Contents
Academy · Research services · Why use the PICO framework
The Rationale for the PICO Framework in Research: Evidence, Limitations and Alternatives
PICO specifies a comparative question, which is a genuine benefit. It is not, however, a search strategy, and the evidence that it improves searching consists of one review of three studies. This article sets out the case for PICO and the alternatives to it.
Definition and scope of the PICO framework
PICO is a structure for formulating a question. It requires four elements to be named: the population under study, the intervention of interest, the comparison against which it is measured, and the outcome by which it is judged. It was introduced in a 1995 editorial by Richardson and colleagues, who were concerned with making clinical questions specific enough to be answered1.
The case for PICO at a glance
- What it is
- A structure for stating a comparative question: Population, Intervention, Comparison, Outcome1
- What it is not
- A search strategy; it does not choose databases, vocabulary or how terms are combined10
- Evidence it improves searching
- One systematic review of three studies, with no pooled analysis2
- Real benefit
- Four decisions made before the work starts, which make a question easier to critique and harder to change quietly
- Use something else for
- Exposures (PECO, PEO), qualitative questions (SPIDER, PEO) and service questions (SPICE)
- Reading time
- About 14 minutes, with 3 figures and 4 tables
It is worth being blunt about what PICO is not, because the confusion is common enough that a nursing education journal published a paper about it10. PICO is not a search strategy. It does not tell you which database to use, how to combine terms, or which controlled vocabulary applies. It tells you what your question is. The translation from a well-formed question into a well-formed search is a separate skill, and a question can be perfectly PICO-shaped and still produce a useless search if that translation is done mechanically.
Summary of the argument
A four-part question is a conceptual tool. The value of PICO lies in the four decisions it requires before work begins, not in any demonstrated improvement in search results. The next section explains why this distinction matters.
One of the test questions in a published search studyIs an ACE inhibitor alone better than a diuretic alone for reducing hypertension in African American patients?
Evidence on the effectiveness of PICO
PICO is recommended almost universally, so it is reasonable to ask what the evidence for it looks like. It is thinner than its ubiquity suggests, and we would rather tell you that than cite papers that do not say what we need them to say.
There is one systematic review. Eriksen and Frandsen set out to establish whether using PICO as a search strategy tool improves the quality of a literature search. They found three eligible studies and carried out no quantitative analysis. A review of three heterogeneous studies with no pooled estimate cannot establish an effect, and the authors do not claim it does2.
The most widely cited individual study had thirty-one participants. Schardt and colleagues compared a PICO-structured search interface against standard PubMed searching. Precision was higher with the PICO interface, but the difference was not statistically significant. This paper is widely cited as demonstrating that PICO improves searching. It does not3.
Table 1. The quantitative evidence base for PICO as a search tool, in full
| What exists | |
|---|---|
| Systematic reviews | One (Eriksen & Frandsen, 2018) |
| Studies included in that review | Three |
| Pooled or quantitative analysis | None performed |
| Most-cited primary study | Thirty-one participants (Schardt et al., 2007) |
| Result of that study | Precision higher with PICO; difference not statistically significant |
Table 2. The primary studies that have tested PICO as a search tool
| Study | Design and searchers | What was compared | Main finding |
|---|---|---|---|
| Schardt et al., 20073 | Randomised pilot; 31 interns and residents, 3 clinical questions | Two PICO search templates versus standard PubMed | Higher precision with the PICO templates on every question; confidence intervals overlapped |
| Agoritsas et al., 201212 | 30 clinical questions from Cochrane reviews; 15 PubMed strategies | PIC versus PICO queries, with and without Clinical Queries filters, limits and related-articles links | The narrow Clinical Queries filter with PICO-based queries performed best on the first two result pages, with median sensitivity of about 50% |
| Hoogendam et al., 201213 | Randomised crossover trial; 22 residents and specialists, 434 queries in 5-minute searches | PICO queries versus standard queries | Precision 3.44% with PICO and 4.02% without; recall 13.62% and 12.27%; neither difference significant |
| Methley et al., 20145 | One qualitative review question, three databases | PICO versus PICOS versus SPIDER, using identical terms | PICO found all 18 relevant studies; PICOS and SPIDER found 13 each, from far fewer records |
Each row is described from the study’s abstract or full text, retrieved from PubMed in November 2026. The last three are the studies included in the Eriksen and Frandsen review; Schardt et al. was not among them.
