses are planned and what each addresses.
The PICO of the included studies describes what the studies actually did: the populations recruited, the doses given, the comparators used, and the outcomes measured and their timepoints. This level is descriptive rather than planned, and it is the level that explains the results. When a pooled estimate is more heterogeneous than expected, the explanation is usually apparent in the gap between the first and third levels: the review admitted a range of studies that the synthesis then treated as interchangeable.
Writing all three levels in the protocol has a specific benefit: it makes the characteristics-of-included-studies table a planned output rather than an afterthought, and it indicates in advance which study features must be extracted to explain heterogeneity later. See what data to extract for the extraction consequences.
The three levels of PICO, and what each one is for
| Level | What it specifies | Written when | What it governs |
|---|---|---|---|
| Review PICO | The broadest populations, interventions, comparators and outcomes the review will admit | In the protocol, before screening | Eligibility criteria and the scope the search must cover |
| PICO for each synthesis | The specific question one analysis answers: one forest plot, one row of a summary of findings table | In the protocol, as a planned set of analyses | Which studies are grouped together, and what each pooled estimate means |
| PICO of the included studies | What the studies actually recruited, delivered, compared and measured | Descriptively, during extraction | The characteristics table, and the explanation of heterogeneity |
Structure as distinguished in the Cochrane Handbook for Systematic Reviews of Interventions, version 6.5 (2024).
The structure at which PICO is specified, as set out in Chapter 3 of the Cochrane Handbook for Systematic Reviews of Interventions, version 6.5 (2024).
One review question at the three levels
| Element | Review scope | One synthesis | One included study |
|---|---|---|---|
| Population | Adults aged 60 or over with musculoskeletal pain for 3 months or more | The same population | Women aged 65 to 80 with knee osteoarthritis |
| Intervention | Exercise programmes of 4 weeks or more | Any eligible programme | 12 weeks of twice-weekly strength classes |
| Comparison | None, waiting list, usual care or attention control | Usual care only | GP advice and analgesia as needed |
| Outcome | Pain and function (critical); adverse events, withdrawals | Pain intensity nearest 12 weeks | Validated pain subscale at 12 and 26 weeks |
| Study design | Parallel and cluster randomised trials | The same designs | Two-arm parallel randomised trial |
Illustrative example written for this guide, built on the eligibility specification later in this article. The third column shows the kind of detail that is extracted, not a real study.
Defining the population without presupposing the answer
Three failures recur. The first is defining the population by a feature that is in fact an outcome: “patients who responded to initial therapy” makes eligibility conditional on something the intervention may have caused, and a review constructed in this way cannot answer an effectiveness question. The second is a definition that depends on information that abstracts never contain, such as a biomarker threshold or a staging subcategory, which forces every record to full-text review and multiplies the screening workload without any gain in precision. The third is leaving the setting and care context implicit, so that primary care and tertiary referral populations are combined in the same analysis without any explicit decision that they should be.
Population that presupposes the answer
“Patients with depression who responded to initial therapy.”
Eligibility depends on something the intervention may have caused, so the review cannot answer an effectiveness question.
Population a screener can apply
“Adults aged 18 or over with major depressive disorder diagnosed by a structured clinical interview, in primary or secondary care.”
Condition, ascertainment, age and setting are fixed before any outcome is known, and each can be read from an abstract.
The population should be specified in three parts, the last of which, the rule for mixed populations, prevents the most disputes 2. Any of the handling rules is defensible; the absence of a rule is not.
Population specification checklist
- The condition, and how its presence was ascertained.
- The bounding characteristics: age, severity, comorbidity, setting, and country or health system where relevant.
- The mixed-population rule, stated in advance: include if a stated proportion meets the definition, include only if the relevant subgroup is reported separately, or exclude.
- A default for records that do not report the information: exclude, or take to full text.
Where the population is defined by a diagnostic process rather than a label, the acceptable reference standard should be stated. That single sentence prevents a review of a condition from becoming, unnoticed, a review of a condition as diagnosed by whatever method each study happened to use.
Specification of the intervention and comparator
The intervention should be specified at the level of detail that could change the effect: the agent or procedure, the dose or intensity, the route, the schedule, the duration, the provider and their training, and the permitted co-interventions. For complex and behavioural interventions this list becomes longer, not shorter, and a structured intervention description framework is useful for keeping it consistent across studies 22.
The comparator is the element at which reviews most often lose the ability to reach a conclusion. “Compared with control” is not a comparator. Placebo, no treatment, waiting list, usual care, attention control and an active alternative represent distinct questions, and they do not yield the same answer. Usual care is the most problematic, because it differs across health systems and decades, and a review that pools against “usual care” without describing it has pooled across an unmeasured intervention.
Three decisions should be made and recorded: which comparators are eligible; whether studies with multiple eligible arms will contribute all arms, one arm or a combined arm, and if combined, how; and whether comparisons between two active interventions are in scope, because that decision determines whether the review is a pairwise review or a network meta-analysis, which is a different design with different assumptions and cannot be adopted retrospectively.
