Gå till huvudinnehållet

Search strategies for systematic reviews

Structure your research question with PICO, SPICE or PEO

A well-defined question is important to make sure that you know which studies to include and the search results are manageable. To structure your research question, you can use a models such as PICO, SPICE or PEO. PICO is used for quantitative studies, SPICE or PEO for qualitative studies.

PICO

P Population
I Intervention
C Control/comparison
O Outcome

Exemple:

Which treatment is effective for long term Covid-19?
Population
Persons with confirmed covid-19 and with long-term symptoms
Intervention Treatment for covid-19
Control/comparison No treatment or other treatments
Outcome Change of symptoms, quality of life, function etc.

SPICE

S Setting Where? The context for the question
P Perspective For Whom? The users, potential users, or stakeholders of the service
I Intervention What? The action taken for the users, potential users, or stakeholders
C Comparison Compared to what? The alternative actions or outcomes
E Evaluation Whit what result or how well? The result or measurement that will determine the success of the intervention

PEO

P Population Who?
E Exposure What problem?
O Outcome What in relation to the problem?

Exemple:

Nurses experience of psychological stress when caring for patients with covid-19.

Population Nurses
Exposure Psychological Stress
Outcome Experiences

Other examples is SPIDER (Setting, Perspective, Intervention, Comparison, Evaluation)

What is a controlled vocabulary?

In many of the subject databases the article content is described using a controlled vocabulary, thesaurus. A concept can be described with different synonyms, but the controlled vocabulary is set.

Example:

Heart attack

Studies ont heart attack, using different words to describe it, will be tagged with the subject heading Myocardial infarction. Thus searching the subject heading will retrieve articles using different synonyms, but they have all been tagged with the same subject heading.

Synonyms Subject Heading
Heart attack Myocardial infarction
Heart infarction  
Cardiac arrest  
Cardiovascular disease  
Myocardial infarction  

Note that the interdisciplinary databases like Web of Science or Scopus doesn't have a controlled vocabulary

Which parts of the research question should be included in the search?

Sometimes it can be difficult to decide which parts of PICO or SPICE to include in the search, it can also be differ between research questions. With a model like PICO, it is common to only search on Population and Intervention, because the Control/Comparison might not be mentioned in the abstract, and you don't want to define the outcome in your search strategy, but rather in your findings.

Example:

Which treatment is effective for long term Covid-19?
Population Persons with confirmed covid-19 with long term symptoms
Intervention Treatment for covid-19
Control/comparison No treatment or other treatment
Outcome Change in symptoms, quality of life or function
In the example above, a search would consist of the search "persons with covid-19" and "treatment for covid-19".
We would not search for Control/comparison or Outcome, but got back to this when it is time to do the relevance screening.Then "No treatment" or "other treatment" and "Change in symptoms", "quality of life" and "function" will be our inclusion criteria.