Skip to Main Content


This guide contains resources from the University Library to support your study and research in Medicine.

Evidence-based medicine

Evidence-based medicine (EBM) refers to the application of the best available research to clinical care, which requires the integration of evidence with clinical expertise and patient values. By best available research, we mean clinically relevant (i.e. patient oriented) research that:

  • Illuminates the accuracy and precision of diagnostic tests,
  • Highlights the importance of prognostic markers,
  • Establishes the efficacy and safety of therapeutic, rehabilitative, or preventive healthcare strategies, or
  • Seeks to understand the patient experience.

For more details on evidence-based medicine see the BMJ Best Practice webpage below:

Core EBM databases:

Asking clinical questions:

One of the basic skills required for practicing EBM is developing of well-built clinical questions. These questions need to be relevant to patients’ problems and phrased in ways that facilitate your acquisition of relevant and precise answers. Well-built clinical questions usually contain up to four elements. PICO is an acronym/mnemonic of these elements and it identifies and organizes the key aspects of a complex patient presentation:

  • Patient or Population and Problem
  • Intervention or Indicator
  • Comparison or Control (not part of all questions)
  • Outcome

Adding the two T's (Type of Question, Type of Study) to the PICO framework addresses that different types of study designs are used to answer different types of questions. 

When to use PICO?

Background questions concern general knowledge. These types of questions generally have only 2 parts: A question root (who, what, when, where, how, why) and a disorder, test, treatment, or other aspect of health care. Often these questions can best be answered by using a textbook or consulting a clinical database.

Foreground questions are specific knowledge questions that affect clinical decisions, including a broad range of biologic, psychological, and sociologic issues. These are the questions that generally require a search of the primary medical literature and that are best suited to the PICO format.

What does it mean to search for evidence-based literature? Isn't searching Google enough? The answer is No. Google, and other search engines "scrape" information on the Internet using proprietary algorithms. This means that the search engines control your search, you don't. This may not be a problem if you're searching for things to buy, consumer reviews, or websites of organizations. But since search engines find information of all kinds, much of it opinion based, you need to use another tool when you're conducting evidence-based research. Article databases focus on the professional, scientific literature and are usually indexed, so that you can create a search to more precisely find reliable scientific evidence on your topic.

The evidence hierarchy pyramid shows you how reliable different types of research are. The lower levels contain information that is based on less evidence, such as expert opinion or very early experiments. Higher up the pyramid, the amount of literature decreases, but it is based on more evidence and thus is more reliable. 

Primary Literature (Unfiltered Literature)

Case series / Case reports: reports on the treatment of one or more individual patients. There are no control groups to compare outcomes. This type of research has little statistical validity.

Case control studies: patients who already have a specific condition are compared with people who do not have the condition, looking back in time to identify factors that might be associated with the illness. Case control studies often rely on medical records & patient recall for data collection, so their reliability is limited.

Cohort studies: compare a group of patients who are already under a specific treatment with a similar group not affected by the treatment. Both groups are are followed over time. Cohort studies are observational, so their reliability is limited.

Randomized controlled clinical trials (RCTs): carefully planned experiments that look at the effects of a treatment on real patients in real time. To reduce the potential for bias, randomization and blinding are used in the selection of patients and treatments. Patients are assigned to treatment and control groups randomly; patients and clinicians (& sometimes laboratories) do not know who is in which group and thus who is receiving treatment or not. This allows the groups to be directly compared. RCTs can provide sound evidence of cause and effect and are considered the gold standard for experimental research.

Secondary Literature (Filtered Literature)

Systematic Reviews: answer a specific question, using the results of an extensive literature search that identifies studies with sound and similar methodologies. The studies are reviewed, assessed for quality and the results are summarized according to the standards of the question that is being reviewed.

Meta-analyses: thoroughly examine a number of valid studies on a topic and mathematically combine the results using statistical methodology to report the results of the analysis as if it were one large study. 

Searching for evidence-based medicine in PubMed (Medline)
*Note: Medline is included in PubMed. When you search PubMed, you are searching Medline

When you have performed a search on a clinical topic in PubMed you will find that there are several publication types that indicate that an article is useful for evidence based practice. If you limit your search to include each of these publication types you should get references to most articles indexed in PubMed that are appropriate for evidence based practice.

  • Clinical Trial, Phase III
  • Clinical Trial, Phase IV

Both of these experimental designs use a control group, as does

  • Controlled Clinical Trial

is a statistical analysis that allows the results of several studies to be pooled to provide more reliable EBP data.

Practice Guideline
is a list of evidence based serial statements to assist in diagnosis and treatment of particular conditions.

Randomized Controlled Trial
is the gold standard for evidence based practice.

Four study categories are provided:

  • Therapy
  • Diagnosis
  • Etiology
  • Prognosis

You can choose the emphasis to be either specificity or sensitivity

  • Specificity (a narrow search - mostly relevant articles but probably omitting a few)
  • Sensitivity (a broad search - most relevant articles but probably some less relevant ones too).

For details see the PubMed Clinical Queries Filter Table which explains sensitive (broad) and specific (narrow) searches and approximate equivalents in the PubMed query language.

There is also a Systematic Reviews filter (a pre-determined search strategy) which can be appended to any search.