Pre-specification and ranking of outcomes
Outcomes should be selected because they matter to patients and decision-makers, not because they are the outcomes that studies report 2. These two sets differ, and the difference is a finding in its own right; a review that reports it provides the reader with useful information about the field.
A small number of critical outcomes and a slightly larger number of important outcomes should be specified and ranked in the protocol 6. The ranking has substantive consequences: it determines which outcomes drive the summary of findings table, which undergo certainty assessment, and which can be given lower priority without the reader suspecting that the choice was made after the results were seen. Harms expected to be poorly reported should be included; omitting them because data will be sparse ensures that the review can say nothing about them.
For each outcome, the protocol should state how it is defined, how it is measured and at what timepoint or window, and how a study reporting the same construct on a different instrument or at a different time will be handled 23. A review with “pain” as an outcome and no rule on instrument or timepoint will produce a pooled estimate that no one, including its authors, can interpret. Each outcome to be extracted also has a fixed internal structure, which is described in what data to extract, and the numerical form to which it must be reduced is described in extracting quantitative data.
PICO variants and alternative element sets
PICO has accumulated two decades of variation and application, and the variants are not interchangeable stylistic choices; each adds an element because a particular review type required it 12. Adding S for study design turns PICO into an eligibility filter; adding T for timeframe forces the timepoint decision that the outcome element often leaves vague. Frameworks developed for qualitative and service-delivery questions substitute elements for context, setting, perspective and phenomenon of interest, because there is no intervention-and-comparator structure to describe.
Diagnostic accuracy reviews require a genuinely different set of elements: the population, the index test, the reference standard against which it is judged, and the target condition being detected 10. Forcing this structure into PICO produces the familiar error of treating the reference standard as a “comparator”, which it is not; it defines the truth rather than constituting an alternative intervention. Prognosis and prediction-model reviews require the population, the predictors or model, the outcome being predicted and the horizon over which it is predicted.
More generally, published typologies of review methodology describe many distinct review types (one widely cited analysis sets out 14, each with its own methods), and PICO was designed for one of them 25. Selecting an element set that fits the question is not a deviation from good practice; applying PICO to a question it does not fit is.
How published reviews state their PICO or PICOS
| Review | Population and intervention | Comparison | Outcome and study design |
|---|---|---|---|
| Zhang et al., 2026, Frontiers in Physiology29 | P: Insufficiently active older adults I: Short bouts of exercise done remotely (“exercise snacking”) | No specific intervention | O: Lower-limb muscle function (primary); balance, adherence or acceptability (secondary) S: Randomised crossover or randomised controlled trials |
| Li et al., 2025, Frontiers in Pharmacology30 | P: Patients with herpes zoster or postherpetic neuralgia I: A Chinese polyherbal preparation as adjunctive therapy | Conventional treatment only | O: Primary: vesicle cessation, scab formation, pain score and postherpetic neuralgia incidence S: Randomised controlled trials |
| Terzoni et al., 2026, Neurourology and Urodynamics31 | P: Patients undergoing radical prostatectomy I: Preoperative pelvic floor muscle training | Any postoperative intervention or monitoring | O: Continence recovery S: Not part of the stated PICO |
| Li et al., 2026, BMC Surgery32 | P: Elderly patients with intertrochanteric femoral fractures I: Proximal femoral bionic nail | InterTAN nail | O: Operative time, blood loss, hospital stay, weight-bearing time, Harris Hip Score, fixation complications S: PICOS cited; the 6 included studies were 2 prospective and 4 retrospective cohorts |
Each row is taken from the review’s published abstract, retrieved from PubMed in November 2026. A selection, not a complete list. Note how far the comparators differ in breadth: “any postoperative intervention or monitoring” is a defensible scope only because it was declared.
Should every element become a search block?
Not necessarily. The eligibility criteria need all the elements; the search does not. Eriksen and Frandsen’s systematic review found only three studies that tested PICO as a search tool, too different to pool. Across them, adding concept blocks reduced the number of records retrieved and lowered sensitivity, which is the expected trade-off. The common advice to leave outcome terms out of the search, because outcomes are often missing from abstracts, is reasonable but was not substantiated by the limited data they found34. Decide the search blocks separately from the eligibility criteria, and test the search against a set of known relevant studies.
Translation of PICO into applicable screening criteria
The critical protocol step is the conversion of the four elements into a decision rule 2,27. A workable rule contains, for each element, an inclusion statement, an exclusion statement where the inverse is not obvious, and a default for cases in which the record is silent. The default column is what prevents two screeners from inferring the team’s intention differently.