What the evidence does support
Taken together, these studies show that the number of concept blocks in a search changes what it retrieves: more blocks give fewer records and lower sensitivity. They do not show that writing the question as PICO, by itself, produces a better search2. Claim the structure for what it does: specify the question.
This does not mean PICO is worthless. It means the honest claim for it is a claim about clarity of thought, not about retrieval performance. A question that names its population, its comparator and its outcome is easier for a colleague to critique, easier to convert into eligibility criteria, and harder to quietly change halfway through a project. Those are real benefits. They are simply not the benefit that is usually advertised.
Questions unsuited to the PICO structure
The most common problem we see in draft protocols is not a badly written PICO question. It is a question that was never PICO-shaped being forced into four slots. The giveaway is an I element that nobody assigned, or a C element invented to fill the gap.
Table 3. Question type and the framework that fits it
| Question type | Example | What fits |
|---|---|---|
| Therapy or prevention | Does drug A lower mortality more than drug B? | PICO fits. This is the purpose for which it was designed. |
| Aetiology or risk | Does occupational solvent exposure raise the risk of Parkinson disease? | PECO or PEO. There is an exposure, not an intervention. Entering solvent exposure in the I element implies that it was assigned. |
| Prognosis | Among patients with early cirrhosis, what predicts decompensation within five years? | Awkward in PICO. There is often no comparator and no intervention; the question is about prediction. |
| Diagnostic accuracy | How well does transient elastography identify advanced fibrosis? | Awkward in PICO. The C slot holds a reference standard rather than an alternative treatment, and the O slot holds accuracy rather than a patient outcome. |
| Qualitative | How do patients on long-term dialysis experience decisions about withdrawal? | Not PICO. There is no intervention, no comparator and no measured outcome. SPIDER or PEO is the appropriate structure. |
| Service or policy | Should a clinic run nurse-led or pharmacist-led anticoagulation review? | SPICE. Setting and perspective are part of the question. |
Two of these require further explanation. An aetiology question concerns an exposure, not an intervention: nobody is assigned to smoke, to be poor, or to live near a motorway. Writing an exposure into the I slot implies a design that does not exist, and it will send you looking for trials where only cohort studies are possible. A qualitative question has no comparator and no measured outcome at all; it has a phenomenon to be understood. Those questions are not badly formed PICO questions, they are a different kind of question.
An exposure forced into PICO
“In adults, does the intervention of occupational solvent exposure, compared with no exposure, cause Parkinson disease?”
Nobody assigns solvent exposure. Writing it as an intervention implies a trial that cannot exist and sends the search towards the wrong designs.
The same question as PECO
“Among adult workers, is occupational solvent exposure, compared with no such exposure, associated with a higher risk of Parkinson disease?”
Named as an exposure, the question points to cohort and case-control studies, which is where the answer will be.
