Section 3.3 What are Evidence Taxonomies?

This section is on exploring the concept of “levels of evidence” or evidence taxonomies in describing types of literature and understanding why levels or hierarchies of evidence are important within evidence based medicine.  

The learning objectives are to recognize the relationship between types of literature in terms of evidence, explain the differences in evidence taxonomies, and understand that the best evidence may not be the highest literature type in the hierarchy.

Evidence taxonomies, often referred to as levels or hierarchies of evidence, have been created to better understand the quality and strengths of different types of literature, particularly types of research studies. They describe a rating or classification system of the research literature and are used to encourage clinicians to find and apply the highest level of knowledge or evidence to answer their clinical questions. (Burns, et al., 2011) They assist in selecting the best research evidence for answering specific clinical questions and patient situations. Some “evidence levels” or “hierarchies of evidence” provide a visual representation of the layered or stepped transition between low levels of evidence and high levels of evidence. 

For instance, the Evidence Pyramid from Dartmouth Biomedical Libraries shows expert opinion as the base of the pyramid and meta-analysis as the top of the pyramid, noting a distinction between primary research and synthesized research. 

[Image of the Dartmouth Evidence Pyramid: A diagram of scientific research represented as pyramid by Dartmouth Biomedical Libraries. The lower portion of the evidence pyramid reflects primary research, and the upper portion reflects synthesized research.]

In comparison, the Duke University Medical Center Library Evidence Pyramid, has animal and laboratory studies at the base of the pyramid and meta-analysis at the top. The Duke pyramid also notes the differences between observational studies, experimental studies and synthesized evidence.

[Image of the Duke Evidence Pyramid: An illustration of the Evidence Pyramid by Duke University Medical Center Library and Archives, highlighting the quality of information, progressing from observational studies through experimental studies to synthesized evidence.]

What is important with any evidence taxonomy illustrating levels or hierarchies is that there is a distinction between the types of literature and synthesis of research reflective of the strength of the research. In terms of evidence, the synthesis of research studies is considered to provide higher evidence. Evidence syntheses, like systematic reviews and meta-analysis, are often used in establishing the standard of care or considered as the “gold standard” of practice. DiCenso, Bayley, and Haynes, further reinforce the value of syntheses of studies, synopses of syntheses, and even summaries of evidence in their 6S Pyramid of Preappraised Evidence.

[Image of 6S Pyramid of Preappraised Evidence: Image of the DiCenso, Bayley, Haynes 6S Hierarchy Pyramid of Preappraised Evidence, moving from original studies at the base through systems at the top, with examples of sources for each represented “S” in the pyramid.]


In some taxonomies, evidence is presented as “grades,” like the recommendation grades outlined by the United States Preventive Services Task Force that considers the certainty of benefit of a service or prevention method, balanced against the potential harms of the same service. An “A” grade represents a “high certainty that the net benefit of the service is substantial,” whereas an “D” grade recommends against the service due to a “moderate or high certainty that the service has no net benefit or that the harms outweigh the benefits.”

[Table image of USPSTF Recommendations: Table of US Preventative Services Task Force Recommendations of Services, starting with high certainty that the net benefit is substantial, Grade A, through recommendations against, Grade D, and insufficient evidence, Grade I.]

However, the Strength of Recommendation Taxonomy, also known as SORT, defines the grades “A,” “B,” and “C,” differently. SORT recommendations focus on the consistency of the recommendation and the quality of patient-oriented evidence, that is, evidence that matters to patients, such as “morbidity, mortality, symptom improvement, cost reduction, and quality of life.”

[Table image of SORT: Recommendation Taxonomy (SORT) table representing grades A, B, and C, based on consistency or inconsistency of evidence, or consensus, disease-orientated evidence.]

Recognizing the differences in what the evidence taxonomy represents is essential for understanding the significance or confidence of the evidence being reported on.

A more recent taxonomy, the Evidence Funnel, developed by the University of Iowa Hospitals and Clinics, focuses on the evaluation and application of the evidence, moving from the quality of research studies, the strength of the recommendations, and certainty of benefit to the resulting change in practice based on the evidence.

[Image of the Evidence Funnel: An image of the Evidence Funnel from the University of Iowa Hospitals and Clinics inverting a standard evidence pyramid into a funnel, narrowing the quality of research and strength of recommendations to a point of practice change through evaluation and application of the evidence.]

Regardless of how the relationship between research studies is described, it is important to be mindful of what each taxonomy represents and how the evidence is being characterized. Additionally, being aware that a particular clinical tool or resource may refer to more than one taxonomy or level of evidence is equally important. For example, Essential Evidence Plus, an evidence-based reference tool that includes evidence-graded recommendations, makes use of four different evidence grading taxonomies– US Prevention Services Task Force, SORT, Centre for Evidence-Based Medicine, and GRADE. Also, from one clinical practice guidelines to another, different evidence taxonomies may be applied to the recommendations. Thus, it is essential to recognize which evidence taxonomy is being referenced in a clinical tool or resource to be able to accurately interpret the recommendations provided for clinical practice.

While seeking the highest level of evidence is considered best practice, keep in mind that the best evidence for a particular patient situation may not be found at the top of the evidence taxonomy. In some cases, higher study types, like randomized clinical trials, are not possible due to ethical considerations in the research design, rarity of the condition or disease, or lack of original research studies. Therefore, a retrospective study or even a systematic review of retrospective studies may represent the best evidence available for the clinical situation.

This section has explored different evidence taxonomies representing the relationship between types of literature in terms of evidence, as well as the importance of recognizing which evidence taxonomy is being referenced within a specific resource or practice guideline. Additionally, this section has highlighted that the best evidence for a particular clinical situation may not be what is considered the highest level of evidence in a hierarchy.  

The resources referenced in this section are provided in this reference list.