Two practical principles apply. At the title and abstract stage, each criterion should be answerable from an abstract or else explicitly deferred to full text; a criterion that cannot be assessed at screening should be designated a full-text criterion rather than applied tacitly by whichever screener knows the literature. In addition, the number of criteria should be small enough to be held in mind; if a screener must consult a two-page document for every record, the rule set requires restructuring rather than additional training.
The rules should be piloted on a set of records assessed by both screeners, the results compared, and the rules rather than the screeners revised. Disagreement at this stage is information about the specification. Our page on handling discrepancies describes how to handle disagreements that persist after the pilot.
Converting each element into a screening rule
| Element | The question the rule must answer | The default that must be stated |
|---|---|---|
| Population | Which condition, ascertained how, in which setting and age range | How a mixed population, and a record that does not report the relevant information, should be handled |
| Intervention | Which agent or procedure, at what dose, intensity and duration, and delivered by whom | Whether co-interventions are permitted, and whether dose is a criterion or a characteristic |
| Comparison | Which comparators are eligible, named individually rather than as “control” | How multi-arm studies contribute, and whether active comparisons are in scope |
| Outcome | Which critical and important outcomes, on which instruments, at which timepoints | Whether an outcome is an eligibility criterion at all, or only an extraction target |
| Study design, if used | Which designs are admitted, and whether non-randomised evidence is in scope for all outcomes or only harms | Whether design is applied at screening or recorded and handled in analysis |
Compiled as a working rule set; the element structure follows the Cochrane Handbook, version 6.5.
The review question behind the specification belowIn adults aged 60 or over with musculoskeletal pain of at least three months, do supervised or prescribed exercise programmes of at least four weeks, compared with no intervention, waiting list, usual care or attention control, improve self-reported pain and physical function at about 12 weeks?
An eligibility specification, as written and as it needs to be
Cannot decide a borderline record
We included randomised controlled trials of exercise interventions in older adults with chronic pain, compared with control, reporting pain and function outcomes. Studies in mixed populations were considered on a case-by-case basis.
“Chronic pain” is undefined. Three months, six months, or whatever the authors described as chronic: two screeners will not reach the same decision, and the population of the review will be determined by whatever the records happened to state.
“Compared with control” names no comparator. Waiting list, usual care and attention control yield different effect sizes; pooling across them without stating this pools across an unmeasured intervention.
No timepoint on the outcomes. “Pain and function outcomes” admits every measurement from two weeks to two years into the same analysis.
“Case-by-case basis” is the absence of a rule. It converts a specification problem into an arbitration problem, repeated for every mixed-population record.
No default for silent records. A screener assessing an abstract that does not state age must guess what the team would want, and two screeners will guess differently.
No handling rule for multimodal interventions. A trial of exercise plus education compared with no intervention is not a trial of exercise, yet nothing in this specification excludes it.
Decides a borderline record without further consultation
We included parallel-group and cluster randomised trials in adults aged 60 years or over with musculoskeletal pain of at least three months’ duration, ascertained by the study investigators by any method, in community or outpatient settings. Eligible interventions were supervised or prescribed exercise programmes of at least four weeks’ duration delivered by any provider; programmes in which exercise was one component of a multimodal package were eligible only where the comparator received the same non-exercise components. Eligible comparators were no intervention, waiting list, usual care as described by the investigators, or an attention control; trials comparing two exercise modalities without a non-exercise arm were excluded. Critical outcomes were self-reported pain intensity on any validated scale and self-reported physical function, both at the timepoint closest to 12 weeks after randomisation; important outcomes were adverse events and withdrawal for any reason. Where fewer than 80 per cent of participants met the age criterion, the trial was eligible only if results for participants aged 60 or over were reported separately. Records that did not state participant age were taken to full text rather than excluded.
Every element carries an operational threshold, so that a borderline record is decided by the criterion rather than by the individual screener.
Comparators are named individually, which makes the planned syntheses visible and ensures that usual care is described rather than assumed.
The outcome specification fixes the construct, instrument class and timepoint, so that the pooled estimate has a single interpretation.
The mixed-population rule is stated as a proportion with a fallback, which eliminates case-by-case negotiation.
Records that are silent on a criterion are explicitly referred to full text, so that the screen is sensitive by design rather than by individual judgement.
The multimodal rule converts a known confounder into an eligibility decision made before any results are seen.
A reviewer reading the second version can determine what would and would not have been included without asking any questions, which is the decisive test of an eligibility specification. It also requires no additional effort at submission if it was written at the protocol stage, because it is the same text.
Placement and timing of the PICO specification
The review PICO should be recorded in a registered protocol before screening begins 6. This is not a bureaucratic preference. A PICO written after the yield is known cannot be distinguished by any reader from a PICO fitted to the yield, and this suspicion is easy to avoid but difficult to dispel. Prospective registration with a dated record of the four elements and the ranked outcomes is what makes any later change identifiable as a change.
