The Evidence Hierarchy
It is helpful to see the landscape of study designs before diving into their definitions as they relate to clinical questions (like PICO).
The Evidence Hierarchy is one way* to view some (not all) study designs, where they fit in, and the level of evidence they might provide.
Systematic Reviews and Meta-analyses are typically positioned at the top because of the rigorous nature of their methodology in order to answer clinical questions.
*There is much debate over the use of the Evidence Hierarchy in the research community. One of the top reasons for the debate is the lack of fluidity in the structure of the hierarchy and the depiction of quality generally increasing or decreasing based on study type. Please note that this is just one way to view study designs in relation to one another and that quality is subjected to individual studies, their study teams, and the rigor with which the study is conducted.
Study designs
Here are a few study designs with their descriptions as they relate to the concepts in a PICO question and the type of clinical questions they can answer. In each case, a well-conducted SR of all available studies that address a clinical question is better than an individual study.
| Design | Description (P: Population, I: Intervention; O: Outcome) |
|---|---|
| SR / MA | Involve systematic literature searches to identify, appraise, and synthesize all relevant research relative to a specific P, I, or O. |
| RCT | Cohort studies where allocation to treatment and control groups is achieved by a random process. P → I → O(s) |
| Cohort | Begins with a group of people who do not have a condition, takes baseline measurements, then follows them over time to determine whether or not they develop the condition. I → O |
| Case-Controlled | Identify factors that may contribute to an existing medical condition by comparing subjects with that condition (cases) with patients who do not have the condition but are otherwise similar (controls). O → I |
| Case Series | Coherent / consecutive set of cases of a condition which derive from the practice of one or more health care professionals / settings. |
Clinical question types
Here are the clinical question types addressed above, an example to contextualize each, and the appropriate study design to answer the question.
| Question Type | Example | Study Design for Question* |
|---|---|---|
| Therapy | Is this intervention more effective than another? | RCT |
| Diagnostic Test | How accurate is this diagnostic test? | independent, prospective blind comparison to a gold standard |
| Prognosis | What is the likely outcome, progression, or survival time for this condition? | cohort study → case-control → case series |
| Harm / Etiology | What are the possible causes of this condition or state of affairs? | RCT** → cohort study → case-control → case series |
| Prevention | How to reduce the risk of this disease? | RCT → cohort study → case-control |
*In each case, a well-conducted SR of all available studies that address a clinical question is better than an individual study.
**RCTs can answer questions of harm when studies report adverse events, but these studies are never chosen with intent to cause harm.