Evidence Based Medicine

You should cite this article if you use its information in other circumstances. An example of citing this article is:
Ronny Gunnarsson. Evidence Based Medicine [in Science Network TV]. Available at: http://science-network.tv/evidence-based-medicine/. Accessed May 27, 2017.

(This page is under construction)

The five steps in Evidence Based Medicine

  1. Define questions and level of evidence
  2. Systematic retrieval of publications
  3. Critical appraisal of evidence – If possible a meta-analysis
  4. Apply results in clinical practice
  5. Evaluate performance in clinical practice

Possible aims with a Meta-analysis

  • Show effect that single studies can’t (Forrest plots)
  • Show more studies not needed
  • Explain why studies give different results.

The staircase of evidence

Some study designs are more reliable than others in proving that a specific treatment is effective (or proving cause and effect):

Level Trustworthiness Study design
I Very low risk for bias (and incorrect conclusions) – Systematic review with meta analysis of high quality Randomized Clinical Trials (RCT)
II Low risk for bias (and incorrect conclusions) – High quality Randomized Clinical Trial (RCT)
III-1 Some potential for bias (and incorrect conclusions) Low quality Randomized Clinical Trial (RCT)
III-2 Clear potential for bias (and incorrect conclusions) – Controlled Clinical trial (CCT)
– Cohort study
– Case-Control Study
III-3 Considerable potential for bias (and incorrect conclusions) – Multiple Single Case Research Experimental Design – SCRED
– Historic cohort study
– Cross-sectional studies
– Ecological studies
IV May provide ideas but does not prove anything – One Single Case Research Experimental Design – SCRED
– Case studies / case series

There may of course be studies within each level of trustworthiness being of very good or very poor quality. Hence, a very good controlled clinical trial might be more trustworthy than a very poor quality RCT. However, as a general rule most randomized controlled trials are more trustworthy than most controlled clinical trials.

Potential problems with EBM

  1. Relevant study population?
  2. Publication bias?
  3. Studies described well enough?
  4. Sensible scoring – evaluation? (ex blinding)
  5. Extracting the right data?
  6. Managing conflict of interest?
  7. Presenting findings appropriately?

Do we have to require randomized controlled trials for everything  ?

Useful links

References

You should cite this article if you use its information in other circumstances. An example of citing this article is:
Ronny Gunnarsson. Evidence Based Medicine [in Science Network TV]. Available at: http://science-network.tv/evidence-based-medicine/. Accessed May 27, 2017.

Comments are closed.