Changes are permitted. Reviews legitimately discover that a comparator does not exist in the literature, that an outcome is universally reported on an instrument the protocol did not anticipate, or that the population definition admits an unforeseen study type. The change should be reported, together with when and why it was made and whether it was made before or after the results for that comparison were visible. A declared amendment requires a single sentence; an undeclared amendment, discovered by a reviewer comparing the protocol with the manuscript, can undermine the credibility of the entire review.
The PICO must also be carried through into the report. The eligibility criteria section should correspond recognisably to the criteria in the protocol, the characteristics table should be organised by the same elements, and the analyses should map onto the per-synthesis questions declared in advance. The reporting requirements for all of this are set out in the PRISMA 2020 statement, which requires reporting of the eligibility criteria and of the rationale for grouping studies for synthesis 11.
Limitations of the PICO framework
A precise PICO makes a review answerable, but it does not make it informative. Three limitations should be stated explicitly.
A narrow PICO can produce an empty or near-empty review. This is a legitimate result, since an empty review accurately maps where evidence is absent, but it is better anticipated than discovered, which argues for a scoping search before the protocol is finalised rather than afterwards 14,16.
A precise PICO does not create comparability. If the included studies differ in dose, population severity and outcome timepoint, the fact that all satisfy the eligibility criteria does not make their effects exchangeable. The specification controls what enters the review; it does not determine whether pooling is appropriate, which is a separate judgement discussed in meta-analysis methodology.
Moreover, a PICO conveys nothing about the quality of the evidence it admits. Specifying the population, intervention, comparator and outcomes tells a reader what question the review asked. Whether the studies answering it were well conducted, whether their reported outcomes match their protocols, and whether the published record constitutes the full evidence record are separate problems requiring separate instruments 15,18. Systematic review support addresses those decisions alongside the formulation of the question.
Key points
Key points
PICO is a specification for eligibility decisions, not a title formula, and the Cochrane Handbook distinguishes three levels at which it must be written: the review PICO that sets the scope, the PICO of each planned synthesis, and the descriptive PICO of the included studies. All three should be written in the protocol. The comparator should be named rather than described as control, the outcome instrument and timepoint fixed, outcomes ranked in advance, and a default stated for records that are silent on a criterion. The criteria should then be tested by having two people screen the same hundred abstracts; disagreement indicates that the criteria, not the screeners, require revision. Where the question concerns diagnostic accuracy, prognosis, qualitative experience or methods, the element set designed for that question type should be used rather than PICO.
Application to your study
Application to your study
- The three levels should be written separately. One paragraph should address the review scope, a numbered list the planned syntheses, and a note the study characteristics to be extracted to explain heterogeneity within them.
- Every comparator should be named. The word “control” should be replaced in the protocol with the specific comparators to be admitted, including a description of how usual care will be defined.
- Every outcome should be assigned a timepoint and an instrument class. Outcomes should then be ranked as critical or important, with a statement of which will undergo certainty assessment.
- The eligibility table should include a default column. For each criterion, the action to be taken when the record is silent should be stated: exclusion or progression to full text.
- The mixed-population rule should be expressed numerically. A stated proportion with a subgroup-reporting fallback, rather than “case by case”.
- The criteria should be piloted on a shared set before the main screen. Two screeners assess the same records, and the criteria that produced disagreements are then rewritten.
- The PICO should be registered before screening, and any change dated. Each change should be reported as made before or after the relevant results were visible.
Frequently asked questions
Population, Intervention, Comparison and Outcome: the four elements that the Cochrane Handbook treats as encapsulating the question of an intervention review. Population denotes the people and setting in which the question is asked; Intervention is the treatment or procedure of interest, specified at the level of detail that could change its effect; Comparison is the alternative against which the intervention is evaluated, named specifically rather than described as control; and Outcome is the set of results by which the comparison is judged, each with a definition, an instrument and a timepoint. The elements constitute a specification for eligibility decisions rather than a template for the review title.
The Cochrane Handbook distinguishes the review PICO, the PICO for each synthesis, and the PICO of the included studies. The review PICO is the planned scope and determines the eligibility criteria and the search. The PICO for each synthesis is the narrower question that one analysis answers, so a review with several comparators has several such questions and should declare them in advance. The PICO of the included studies is descriptive: it records what the studies actually recruited, delivered and measured. It is important because unexplained heterogeneity usually lies in the gap between the first and third levels: the review admitted a range of studies that the analysis then treated as interchangeable.
Specific enough for two screeners to agree, and no more specific than the information available to them. In practice, this requires three components: the condition and how its presence was ascertained; the relevant bounding characteristics, such as age, severity, comorbidity, setting and health system; and an explicit handling rule for mixed populations. The population should not be defined by anything that could be a consequence of the intervention, such as response to initial therapy, because the review then cannot answer an effectiveness question. Criteria that no abstract reports, such as a biomarker threshold, should be avoided unless those records are deliberately referred to full text.