Alternative frameworks and their purposes
Table 4. Question frameworks in common use
| Framework | Elements | Suited to | Source |
|---|---|---|---|
| PICO | Population, Intervention, Comparison, Outcome | A quantitative question comparing two courses of action. | Richardson et al., 1995 |
| PICOT | adds Time | Where the time horizon forms part of the question; HbA1c at twelve weeks is a different question from HbA1c at five years. | Variant in common use |
| PICOS | adds Study design | Where eligibility is restricted by design, which is usual in a systematic review. | Variant in common use |
| PECO | Population, Exposure, Comparison, Outcome | Aetiology and risk questions. Smoking is not assigned; it is an exposure, not an intervention. | Variant in common use |
| PEO | Population, Exposure, Outcome | Risk and aetiology questions with no comparator group. | In wide use; no originating paper was located |
| SPIDER | Sample, Phenomenon of Interest, Design, Evaluation, Research type | Qualitative and mixed-methods questions, where there is a phenomenon to understand rather than an intervention to compare. The caveat below should be read before the framework is used. | Cooke, Smith & Booth, 2012 |
| SPICE | Setting, Perspective, Intervention, Comparison, Evaluation | Questions about a service or a way of working, where the perspective of the enquirer and the setting alter the answer. | Booth, 2006 |
Two notes on provenance, because reference lists in this area are often loose. SPICE comes from a 2006 paper by Booth, and Booth’s own glosses for the five elements are “Setting – where?”, “Perspective – for whom?”, “Intervention – what?”, “Comparison – compared with what?” and “Evaluation – with what result?”. The frequently repeated claim that SPICE is specifically for qualitative or service-evaluation questions is a library-guide convention; Booth does not say it, so we do not attribute it to him6. PEO is in wide use and we could not open a paper that originates or defines it, so it is described here as a widely used variant rather than attributed to anyone.
Evidence on SPIDER as an alternative
SPIDER is the framework usually recommended for qualitative evidence synthesis, and it is recommended on a great many pages that do not mention what happened when somebody tested it. Two things are worth knowing before you choose it.
Its own authors were cautious. Cooke, Smith and Booth introduced SPIDER in 2012 and concluded that “To constitute a viable alternative to PICO, SPIDER needs to be refined and tested on a wider range of topics.” That is the originating paper’s own assessment, not a critic’s4.
The only head-to-head comparison found against it. Methley and colleagues ran PICO, PICOS and SPIDER against the same qualitative review question across three databases. PICO retrieved all 18 relevant articles. SPIDER and PICOS retrieved 13 each. SPIDER was by far the most specific — 239 hits against 23,758 for PICO — which is exactly its appeal, and exactly the problem. Their conclusion was that “Based on these limited findings the SPIDER tool would not be recommended due to the risk of not identifying relevant papers.”5
Implications for systematic reviews
SPIDER restructures a qualitative question in a way that PICO cannot, and it produces a far more precise search. However, the only evaluation that could be verified advises against its use for a systematic review, because the gain in precision was achieved at the cost of missing relevant studies. Where SPIDER is used to structure a qualitative question, it should not be used alone to build the search. That position is defensible; the claim that “SPIDER is better than PICO for qualitative work” is not.
Critiques of the PICO framework
There is a small literature arguing that PICO is the wrong tool, and it is worth reading rather than ignoring — particularly if your question is about a problem rather than a treatment.
Waldrop and Dunlap propose replacing PICO with PPCO — problem, population, change, outcome — putting the problem first, before any intervention is named. Their objection is that “the traditional PICO question favors experimental research evidence and that from medical interventions.” and that “it supports a premature bias toward an already-known intervention or practice change.” Where the difficulty is that nobody has yet characterised the problem, they argue that “the PICO question doesn't always address the need to describe a problem thoroughly”9
Tucker and colleagues make a related argument in a 2023 paper and propose an alternative they call SEARCH. A companion paper by the same group, led by Cullen, proposes a different replacement again, which they call the PURPOSE statement. We cite both because the existence of two competing replacements from one research group tells you something useful about how settled this is. We do not spell out what either acronym stands for: the articles themselves were not reachable from here, and we would rather leave a gap than guess at an expansion7,8.
Schiavenato and Chu argue that the deeper problem is that PICO is widely taught as a search method rather than a question structure. That is the same conflation this page opened by correcting, and it is the one worth guarding against in your own writing10.
Selection of a question framework
Decide what kind of question you have before you decide which letters to use. Work down the questions below and stop at the first “yes”; if time or design is part of a PICO question, add T or S.
Whichever you choose, the framework is the start of the protocol, not the end of it. Hanson’s step-by-step account of clinical research treats the four-part question as the thing you refine first and then convert into primary and secondary aims, which is the right order11.
