PICO Question Generator
Enter the Population, Intervention or exposure, Comparison and Outcome — and, if you want, a Time frame or Study design — to get a focused clinical question phrased for therapy, prevention, diagnosis, prognosis, aetiology or meaning, plus a search block for each element you can send straight to the Boolean Search String Builder. Honest templates, no AI. Free, no sign-up.
Does an intervention work better than the alternative? Usually answered by randomised controlled trials, and systematic reviews of them.
Your PICO elements
Leave blank to search for the element as written above.
Leave blank to search for the element as written above.
Leave blank to search for the element as written above.
Leave blank to search for the element as written above.
Leave blank to search for the element as written above.
Leave blank to search for the element as written above.
Your PICO question
In [population], how does [intervention] affect [outcome]?
The same elements, as every question type
Therapy — In [population], how does [intervention] affect [outcome]?
Prevention — In [population], does [intervention] reduce the risk of [outcome]?
Diagnosis — In [population], how accurate is [intervention] for diagnosing [outcome]?
Prognosis — In [population], how does [intervention] change the likelihood of [outcome]?
Aetiology or harm — Are [population] exposed to [intervention] at greater risk of [outcome] than those who are not?
Meaning — How do [population] experience [intervention]?
Search terms
Fill in an element marked “Use in the search string” to see its search block.
How to Use the PICO Question Generator
- 1. Pick the question type. Therapy, prevention, diagnosis, prognosis, aetiology or meaning. The labels change to match: in a diagnosis question the I is the test, in an aetiology question it is an exposure.
- 2. Write each element as it should read in the question. "Older adults living at home", not "elderly"; "falls", not "health". Add a time frame for PICOT or a study design for PICOS.
- 3. Add search terms. Under each element, list the keywords and synonyms a database would use, one per line. Untick the elements you do not want in the first search — comparison, time and design usually narrow it too far.
- 4. Copy the question, then build the search. The question updates as you type. Copy it, then follow the link to turn the search blocks into PubMed, Scopus, EBSCO, ProQuest, Web of Science and Google Scholar syntax.
What PICO, PICOT and PICOS Stand For
PICO comes from a 1995 editorial by Richardson, Wilson, Nishikawa and Hayward, which argued that a well-built clinical question names the patient or problem, the intervention or exposure, the comparison where one is relevant, and the outcomes that matter. Two extensions are common: nursing curricula add a Time frame (PICOT), and systematic reviews add the Study design they will include (PICOS).
| Letter | Stands for | Example |
|---|---|---|
| P | Population or problem | older adults living at home |
| I | Intervention, exposure, test or factor | a home-based exercise programme |
| C | Comparison (optional) | usual care |
| O | Outcome | falls |
| T | Time frame (PICOT) | 12 months |
| S | Study design (PICOS) | randomised controlled trials |
Six Kinds of Clinical Question
| Type | Asks | I and C are | Usually answered by |
|---|---|---|---|
| Therapy | Does an intervention work better than the alternative? | Treatment vs alternative | Randomised controlled trials |
| Prevention | Does an intervention lower the risk of a condition? | Preventive action vs alternative | Randomised controlled trials |
| Diagnosis | How accurate is a test? | Index test vs reference standard | Cross-sectional accuracy studies |
| Prognosis | What is likely to happen? | Prognostic factor vs its absence | Cohort studies |
| Aetiology or harm | Does an exposure cause harm? | Exposure vs no exposure | Cohort and case-control studies |
| Meaning | How do people experience something? | Phenomenon of interest; no comparison | Qualitative studies |
The design column follows the logic of the Oxford Centre for Evidence-Based Medicine's 2011 levels of evidence, where the best individual study depends on the question: randomised trials for treatment, cross-sectional studies against a reference standard for diagnosis, inception cohorts for prognosis.
Worked Example: Preventing Falls
With the elements in the table above, the generator labels the question PICOT and phrases it as a prevention question like this:
Prevention (PICOT)
In older adults living at home, does a home-based exercise programme compared with usual care reduce the risk of falls over 12 months?
With search terms "older adults, elderly, aged" for P, "exercise*, physical activity, strength training" for I and "fall*, accidental falls" for O — and the comparison and time left out of the search — the combined string reads:
Combined search string
("older adults" OR elderly OR aged) AND (exercise* OR "physical activity" OR "strength training") AND (fall* OR "accidental falls")
A diagnosis question uses the same slots differently. Enter "adults with suspected deep vein thrombosis", "D-dimer testing", "compression ultrasound" and "deep vein thrombosis", choose Diagnosis, and you get: In adults with suspected deep vein thrombosis, how accurate is D-dimer testing compared with compression ultrasound for diagnosing deep vein thrombosis?
Common Mistakes
- A population that is everyone. "Patients" or "adults" gives you thousands of irrelevant studies. Add the condition, setting or age that makes the question yours.
- An outcome you cannot measure. "Wellbeing" is a topic; "falls", "HbA1c" or "30-day readmission" are outcomes.
- Forcing a comparison. Many prognosis and meaning questions have none, and that is fine — the C is optional for a reason.
- Searching for every element. Each extra AND removes records. Start with P, I and O, and add C or the design only if the results are unmanageable.
- Searching with the question's wording. Databases index "aged" and "exercise", not "a home-based exercise programme". List keywords and synonyms instead.
Frequently Asked Questions
- What is a PICO question?
- A PICO question is a focused clinical question built from four elements: the Population or problem, the Intervention or exposure, the Comparison, and the Outcome. Naming each element makes the question answerable and tells you what to search for. For example: in older adults living at home, does a home-based exercise programme compared with usual care reduce the risk of falls?
- What is the difference between PICO and PICOT?
- PICOT adds a Time frame — the period over which the outcome is measured, such as "over 12 months" or "within 30 days of discharge". It is common in nursing education and evidence-based practice projects. The generator labels a question PICOT as soon as you enter a time.
- What is PICOS used for?
- PICOS adds the Study design, and it is mostly used to set eligibility criteria for a systematic review — for example, including only randomised controlled trials. The design usually belongs in your inclusion criteria rather than in the question sentence, which is why the generator lists it separately.
- Does a PICO question always need a comparison?
- No. The comparison is included when relevant — placebo, usual care, another test. Prognosis questions often compare people with and without a factor, and meaning questions have no comparison at all. Leave C empty and the generator writes the question without it.
- Can PICO be used for qualitative research?
- Partly. The meaning question type covers experiences, but many qualitative reviews prefer frameworks built for them: PICo (Population, phenomenon of Interest, Context) or SPIDER (Sample, Phenomenon of Interest, Design, Evaluation, Research type), introduced by Cooke, Smith and Booth in 2012.
- How do I turn a PICO question into a search strategy?
- Treat each element as a concept. List its keywords and synonyms, join them with OR, then join the concepts with AND — usually P, I and O only. Add subject headings such as MeSH where the database has them, and adapt the syntax to each database; the Boolean Search String Builder does that last step for six of them.