Yes, and most non-trivial reviews do. A review whose scope admits several interventions or several comparators contains a structured set of questions, one per planned synthesis, and the protocol should list them rather than imply a single question. This differs from adding comparisons after the data have been seen. The criterion a reader will apply is whether each synthesis was declared before its results were visible. The planned analyses should be declared, the primary analyses identified, and any later addition reported as an amendment with its date and rationale.
The study’s description should be recorded without embellishment, and the inconsistency treated as a finding. Usual care is the most common example: it differs by health system and by decade, and studies frequently describe it in a single sentence or not at all. Whatever description exists should be extracted as a study characteristic, and studies should be grouped for synthesis only where the comparators are plausibly similar; where they are not, this should be stated and the studies either analysed separately or not pooled. Where the comparator description is genuinely absent, contacting the study authors is legitimate, as is reporting that the comparator could not be characterised; a reader needs this information to interpret the pooled estimate.
When the question does not concern the effect of an intervention. Diagnostic accuracy reviews require the population, the index test, the reference standard and the target condition; forcing these into PICO leads to the common error of treating the reference standard as a comparator. Prognosis and prediction-model reviews require the population, the predictors or model, the outcome predicted and the prediction horizon. Qualitative and service-delivery questions require elements for context, setting, perspective and phenomenon of interest, since there is no intervention-versus-comparator structure to describe. Published typologies of review methodology cover many distinct review types (one widely used analysis sets out 14), and PICO was designed for one of them.
Usually not, or only for the critical outcomes, and the choice has consequences that should be considered deliberately. Requiring a study to report a given outcome for inclusion makes the review vulnerable to selective outcome reporting within the studies: trials that measured the outcome and obtained an unfavourable result are excluded by their own silence. The safer default is to define eligibility on population, intervention, comparator and design, to include studies that meet those criteria, and then to report how many reported each outcome. Where an outcome is made an eligibility criterion, which is sometimes unavoidable in a narrowly defined review, this should be stated explicitly and the risk acknowledged.
Yes, provided that the change is reported as a change. Reviews legitimately discover that a planned comparator does not exist in the literature, that every study uses an instrument the protocol did not anticipate, or that the population definition admits an unforeseen study type. An amendment is acceptable when it is dated, explained and declared in relation to the results: a reader needs to know whether the change was made before or after the data for that comparison were visible. An undeclared change that a reviewer identifies by comparing the registered protocol with the manuscript is interpreted as post hoc fitting, regardless of how reasonable the change was.
References
- Ostengaard L, Korfitsen CB, Hrobjartsson A, Laursen DRT, Nejstgaard CH. PRISMA 2020 Is a Reporting Guideline for Systematic Reviews but Is Often Misused to Guide Review Content. Cochrane Evidence Synthesis and Methods. 2026;4(5):e70095. doi:10.1002/cesm.70095 read 9 October 2026. doi.org/10.1002/cesm.70095
- Thomas J, Kneale D, McKenzie JE, Brennan SE, Bhaumik S. Chapter 2: Determining the scope of the review and the questions it will address. Cochrane Handbook for Systematic Reviews of Interventions, version 6.5. Cochrane. 2024;6.5. read 9 October 2026. cochrane.org
- Higgins JPT, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, Welch VA (editors). Cochrane Handbook for Systematic Reviews of Interventions, version 6.5. Cochrane. 2024;6.5. read 9 October 2026. cochrane.org
- Rethlefsen ML, Kirtley S, Waffenschmidt S, Ayala AP, Moher D, Page MJ, Koffel JB; PRISMA-S Group. PRISMA-S: an extension to the PRISMA Statement for Reporting Literature Searches in Systematic Reviews. Systematic Reviews. 2021;10(1):1-19. read 9 October 2026. doaj.org
