April 13, 2004 Feature

Levels of Evidence

The evidence in evidence-based practice (EBP) may take many forms ranging from expert opinion to meta-analysis. Each form of evidence, though, is not equally persuasive in making the case that a certain clinical procedure should become an aspect of recommended care for members of a certain clinical population. The greater the scientific rigor in producing clinical evidence, the more potent is that evidence for influencing the formation of policies affecting clinical practice.

The literature bearing on a certain clinical procedure is inventoried for: the relevance of findings, the quality of findings, the number of findings, and the consistency of findings for establishing a clear and singular linkage between a certain clinical outcome and a certain clinical procedure applied to members of a certain clinical population.

The terms "levels of evidence" or "strength of evidence" refer to systems for classifying the evidence in a body of literature through a hierarchy of scientific rigor and quality. Several dozen of these hierarchies exist (Agency for Healthcare Research and Quality [AHRQ], 2002b). Some systems comprise three levels and others eight or more. The gradations in some hierarchies are based on randomization and experimental controls. The organizing focus for others may center on magnitude of effect sizes, confidence intervals, number of results, consistency of results, sample size, or Type I and Type II error rates (AHRQ, 2002a). In each application of EBP process, reviewers must select the most relevant levels-of-evidence system for the type of procedure being assessed (e.g., measurement technologies, diagnosis, prognosis, safety, efficacy, and effectiveness).

In the United States, the recognized authority regarding the assessment of scientific clinical research is AHRQ and the system in the "Example Levels of Evidence" table below is one used by AHRQ (2001). Some international organizations adapt AHRQ systems (e.g., Scottish Intercollegiate Guidelines Network); others use wholly different systems (e.g., World Health Organization).

Randall R. Robey, is an associate professor and director of the Communication Disorders Program in the Curry School of Education at the University of Virginia. His research combines two areas of concentration: clinical aphasiology and quantifying valid indices of change brought about by clinical interventions. Contact him by e-mail at rrr7w@virginia.edu.

cite as: Robey, R. R. (2004, April 13). Levels of Evidence. The ASHA Leader.

Example Levels of Evidence

Sources of Evidence 

Meta-analysis of multiple well-designed controlled studies


Well-designed randomized controlled trials


Well-designed non-randomized controlled trial (quasi-experiments)


Observational studies with controls (retrospective studies, interrupted time-series studies, case-control studies, cohort studies with controls)


Observational studies without controls (cohort studies without controls and case series)



Agency for Healthcare Research and Quality. (2001). Making health care safer: A critical analysis of patient safety practices. Evidence Report/Technology Assessment: Number 43. AHRQ Publication No. 01-E058. Rockville, MD: Agency for Healthcare Research and Quality (www.ahrq.gov/clinic/ptsafety/).

Agency for Healthcare Research and Quality. (2002a). Rating the strength of scientific research findings. AHRQ Publication No. 02-P022. Rockville, MD: Agency for Healthcare Research and Quality (www.ahrq.gov/clinic/epcsums/strenfact.htm).

Agency for Healthcare Research and Quality. (2002b). Systems to rate the strength of scientific evidence. Summary, evidence report/technology assessment: Number 47. AHRQ Publication No. 02-E015. Rockville, MD: Agency for Healthcare Research and Quality (www.ahrq.gov/clinic/epcsums/strengthsum.htm).


Advertise With UsAdvertisement