Anesthesiology Residents: EBM Review and Practice

EBM integrates:

• the best available research evidence;
• the clinician's expertise;
• and the patient's values & preferences.

Standard practices can change over ​time
in response to new research, so ​it is very
important to stay current with ​the
research in your field and to search
​journal databases for topics that are
​undergoing re-evaluation.

There are 6 steps in the evidence-based practice process:

ASSESS the Patient

Start with the patient---a clinical problem or question arise from the care of the patient​

ASK the Question​

Construct a well-built clinical question derived from the case using the PICO framework​

ACQUIRE the Evidence​

Select the appropriate resource(s) and conduct a search​

APPRAISE the Evidence​

Appraise the evidence for its relevance, reliability, validity, and applicability​

APPLY the Evidence to the Patient

Integrate the evidence with clinical expertise, patient preferences, and apply it to practice

EVALUATE the Process and Outcomes

Did the evidence help the patient?  Does the process need to be changed or updated?

Some clinical questions can be answered or informed by using textbooks and other synthesized resources. These resources provide a quick way to determine what is known about a topic at a specific point in time, but they may not be as current as journal articles.

Sources for background information:

To turn a clinical question into a search strategy:  

• Ask a specific question with PICO​.
• Use a database related to your topic.​
• Develop a search to answer your question.​
• Identify the levels of evidence​.
• Find full text of the article through Find@UNC.

PICO is a widely-used acronym to assist in remembering the key components of a clinical question.  PICO clarifies and focuses questions that arise during a patient assessment.

It identifies and organizes the main ideas of a complex patient presentation: 

P = Patient or Population;

I = Intervention or Issue;

C = Comparison or Control (not part of all questions);

O = Outcome.

You can also add Type of Question, Type of Study, or Setting to the PICO framework to create PICOTTS.

Not all parts of PICO are required!

PICO is a framework to help you narrow your topic.​ You do not have to use all of the letters in PICOTTS to have a specific question.

There are 38 varieties of question frameworks like PICO.

Question frameworks vary by discipline, methodology, and by the type of question you are asking.  If one framework doesn’t fit, try another!  Learn more about the frameworks in the article Formulating questions to explore complex interventions within qualitative evidence synthesis.

Evidence hierarchies are an estimate to help you filter your searches to the most-likely
best* evidence for the kind of question you are asking.

*Remember that all evidence must be critically appraised. A poorly-conducted or -reported randomized trial does not provide stronger evidence than the results of a well-conducted cohort study.

Filter Search Results by Appropriate Study Design

Look for articles in the search result that have the highest level of evidence for your type of question.

Small result sets can be filtered “by eye”.​ For larger result sets, use search filters.

• For prevention or treatment questions, start by searching for evidence at the top of the list, systematic reviews of randomized trials.

If no evidence is found at the top levels, move down the list looking for systematic reviews and then single studies of first cohort studies and then case-series or case-control studies.

• For etiology/harm or prognosis questions, start with cohort studies.

• Consider the publication date in the selection process.

If the systematic review you find was published a number of years ago and found inconclusive evidence, then look for newer randomized trials as your next step.

Curated evidence resources​

These resources focus on gathering evidence.  You still need to appraise their work!

• Cochrane Database of Systematic Reviews
• TripDatabase.com​
• JAMAevidence​

Evidence "in the wild"​

Sometimes evidence has not been gathered into summary document or a database.  If you use the resources above and cannot locate evidence sources, try:

• Databases such as Pubmed, Embase, PsycInfo, etc.​
• Professional organizations' websites​
• HSL’s Finding Guidelines Guide

Use HSL Resources page to find more health databases.

How should I format my search?

Select words from the PICO concepts to guide the development of a PubMed search. Remember: ​

• Start with key concepts.

You do not need to use all of the concepts or all the words.  Find MeSH terms if applicable.​

• Include alternate word endings, especially for key concepts.

List variations of a word, because the database will search for exactly what you type.  Use an asterisk to search for words that have the same base but different variations at the end.  Don't shorten too much or you may get an error or unexpected results.

For example: (prevent OR preventing OR prevented OR prevention OR preventative)​

For example: ed* will return education terms, but it will also give you edema, emergency department, edition, editor, edge, edible, and other unrelated words.  educat* will search for education, educating, educator, educated, educational

• Include synonyms.

Synonyms will help expand your search and give you more options.  Use the word OR between each synonym.  Use parenthesis around each synonym group to help the database understand how to organize your search.

For example: (therapy OR therapeutic OR treatment)​ AND (cancer OR cancers OR neoplasm*)

• Include acronyms and alternate spellings.

Include possible acronyms as synonyms for your search terms with OR in between.  Beware of acronyms with multiple meanings.  Use alternate spellings (such as hematology OR haematology).

For example: "enhanced recovery after surgery" OR ERAS

For example: MI could refer to "myocardial infarction" or Michigan, mile, middle initial, etc.

• Use double quotes for phrases.

Search for phrases in double quotes so the database keeps them together as one concept.

For example:  "quality of life" or "length of stay"

• Apply filters as needed. 

Filters, such as article types, gender, or age limits, are usually added after a preliminary search to narrow the results.​  You can also search for some of these filters with synonyms

For example: (child OR children* OR pediatric* OR teen* OR preteen* OR adolescen* OR youth*)

Example: Question to PICO to search

Question:  Are electric toothbrushes better at removing plaque from teeth?

P = (humans with teeth)

I = electric toothbrush

C = manual toothbrush

O = plaque

Search:  Electric AND Manual AND Plaque AND (Toothbrushes OR Toothbrushing OR Toothbrush)

How to appraise an article

Once you have gathered the evidence, you will need to appraise the evidence for its relevance, reliability, validity, and applicability​

Ask questions like:

Relevance:  ​

• Is the research method/study design appropriate for answering the research question?​
• Are specific inclusion / exclusion criteria used? ​

Reliability:  ​

• Is the effect size practically relevant? How precise is the estimate of the effect? Were confidence intervals given?  ​

Validity: ​

• Were there enough subjects in the study to establish that the findings did not occur by chance?    ​
• Were subjects randomly allocated? Were the groups comparable? If not, could this have introduced bias?  ​
• Are the measurements/ tools validated by other studies?  ​
• Could there be confounding factors?   ​

Applicability:  ​

• Can the results be applied to my organization and my patient?   ​

Use appraisal tools

Worksheets & Checklists

EBM calculators

Types of common EBM calculations include:

• Control Event Rate (CER)
• Experimental Event Rate (EER)
• Relative Risk (RR)
• Relative Risk Reduction (RRR)
• Absolute Risk Reduction (ARR)
• Number Needed to Treat (NNT)
• Confidence Interval (CI)

You can learn how to calculate these manually through the University of Alberta's EBM Toolkit or use Evidence-Based Medicine Toolbox's EBM Calculators & Vanderbilt's Power & Sample Size Calculator