- Jonnalagadda SR, Goyal P, Huffman MD. Automating data extraction in systematic reviews: a systematic review. Systematic Reviews. 2015;4:78. read 9 October 2026. cochrane.org
- Moher D, Shamseer L, Clarke M, Ghersi D, Liberati A, Petticrew M, et al. Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015 statement. Systematic Reviews. 2015;4(1):1. doi:10.1186/2046-4053-4-1 read 9 October 2026. doi.org/10.1186/2046-4053-4-1
- Ip S, Hadar N, Keefe S, Parkin C, Iovin R, Balk EM, Lau J. A web-based archive of systematic review data. Systematic Reviews. 2012;1:15. read 9 October 2026. cochrane.org
- Horsley T, Dingwall O, Sampson M. Checking reference lists to find additional studies for systematic reviews. Cochrane Database of Systematic Reviews. 2011;2011(8):MR000026. doi:10.1002/14651858.MR000026.pub2 read 9 October 2026. doi.org/10.1002/14651858.MR000026.pub2
- Nishikawa-Pacher A. Research Questions with PICO: A Universal Mnemonic. Publications. 2022;10(3):21. doi:10.3390/publications10030021 read 9 October 2026. doi.org/10.3390/publications10030021
- Luijendijk HJ. How to PICO questions about medical tests. BMJ Evidence-Based Medicine. 2021;26(4):155-157. doi:10.1136/bmjebm-2021-111676 read 9 October 2026. doi.org/10.1136/bmjebm-2021-111676
- Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71. doi:10.1136/bmj.n71 read 9 October 2026. doi.org/10.1136/bmj.n71
- Shokraneh F. PICO Framework: Two Decades of Variation and Application. F1000Research. 2019;8:1419. doi:10.7490/f1000research.1117334.1 read 9 October 2026. doi.org/10.7490/f1000research.1117334.1
- Bramer WM, de Jonge GB, Rethlefsen ML, Mast F, Kleijnen J. A systematic approach to searching: an efficient and complete method to develop literature searches. Journal of the Medical Library Association. 2018;106(4):531-541. doi:10.5195/jmla.2018.283 read 9 October 2026. doi.org/10.5195/jmla.2018.283
- Tricco AC, Lillie E, Zarin W, O'Brien KK, Colquhoun H, Levac D, et al. PRISMA Extension for Scoping Reviews (PRISMA-ScR): Checklist and Explanation. Annals of Internal Medicine. 2018;169(7):467-473. doi:10.7326/M18-0850 read 9 October 2026. doi.org/10.7326/M18-0850
- Mayo-Wilson E, Li T, Fusco N, Dickersin K. Practical guidance for using multiple data sources in systematic reviews and meta-analyses (with examples from the MUDS study). Research Synthesis Methods. 2018;9:2-12. read 9 October 2026. cochrane.org
- Munn Z, Peters MD, Stern C, Tufanaru C, McArthur A, Aromataris E. Systematic review or scoping review? Guidance for authors when choosing between a systematic or scoping review approach. BMC Medical Research Methodology. 2018;18(1):143. doi:10.1186/s12874-018-0611-x read 9 October 2026. doi.org/10.1186/s12874-018-0611-x
- Paez A. Gray literature: An important resource in systematic reviews. Journal of Evidence-Based Medicine. 2017;10(3):233-240. doi:10.1111/jebm.12266 read 9 October 2026. doi.org/10.1111/jebm.12266
- Baudard M, Yavchitz A, Ravaud P, Perrodeau E, Boutron I. Impact of searching clinical trial registries in systematic reviews of pharmaceutical treatments: methodological systematic review and reanalysis of meta-analyses. BMJ. 2017;356:j448. read 9 October 2026. cochrane.org
- Bramer WM, Giustini D, de Jonge GB, Holland L, Bekhuis T. De-duplication of database search results for systematic reviews in EndNote. Journal of the Medical Library Association. 2016;104(3):240-243. doi:10.3163/1536-5050.104.3.014 read 9 October 2026. doi.org/10.3163/1536-5050.104.3.014
- Rethlefsen ML, Farrell AM, Osterhaus Trzasko LC, Brigham TJ. Librarian co-authors correlated with higher quality reported search strategies in general internal medicine systematic reviews. Journal of Clinical Epidemiology. 2015;68(6):617-626. doi:10.1016/j.jclinepi.2014.11.025 read 9 October 2026. doi.org/10.1016/j.jclinepi.2014.11.025
- Stewart LA, Clarke M, Rovers M, Riley RD, Simmonds M, Stewart G, Tierney JF; PRISMA-IPD Development Group. Preferred reporting items for a systematic review and meta-analysis of individual participant data: the PRISMA-IPD statement. JAMA. 2015;313:1657-1665. read 9 October 2026. cochrane.org
- Squires J, Valentine J, Grimshaw J. Systematic reviews of complex interventions: framing the review question. Journal of Clinical Epidemiology. 2013;66:1215-1222. read 9 October 2026. cochrane.org
- Zarin DA, Tse T, Williams RJ, Califf RM, Ide NC. The ClinicalTrials.gov results database: update and key issues. New England Journal of Medicine. 2011;364:852-860. read 9 October 2026. cochrane.org
