Inclusion and exclusion criteria are the rules that decide which studies a systematic review keeps. This builder turns your question into those rules, one element at a time, with tables for PICO (effects), PCC (scoping reviews) and SPIDER (qualitative evidence), each with a worked example. It is for anyone writing a protocol or screening form.
Start from a clear question. If yours is still a topic, work through how to write a research question first. Note that clinical trials also use "inclusion and exclusion criteria" for participants; this page is about criteria for studies in a review.
Which framework to use
| Your question | Framework | Typical review |
|---|---|---|
| Does X work, or what is its effect on Y? | PICO: population, intervention, comparison, outcome | Systematic review, meta-analysis, rapid review |
| What is known about X, and how has it been studied? | PCC: population, concept, context | Scoping review, mapping review |
| How do people experience or view X? | SPIDER: sample, phenomenon of interest, design, evaluation, research type | Qualitative evidence synthesis |
The scoping review vs systematic review guide explains the choice in more detail.
The criteria builder (copy this)
Copy the table for your framework. Write one include line and, only where needed, one exclude line per row. Add the last four rows to every review.
PICO builder
| Element | Include | Exclude | How a screener tests it |
|---|---|---|---|
| Population | Who was studied, age, condition, setting | ||
| Intervention or exposure | What was done or what people were exposed to | ||
| Comparison | What the intervention group was compared with | ||
| Outcomes | Which outcome, measured how, and when | ||
| Study design | From the methods section | ||
| Publication type | Full paper, abstract only, editorial, protocol | ||
| Language | Only if limited | ||
| Dates | Only if limited, with a reason |
PCC builder
| Element | Include | Exclude | How a screener tests it |
|---|---|---|---|
| Population | Who the source is about | ||
| Concept | The idea, intervention or phenomenon being mapped | ||
| Context | Setting, country, sector or time | ||
| Source types | Studies only, or also reports, guidelines, theses |
SPIDER builder
| Element | Include | Exclude | How a screener tests it |
|---|---|---|---|
| Sample | Who took part | ||
| Phenomenon of interest | The experience, behavior or decision studied | ||
| Design | Interviews, focus groups, ethnography, observation | ||
| Evaluation | Views, experiences, attitudes, meanings | ||
| Research type | Qualitative or mixed methods with a separate qualitative analysis |
Worked examples (invented)
All three are invented examples for illustration.
PICO: night-shift naps
Question: In adult shift workers, does a planned nap during the night shift improve end-of-shift alertness compared with no nap?
| Element | Include | Exclude |
|---|---|---|
| Population | Adults (18+) working night or rotating shifts | Simulated shifts with no work duties |
| Intervention | A planned nap of any length during the shift | Naps taken only before the shift |
| Comparison | No nap, or breaks without sleep | No comparison group |
| Outcomes | Alertness at end of shift (PVT, KSS) | |
| Study design | Randomized and non-randomized controlled studies, including crossover | Case series, cross-sectional surveys |
| Publication type | Full reports in journals or theses | Conference abstracts with no full paper |
PCC: digital tools for medication adherence
Question: What digital tools have been studied to support medication adherence in older adults living in the community? (The topic of the scoping example in scoping review vs systematic review.)
| Element | Include | Exclude |
|---|---|---|
| Population | Adults aged 65 or over | Studies where results for over-65s cannot be separated |
| Concept | Any digital tool aimed at medication adherence: apps, text reminders, smart pill boxes | Tools for prescribing staff only |
| Context | Community and home settings | Hospital inpatients, care homes |
| Source types | Primary studies of any design, reviews, government reports | Opinion pieces |
SPIDER: nurses' experience of night-shift naps
Question: How do nurses experience taking planned naps during night shifts?
| Element | Include | Exclude |
|---|---|---|
| Sample | Registered nurses working night shifts | Mixed staff groups where nurses' views cannot be separated |
| Phenomenon of interest | Taking, or being unable to take, planned naps at work | Sleep outside work hours |
| Design | Interviews, focus groups, open-ended survey questions analyzed qualitatively | Closed-question surveys |
| Evaluation | Experiences, views, barriers and facilitators | |
| Research type | Qualitative and mixed methods with a separate qualitative analysis | Quantitative only |
Rules that make criteria work
- One idea per line. "Adults with insomnia treated in primary care" is three criteria. Split them so a screener can tell which one failed.
- Testable from the text. A screener must be able to answer Yes, No or Unclear from the abstract or methods. "Clinically relevant outcomes" is not testable; "alertness measured with the PVT or KSS" is.
- No quality criteria. Don't screen for "high-quality studies". Assess risk of bias after inclusion, so readers see the whole evidence base. See levels of evidence for how designs are ranked.
- Exclusions add information. Use them for edge cases, not to repeat the inclusion line in the negative.
- Justify limits. State and justify every language, date or country limit in the protocol.
- Pilot, then fix. Two reviewers apply the criteria to 50 to 100 records. Rewrite any criterion they disagree on before formal screening, and date the change.
How to use the criteria in AI for Research
In the Systematic Review workspace, the criteria you enter drive the protocol draft, the AI screening suggestions and the methods draft.
- Choose New review, enter the question and pick the Type of review. The framework follows the type: PICO for systematic, rapid, meta-analysis and living reviews; PCC for scoping, mapping and realist reviews; SPIDER for qualitative evidence synthesis. You can switch it.
- Fill the framework fields. PICO gives Population, Intervention or exposure, Comparison and Outcomes; PCC gives Population, Concept and Context; SPIDER gives Sample, Phenomenon of interest, Design, Evaluation and Research type.
- Type your criteria under Include when and Exclude when, one per line, up to 40 of each.
- On the Search tab, the PubMed search box starts from your framework fields, each in parentheses, joined by AND. Rewrite it into a proper search string with synonyms and subject headings.
- When you screen, Suggest for the next 25 (1 credit) reads each record against these criteria and suggests include, exclude with a reason, or maybe.
| What you enter | What uses it | Credits |
|---|---|---|
| Question, framework fields, Include when, Exclude when | The protocol draft in PROSPERO's headings | 3 per draft |
| The same criteria | AI screening suggestions with a reason | 1 per 25 records |
| Framework fields | The starting PubMed search string | 0 |
| Criteria, searches and counts | The methods section draft | 3 per draft |
Copy the finished criteria into the systematic review protocol template and the screening forms.
What this builder and the workspace do not do
- They do not choose your criteria. The builder gives structure and the AI drafts from what you type; the eligibility decisions are yours and your team's.
- The starting search string is a rough start. It joins your framework words with AND; it does not add synonyms, MeSH terms or field tags. See systematic review search strategy.
- AI suggestions read titles and abstracts only. They can't apply a criterion that needs the methods section, so expect more Maybe suggestions on design and outcomes.
- Other frameworks are not built in. PICOS, PICOT, PICo, ECLIPSE and SPICE aren't framework options in the workspace; put their extra elements in the Include when and Exclude when lines.
Who it is for
Master's students planning a review dissertation (AI for a master's thesis), PhD candidates (AI tools for PhD students) and clinicians writing a review protocol (AI for medical research). The full workflow is on the systematic review software page.
Learn more
- How to write a research question: PICO, FINER and SPIDER with examples.
- What is a systematic review: where criteria sit in the eight steps.
- Scoping review vs systematic review: PCC and when to map instead of answer.
Last updated 2026-11-12