PICO, PEO or SPIDER: which framework fits my question?
Use PICO when something is done to people and you have a comparator. Use PEO or PECO when you are studying an exposure nobody assigned, or measuring how common something is, where there may be no comparator at all. Use SPIDER for qualitative questions about experience. The framework determines your search and your eligibility criteria, so choosing the wrong one costs you at every later stage.
Pick by the shape of the question
- PICO — Population, Intervention, Comparator, Outcome
- For effectiveness questions where an intervention is deliberately given and compared against something. The comparator is not optional: "how effective is this drug?" is not a PICO question until you say "compared with what?"
- PEO — Population, Exposure, Outcome
- For questions about exposures rather than interventions, and for prevalence questions where there is nothing to compare against. Nobody randomised anyone to the exposure, which is also why confounding dominates the appraisal.
- PECO — Population, Exposure, Comparator, Outcome
- PEO where a comparison group does exist — exposed versus unexposed. Common in environmental and occupational aetiology reviews.
- SPIDER — Sample, Phenomenon of Interest, Design, Evaluation, Research type
- For qualitative and mixed-methods questions about experience, acceptability or meaning, where "outcome" is the wrong noun and the design is part of what you are selecting on.
The cost of choosing wrong
The framework is not paperwork; it is the skeleton of the search. Forcing a prevalence question into PICO produces an empty comparator concept, and a search built on it either retrieves nothing or quietly restricts to comparative studies you never wanted.
Forcing a qualitative question into PICO is worse, because the outcome concept becomes a set of measured endpoints, and the search then excludes exactly the interview-based studies that would have answered it.
If a concept is genuinely absent — no comparator in a prevalence review — leave it out rather than inventing one. An honest three-part question beats a padded four-part one.
Sharpening the elements
Each element should be specific enough to make a screening decision without a conversation. "Adults" is not a population; "adults over 65 with stage 2 or 3 chronic kidney disease" is. The test is whether two reviewers reading the same abstract would agree on include or exclude from the criteria alone — if not, the element is still too vague, and you will discover it as conflicts during screening.
Where this answer stops
A framework organises a question; it does not make it answerable. Any of these can be filled in completely and still describe a review that no existing literature can support.