- Papaioannou D, Sutton A, Carroll C, Booth A, Wong R. Literature searching for social science systematic reviews: consideration of a range of search techniques. Health Information and Libraries Journal. 2010;27(2):114-122. doi:10.1111/j.1471-1842.2009.00863.x read 9 October 2026. doi.org/10.1111/j.1471-1842.2009.00863.x
- Grant MJ, Booth A. A typology of reviews: an analysis of 14 review types and associated methodologies. Health Information and Libraries Journal. 2009;26(2):91-108. doi:10.1111/j.1471-1842.2009.00848.x read 9 October 2026. doi.org/10.1111/j.1471-1842.2009.00848.x
- Schardt C, Adams MB, Owens T, Keitz S, Fontelo P. Utilization of the PICO framework to improve searching PubMed for clinical questions. BMC Medical Informatics and Decision Making. 2007;7:16. doi:10.1186/1472-6947-7-16 read 9 October 2026. doi.org/10.1186/1472-6947-7-16
- Counsell C. Formulating questions and locating primary studies for inclusion in systematic reviews. Annals of Internal Medicine. 1997;127:380-387. read 9 October 2026. cochrane.org
- Richardson WS, Wilson MC, Nishikawa J, Hayward RS. The well-built clinical question: a key to evidence-based decisions. ACP Journal Club. 1995;123(3):A12-A13. doi:10.7326/ACPJC-1995-123-3-A12 read 9 October 2026. doi.org/10.7326/ACPJC-1995-123-3-A12
- Zhang S, Wang M, Lin R, Shuai Z, Lv Z, Wang C, Zhang R, Yang T, Wang Y, Zhang X. Remote exercise snacking and fall-related functional outcomes in older adults: a systematic review including a meta-analysis. Frontiers in Physiology. 2026;17:1709619. doi:10.3389/fphys.2026.1709619 read 13 November 2026. doi.org/10.3389/fphys.2026.1709619
- Li Y, Zhang L, Wang W, Zhao H, Liao X. Efficacy and safety of the commercial Chinese polyherbal preparation Liu Shen Wan as an adjunctive treatment for herpes zoster and postherpetic neuralgia: a systematic review and meta-analysis. Frontiers in Pharmacology. 2025;16:1698753. doi:10.3389/fphar.2025.1698753 read 13 November 2026. doi.org/10.3389/fphar.2025.1698753
- Terzoni S, Maruccia S, Di Prisco L, Parozzi M, Caruso R, Ferrara P, et al. Preoperative pelvic muscle training for continence recovery after prostatectomy: yet another meta-analysis? Neurourology and Urodynamics. 2026;45(1):127-136. doi:10.1002/nau.70167 read 13 November 2026. doi.org/10.1002/nau.70167
- Li H, Deng Y, Xu J, Xiong X, Li X, Zhao T, Ying Q, Zhang J. Proximal femoral bionic nail (PFBN) versus InterTAN nail for intertrochanteric femoral fractures in elderly patients: a systematic review and meta-analysis. BMC Surgery. 2026;26. doi:10.1186/s12893-026-03943-8 read 13 November 2026. doi.org/10.1186/s12893-026-03943-8
- Methley AM, Campbell S, Chew-Graham C, McNally R, Cheraghi-Sohi S. PICO, PICOS and SPIDER: a comparison study of specificity and sensitivity in three search tools for qualitative systematic reviews. BMC Health Services Research. 2014;14:579. doi:10.1186/s12913-014-0579-0 read 13 November 2026. doi.org/10.1186/s12913-014-0579-0
- Eriksen MB, Frandsen TF. The impact of patient, intervention, comparison, outcome (PICO) as a search strategy tool on literature search quality: a systematic review. Journal of the Medical Library Association. 2018;106(4):420-431. doi:10.5195/jmla.2018.345 read 13 November 2026. doi.org/10.5195/jmla.2018.345
Every source above is a peer-reviewed article, a publisher’s record of one, or a standard issued by the body that owns it. Each carries the date it was last read. Where a standard moves, this article is revised rather than left up with a note.
How to cite this article
Pick your referencing style and copy it. Your university or target journal may vary the punctuation — check its guide if it has one.
Pubrica Academy. (2026, October 9). Framing a Systematic Review Question with PICO. Pubrica. https://pubrica.com/academy/systematic-review/the-pico-framework-for-framing-systematic-review-research-questions/
Pubrica Academy. “Framing a Systematic Review Question with PICO.” Pubrica, 9 Oct. 2026, pubrica.com/academy/systematic-review/the-pico-framework-for-framing-systematic-review-research-questions/.
Pubrica Academy (2026) Framing a Systematic Review Question with PICO. Pubrica. Available at: https://pubrica.com/academy/systematic-review/the-pico-framework-for-framing-systematic-review-research-questions/ (Accessed: 15 October 2026).
Pubrica Academy. “Framing a Systematic Review Question with PICO.” Pubrica. October 9, 2026. https://pubrica.com/academy/systematic-review/the-pico-framework-for-framing-systematic-review-research-questions/
Pubrica Academy, ‘Framing a Systematic Review Question with PICO’ (Pubrica, 9 October 2026) <https://pubrica.com/academy/systematic-review/the-pico-framework-for-framing-systematic-review-research-questions/> accessed 15 October 2026.
