Skip to main content

Evidence Based Practice by Elaine Cox

Evidence Pyramid: Dartmouth & Yale

More Pyramids

Take a look at the following websites that offer unique versions of the Evidence Pyramid.

Determing Levels of Evidence

Levels of Evidence
are determined by the following:

1. Strength of Evidence: confidence that the evidence is a true measure

2. Quality of Evidence: minimization of bias

3. Statistical Precision: degree of certainty about the existence of a true measured effect

4. Size of Effect: how much of an effect is above 'No Apparent Effect' for clinically relevant benefits (no harms)

5. Relevance of Evidence: appropriateness of the outcome measure; usefulness in measuring benefits / harms

Highest Level of Evidence -- systematic review of all relevant randomized controlled trials (RCTs).

Next Highest Level of Evidence -- at least one RCT on effectiveness
And so on... e.g. a pseudo-randomized trial  that assigns participants by alternating between groups by date of admission

Lower Levels of Evidence --
    non-randomized studies with control group running concurrently with intervention group
    non-randomized studies with intervention effects which are compared to previous / historical information
    single-case studies

Confidence Interval (CI):
Covers the likely range of true effect. (play of chance can skew results)

Rating the Evidence

EB Practitioners must determine the strength of the evidence they find in the literature. Systems for rating the strength of evidence vary widely.