An evidence-based practice workflow: from PICO to review
Turn a practical uncertainty into a focused question, find and appraise evidence, then assess what happens when it is applied.
Evidence-based practice begins with a concern that really makes a difference in a particular context. That concern may be about whether an intervention has an impact, if a test provides accurate results, how patients/clients experience their care, or how services are organised. The goal is to take that concern and transform it into a researchable question.
Research findings are one part of the solution. Professional expertise, local circumstances and the preferences of affected people also contribute to the decision. Before a research finding is used in a practical decision, it needs interpretation.
Ask a focused question
PICO is a useful structure for many intervention questions. It identifies the population or problem, the intervention, the comparison and the outcomes. Specify timing and setting where they are important. The purpose is to clarify the decision, rather than force every uncertainty into the same template.
For a hypothetical service example, consider adults attending community rehabilitation appointments. The service is considering a text reminder alongside its usual appointment letter, compared with the letter alone. Its main outcome is the proportion of missed appointments during three months. It also wants to consider acceptability and whether the system disadvantages people without reliable phone access.
That is more searchable than asking whether reminders work. It also shows that a favourable attendance result would not settle every concern about introducing the system.
Different questions need different kinds of evidence. A question about lived experience may call for qualitative research. A question about future outcomes may require a prognosis study. A test-accuracy question needs an appropriate comparison with a reference standard, rather than simply the design used for an intervention trial.
Find evidence efficiently
Before examining individual studies, look for a relevant, reliable and sufficiently current synthesis. A systematic review or guideline can help direct the search. Examine its scope, search date and methods, however: a promising title does not establish that it addresses the focused question.
Formulate the search using key concepts from the question and suitable synonyms. Not all PICO elements need to be included; too many restrictions may exclude useful studies. Describe the sources searched, including relevant databases and registers, the date, the main search terms and any limits so others can understand the process.
Describe a short, time-limited search for practice honestly. Do not label it a systematic review unless its methods justify the term. Useful evidence may also be missing, inaccessible or inconsistent.
Appraise before applying
Consider whether the research design fits the question and whether its conduct could have introduced bias. Review the effect size and its uncertainty alongside the methods. Check whether outcomes were predetermined, whether participants were lost to follow-up and whether the statistical analysis was appropriate for the study design.
A comparison between two clinics may be affected by differences in staffing, location or the people attending them. Randomisation may reduce some concerns, but assessment still needs to address missing records, implementation and the definitions of measured outcomes.
When appraising a review, examine how studies were found, selected, assessed and synthesised. Certainty in a body of evidence is a separate judgement from the trustworthiness of individual studies. An impressive pooled estimate can still be based on weak underlying evidence.
Consider the local decision
Compare the research setting with the service where the findings might be used. Identify potential costs, workload, access requirements and unintended consequences. Include the views of people who would use or receive the service.
If the evidence suggests an effect, translate it into absolute terms using an appropriate baseline. Keep the outcome definition and time period visible. A percentage without its denominator or follow-up period can create a misleading impression.
Review what happens
Implementation should include an assessment plan. Decide in advance what will be measured, which groups should be considered and when results will be reviewed. Keep an eye on benefits and burdens that the main outcome may miss.
Suppose the fictional service's missed-appointment rate falls from 25 in 100 to 18 in 100 after reminders begin. That is a useful observation, but the before-and-after change alone does not establish that reminders caused it. Season, scheduling changes or a different patient mix could also contribute.
A practical evidence record should preserve the question, search, appraisal, decision and review plan. It makes later updates easier and allows a team to see which part of the reasoning has changed when new evidence or different circumstances emerge.
Methodology references
- Oxford CEBM: Asking focused questions
- Oxford CEBM: Finding the evidence tutorial
- Sackett et al. Evidence based medicine: what it is and what it is not