Pubrica Academy. Framing a Systematic Review Question with PICO. [Internet]. Pubrica; 2026 Oct 9 [cited 2026 Oct 15]. Available from: https://pubrica.com/academy/systematic-review/the-pico-framework-for-framing-systematic-review-research-questions/
Pubrica Academy, “Framing a Systematic Review Question with PICO,” Pubrica, Oct. 9, 2026. [Online]. Available: https://pubrica.com/academy/systematic-review/the-pico-framework-for-framing-systematic-review-research-questions/
@misc{the_pico_framework_for_framing_systema_2026,
author = {{Pubrica Academy}},
title = {Framing a Systematic Review Question with PICO},
howpublished = {Pubrica},
year = {2026},
month = {oct},
url = {https://pubrica.com/academy/systematic-review/the-pico-framework-for-framing-systematic-review-research-questions/},
note = {Accessed: <span data-cite-acc="iso">2026-10-15</span>}
}This article is written to be learned from. Cite it if you are referring to what it says; if you are using the evidence behind it, cite the sources listed above instead.
PICO-to-eligibility specification worksheet
This table should be completed before screening begins and kept with the protocol. The third column prevents most screening disagreements, and it is the column that is most often omitted.
PICO SPECIFICATION WORKSHEET
=== LEVEL 1: REVIEW PICO (the scope) ===
POPULATION
Condition: ______________________________________
How ascertained (accepted definition / reference standard): ______
Age range: ____________ Setting: ____________________
Other bounding characteristics: ___________________________
MIXED-POPULATION RULE: eligible if at least ____% meet the
definition; otherwise eligible only if the relevant subgroup is
reported separately [ ] / excluded [ ]
DEFAULT IF RECORD IS SILENT: [ ] exclude [ ] take to full text
INTERVENTION
Agent / procedure: ______________________________________
Dose or intensity range eligible: _________________________
Duration: minimum ________ Route / mode: _______________
Delivered by: ___________________________________________
Co-interventions permitted? [ ] yes [ ] no [ ] only if also in comparator
Multimodal packages eligible? [ ] no [ ] yes, if: ______________
DEFAULT IF RECORD IS SILENT: [ ] exclude [ ] take to full text
COMPARISON (name each one — do not write "control")
[ ] placebo / sham
[ ] no intervention
[ ] waiting list
[ ] usual care — will be described as reported by investigators
[ ] attention control
[ ] active alternative, namely: ___________________________
Multi-arm studies: contribute [ ] all eligible arms [ ] one arm,
chosen by this rule: ______________ [ ] combined, by this method: ______
Head-to-head active comparisons in scope? [ ] no [ ] yes
(if yes — is this a network meta-analysis? decide now: ____)
DEFAULT IF RECORD IS SILENT: [ ] exclude [ ] take to full text
OUTCOMES (rank them; do not let the literature choose)
CRITICAL (drive the summary of findings and certainty assessment)
1. Construct: __________ Instrument class: __________ Timepoint: ______
2. Construct: __________ Instrument class: __________ Timepoint: ______
IMPORTANT
3. ________________________________________________
4. ________________________________________________
HARMS (include even if you expect them to be reported badly)
5. ________________________________________________
Is any outcome an ELIGIBILITY criterion? [ ] no [ ] yes, which: ______
(if yes, note the selective-reporting risk this creates: ____________)
Rule for a different instrument measuring the same construct: __________
Rule for a different timepoint: __________________________
STUDY DESIGN (if used as a criterion)
Eligible designs: ________________________________________
Non-randomised evidence: [ ] excluded [ ] harms only [ ] all outcomes
Applied at [ ] screening [ ] extraction and handled in analysis
=== LEVEL 2: PLANNED SYNTHESES (one line each) ===
S1: ____ in ____ versus ____ for ____ at ____ PRIMARY? [ ]
S2: ____ in ____ versus ____ for ____ at ____ PRIMARY? [ ]
S3: ____ in ____ versus ____ for ____ at ____ PRIMARY? [ ]
Pre-specified subgroups, with the reason each is expected to matter:
________________________________________________
=== LEVEL 3: STUDY-LEVEL PICO TO EXTRACT ===
Characteristics to extract so heterogeneity can be explained later:
[ ] actual population severity / case mix
[ ] actual dose, intensity, duration delivered
[ ] actual comparator as described
[ ] actual outcome instrument and measurement timepoint
[ ] setting, country, year of recruitment
[ ] ____________________________________________
=== CALIBRATION ===
Records screened by both reviewers in the pilot: ______
Disagreements: ______ Criteria rewritten as a result: ______________
Date criteria frozen: ____________ Registered where: ______________
=== AMENDMENTS ===
Date | What changed | Why | Were the relevant results visible? (Y/N)
____ | ___________ | ___ | ___ Pubrica AcademyWritten and maintained by Pubrica’s subject matter experts. Articles are revised when the underlying standards change, and each article states the date on which it was last reviewed.
How Pubrica can help
Pubrica works through the PICO with the review team at protocol stage, writes the eligibility criteria as a decision table with explicit defaults, and conducts the two-screener calibration on a sample of the team’s own records so that the criteria are tested before the full screen begins. Where the question proves not to be an intervention question, Pubrica advises on the element set that fits it rather than applying PICO to a diagnostic accuracy or prognosis review.