Written and maintained by CASRAI Editorial Board
Last updated
The SPIDER framework breaks a qualitative research question into five components – Sample, Phenomenon of Interest, Design, Evaluation, Research type – so that a question about experience, meaning, or process can be searched in a bibliographic database the same rigorous way PICO structures a question about an intervention’s effect. It was developed by Cooke, Smith, and Booth and published as “Beyond PICO: The SPIDER Tool for Qualitative Evidence Synthesis” in Qualitative Health Research, volume 22, issue 10, pages 1435–1443, in 2012 (DOI: 10.1177/1049732312452938), specifically because PICO’s Population/Intervention/Comparison/Outcome structure does not map onto qualitative questions – a qualitative evidence synthesis is rarely testing an intervention against a comparator, and forcing that shape onto a question about lived experience distorts both the review question and the resulting search.
This guide is written for qualitative evidence synthesis authors and information specialists building a searchable protocol – not as a general study-skills introduction. It covers the five elements and where authors most often misapply them, a worked conversion from a vague qualitative topic to a full SPIDER question to an actual database search string, what the only published head-to-head evidence (Methley et al., 2014) actually found when SPIDER, PICO, and PICOS were tested against each other, and when SPIDER is the wrong tool in favor of SPICE or PEO.
The Five SPIDER Elements
Each element narrows the search and doubles as an inclusion criterion once studies are being screened for eligibility.
| Element | What it specifies | Where authors go wrong |
|---|---|---|
| S – Sample | Who or what is being studied – not a “population” in the epidemiological sense, but the specific group whose experience or perspective the review is synthesizing (e.g., first-time mothers, newly-diagnosed type 2 diabetics, hospice nurses). | Importing a PICO-style population statement wholesale, without narrowing it to the specific role or standpoint the qualitative literature actually reports from. |
| PI – Phenomenon of Interest | The experience, behavior, or event being explored – replaces PICO’s Intervention, since there is usually no administered intervention to isolate. | Phrasing this as an outcome to be measured (“improves adherence”) instead of a phenomenon to be understood (“experiences of managing adherence”). |
| D – Design | The research design used to collect the data – interview studies, focus groups, ethnography, case study, grounded theory, action research. | Treating Design as optional or leaving it too broad (“qualitative studies”), which recovers far more noise than a stated design term does. |
| E – Evaluation | The outcome of interest, expressed in qualitative terms – experiences, views, perceptions, attitudes, beliefs, satisfaction – not a measured clinical or quantitative outcome. | Reusing a PICO Outcome (a rate, score, or measured endpoint) instead of the subjective or interpretive outcome a qualitative study actually reports. |
| R – Research type | The methodological family the eligible studies must belong to – qualitative, mixed-methods, or (per the original 2012 paper) qualitative and/or quantitative depending on the review’s scope. | Confusing Research type with Design – Design is the specific data-collection method (interviews), Research type is the broader methodological category the study is published as. |
Worked Example: From Vague Topic to Search String
Start with a vague topic: “how new mothers experience breastfeeding support.” Run through the five elements before opening a database:
- Sample: first-time postpartum mothers
- Phenomenon of Interest: receiving breastfeeding support from health professionals or peer supporters
- Design: interview or focus-group studies
- Evaluation: experiences, perceptions, satisfaction
- Research type: qualitative
That becomes an answerable review question: “What are first-time postpartum mothers’ experiences of breastfeeding support, as reported in qualitative interview and focus-group studies?” Each element then becomes a concept block in the search strategy, combined with AND between blocks and OR within a block to capture synonyms and truncation:
(postpartum* OR "new mother*" OR "first-time mother*") AND (breastfeed* OR "breast feeding" OR lactat*) AND (support* OR "peer support" OR counsel*) AND (interview* OR "focus group*" OR qualitative)
Note what the string leaves out on purpose: there is no Comparison block, because SPIDER has none – a qualitative synthesis is not testing the phenomenon against a comparator condition. Evaluation terms (experience*, perception*, view*, satisf*) are often folded into the qualitative/Research-type block rather than built as a separate concept, since they tend to co-occur with the design terms in how these studies are indexed.
SPIDER vs. PICO: What the Evidence Actually Shows
The only published head-to-head comparison is Methley, Campbell, Chew-Graham, McNally, and Cheraghi-Sohi, “PICO, PICOS and SPIDER: a comparison study of specificity and sensitivity in three search tools for qualitative systematic reviews,” BMC Health Services Research 14:579 (2014). The authors ran the same qualitative review question through PICO, the modified PICOS (PICO plus Study design), and SPIDER across multiple databases and compared how many of the studies known to be eligible each search actually retrieved (sensitivity) against how much of what each search retrieved was actually relevant (specificity/precision).