Practical application
- Write the question in full prose first, then attempt to divide it into elements. If an element cannot be completed without invention, this is informative: the question probably does not fit the PICO structure.
- Examine the I element specifically by asking who assigned it. If nobody assigned it, it is an exposure, and PECO or PEO is more appropriate.
- Examine the C element. If a comparator was invented to fill the slot, remove it and consider whether the question is genuinely comparative.
- If the question is qualitative, structure it with SPIDER or PEO, but construct the search more broadly than SPIDER alone would suggest, for the reasons given above.
- Do not claim in a protocol or manuscript that PICO improved the search. State that it specified the question, which is its function and a claim that can be supported.
- When converting the question into a search, follow the worked example in Pubrica’s practical PICO guide rather than combining the four elements with AND.
Key points
PICO is a sound discipline for specifying a comparative question. It is not a search strategy, and the evidence that it improves searching consists of one review of three studies with no pooled analysis plus one non-significant study of thirty-one people. PICO is appropriate where there is an intervention and a comparator; PECO or PEO suit exposures, SPIDER or PEO suit qualitative questions, and SPICE suits questions in which setting and perspective matter. Papers arguing against PICO should be treated as arguments to be answered rather than as citations to be borrowed.
A four-part structure for writing an answerable comparative question: the population, the intervention, the comparison and the outcome. It was introduced in a 1995 editorial by Richardson and colleagues as a way of making clinical questions specific enough to be searched and answered. It is a way of organising the question, not a method for searching a database.
Because it requires four decisions that are easily left vague: which population, receiving which intervention, compared with what, and measured how. A question lacking any of these cannot be answered from the literature without the reader inferring the intended meaning. PICO has not, however, been shown to improve the results of a literature search; the evidence for that claim is limited, and this article states so explicitly.
It converts a general interest into a specification. “Does diet help diabetes” becomes a statement of which patients, which dietary pattern, compared with which alternative, and which measured outcome over what period. If the question becomes a systematic review, each of these four decisions then becomes a line in the eligibility criteria.
In a review, the four elements become the inclusion and exclusion criteria, which is a different function from framing a single clinical question; this is covered in the page on PICO in systematic review questions. For searching, the elements provide the concept blocks, but the translation is not mechanical: synonyms are combined within a concept, concepts are combined across, and the comparator is usually not searched at all. The practical guide works an example through.
The middle letter. PICO has an intervention — something assigned or chosen, which can be compared with an alternative. PEO has an exposure — something a person has encountered, which nobody assigned. Whether air pollution worsens asthma is a PEO question; whether an air filter improves asthma control compared with no filter is a PICO question. Forcing an exposure into the I element is the most common misuse of PICO observed in draft protocols.
References
- 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 15 October 2026. View on ACP Journal ClubThe editorial in which the four-part clinical question was introduced. Verified from the University of Iowa institutional record. No open full text was reachable, so the paper's own wording of the four elements is not quoted anywhere on this site.
- 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 15 October 2026. View on Journal of the Medical Library AssociationThe only systematic review of whether PICO improves searching. It found three eligible studies and performed no quantitative analysis. It is the correct citation for the state of the evidence, and it does not support a claim that PICO measurably improves search quality.
- 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 15 October 2026. View on BMC Medical Informatics and Decision MakingA study of 31 participants comparing a PICO-structured search interface with standard PubMed search. Precision was higher with the PICO interface, but the difference was not statistically significant. Frequently cited as though it demonstrated a benefit.
- Cooke A, Smith D, Booth A. Beyond PICO: The SPIDER Tool for Qualitative Evidence Synthesis. Qualitative Health Research. 2012;22(10):1435–1443. doi:10.1177/1049732312452938 read 15 October 2026. View on Qualitative Health ResearchThe originating SPIDER paper. Its own conclusion is cautious: the authors say SPIDER needs refinement and wider testing before it can be considered a viable alternative to PICO.
- 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 15 October 2026. View on BMC Health Services ResearchThe only head-to-head comparison we could verify. PICO retrieved all 18 relevant articles; SPIDER and PICOS retrieved 13 each. SPIDER was the most specific in every database. The authors advise against SPIDER for systematic reviews.
- Booth A. Clear and present questions: formulating questions for evidence based practice. Library Hi Tech. 2006;24(3):355–368. doi:10.1108/07378830610692127 read 15 October 2026. View on Library Hi TechThe source of SPICE. Note that the paper does not state which question types SPICE suits; the familiar claim that it is for qualitative or service-evaluation questions is a library-guide convention and is not attributed to Booth here.
- Tucker S, Edmonds SW, Cullen L, Hanrahan K, Laures E. The Problem With PICO for Finding the Best Evidence: Fishing With the Wrong Bait and the SEARCH Solution. Journal of PeriAnesthesia Nursing. 2023;38(5):809–812. doi:10.1016/j.jopan.2023.07.002 read 15 October 2026. View on Journal of PeriAnesthesia NursingA critique of PICO proposing an alternative the authors call SEARCH. Bibliographic detail verified from the first author's institutional record; the article text itself could not be opened, so what SEARCH stands for is not stated anywhere on this site.
- Cullen L, Hanrahan K, Tucker S, Edmonds SW, Laures E. The Problem With the PICO Question: Shiny Object Syndrome and the PURPOSE Statement Solution. Journal of PeriAnesthesia Nursing. 2023. doi:10.1016/j.jopan.2023.01.024 read 15 October 2026. View on Journal of PeriAnesthesia NursingA companion critique from the same group, proposing a different replacement again. Relevant because it shows the objection to PICO is not a single dissenting voice.
- Waldrop JB, Dunlap JJ. Beyond PICO—A New Question Simplifies the Search for Evidence. American Journal of Nursing. 2024;124(3):34–37. doi:10.1097/01.NAJ.0001007676.91191.dd read 15 October 2026. View on American Journal of NursingProposes PPCO — problem, population, change, outcome — putting the problem before any named intervention. Its argument is that PICO pushes the author toward an intervention before the problem has been described.
- Schiavenato M, Chu F. PICO: What it is and what it is not. Nurse Education in Practice. 2021;56:103194. doi:10.1016/j.nepr.2021.103194 read 15 October 2026. View on Nurse Education in PracticeA third critique, arguing that PICO is widely misunderstood as a search method rather than a question structure — the same conflation this site's own pages had made.
- Hanson BP. Designing, conducting and reporting clinical research. A step by step approach. Injury. 2006;37:583–594. doi:10.1016/j.injury.2005.06.051 read 15 October 2026. View on InjuryA single-author five-step framework from question to publication, which uses PICO to refine the study question before converting it into primary and secondary aims. Useful as the bridge between the question frameworks and the protocol.
- Agoritsas T, Merglen A, Courvoisier DS, Combescure C, Garin N, Perrier A, Perneger TV. Sensitivity and predictive value of 15 PubMed search strategies to answer clinical questions rated against full systematic reviews. Journal of Medical Internet Research. 2012;14(3):e85. doi:10.2196/jmir.2021 read 13 November 2026. doi.org/10.2196/jmir.2021One of the three studies in the Eriksen and Frandsen review. Tested PIC and PICO queries with and without PubMed Clinical Queries filters on 30 questions drawn from Cochrane reviews.
- Hoogendam A, de Vries Robbé PF, Overbeke AJPM. Comparing patient characteristics, type of intervention, control, and outcome (PICO) queries with unguided searching: a randomized controlled crossover trial. Journal of the Medical Library Association. 2012;100(2):121–126. doi:10.3163/1536-5050.100.2.010 read 13 November 2026. doi.org/10.3163/1536-5050.100.2.010A randomised crossover trial in 22 physicians. PICO queries did not give better recall or precision than standard queries in five-minute searches.
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 writes and critiques research questions and protocols as part of its research design and systematic review services, including an independent assessment of whether a question is appropriately structured before a protocol is committed to it.