The finding runs opposite to what many authors assume going in: PICO had the highest sensitivity (it retrieved the largest share of the known-relevant studies), SPIDER had the highest specificity (a larger share of what it retrieved was actually relevant, at the cost of missing more eligible studies), and PICOS sat in between – lower sensitivity and higher specificity than plain PICO, but lower specificity and higher sensitivity than SPIDER. The authors’ own recommendation for practice: use PICO when a fully comprehensive search is required and the resources exist to screen a larger, noisier result set; use PICOS when time and resources are limited; and treat SPIDER cautiously on its own, since its narrower retrieval carries a real risk of missing eligible studies, even though its precision is useful once a broader search has already run.
In practice this means SPIDER is best used to build the initial question and eligibility criteria – where its structure is genuinely better suited to a qualitative question than PICO’s is – while the search strategy itself may still benefit from broader, PICO-style term expansion (dropping the Design/Research-type filter, or running it as a separate, less restrictive line) rather than relying on a narrowly SPIDER-shaped string as the sole search.
When to Use SPIDER vs. SPICE or PEO
SPIDER is not the only qualitative-question framework, and it is not always the right one:
- SPIDER fits a question about a specific sample’s experience of a phenomenon, reported through a qualitative or mixed-methods design – the most common qualitative evidence synthesis question shape.
- SPICE (Setting, Perspective, Intervention/Interest, Comparison, Evaluation) fits better when the review is evaluating a service, program, or policy from a defined stakeholder perspective – it keeps a Comparison element, which SPIDER deliberately drops.
- PEO (Population, Exposure, Outcome) is a simpler three-element alternative sometimes used for qualitative questions with a clear exposure or context but without the design/methodology specificity SPIDER builds in.
If the review question already involves comparing a service model against an alternative, SPICE is the better fit. If the sample, phenomenon, and desired qualitative outcome are clear but the design doesn’t need to be specified up front, PEO is simpler to work with. SPIDER earns its extra complexity when Design is doing real work – for example, when the protocol specifically wants interview-based evidence and needs to exclude survey-based mixed-methods studies that would otherwise pass a looser filter.
Where SPIDER Fits in a Qualitative Evidence Synthesis Protocol
A SPIDER-formatted question is the starting point, not the whole protocol. PROSPERO registration and PRISMA reporting were both built around PICO-style quantitative reviews, and PROSPERO’s own form accepts a free-text review question rather than requiring the PICO fields to be populated – a SPIDER question registers cleanly the same way. Once studies are retrieved, downstream steps that a SPIDER-shaped search feeds into include appraising and reporting the included studies against qualitative-specific standards (COREQ and SRQR reporting checklists), extending the search into unindexed material (grey literature searching), and translating the resulting string across databases (PubMed-to-Embase search strategy translation). Authors deciding whether a qualitative synthesis is the right review type at all, rather than a standard systematic review, should start with systematic review vs. scoping review and the broader survey of qualitative research methods.
Frequently Asked Questions
What does each letter in SPIDER stand for?
Sample, Phenomenon of Interest, Design, Evaluation, Research type – developed by Cooke, Smith, and Booth (2012) as a qualitative-research alternative to PICO.
Is SPIDER better than PICO for a qualitative review?
“Better” depends on what the search needs to do. The only published head-to-head comparison (Methley et al., 2014) found PICO retrieves more of the truly eligible studies (higher sensitivity), while SPIDER returns a cleaner, more precise result set (higher specificity) at the cost of missing more eligible studies. Neither is unconditionally better – SPIDER’s advantage is a better-shaped question and eligibility criteria for a qualitative topic, not necessarily a better search on its own.
Can I use SPIDER for a mixed-methods review?
Yes – the original 2012 paper describes SPIDER’s Research type element as covering qualitative and/or mixed-methods studies, and the framework was designed with mixed-methods evidence synthesis in scope, not qualitative-only reviews.
Does PROSPERO require a PICO-formatted question to register a review?
No. PROSPERO’s registration form takes a free-text review question and objectives; a SPIDER-formatted qualitative question registers the same way a PICO question does – there is no requirement to force it into PICO’s fields.
When should I use SPICE instead of SPIDER?
When the review is evaluating a service, program, or policy against an alternative and needs a Comparison element – SPICE (Setting, Perspective, Intervention/Interest, Comparison, Evaluation) keeps that element; SPIDER deliberately does not have one.
