Evidence based medicine (EBM) is less than 10%

• The double blind randomized controlled trial (RCT) is mathematically correct. If technically well performed it will exclude inclusion bias (randomised controlled) , exclude placebo effect or woman bias (blinded) , and observer bias (double blinded).
• The value of all other studies, as cohort or others studies needs a close evaluation of material and methods and ranges from excellent to poor.

Conclusion

• What is  proven should not be discussed unless there is a technical problem in design or in blinding
• What is not clear is ‘not proven’, but should not be dismissed as ‘not true’
• The RCT
  • is tedious
  • is slow in adding information
  • Is not suited for rare events; eg a complication occurring in 1% of surgery needs a RCT of 6000 women in order to have 30 cases in the control group and  30 in the test group. This is rarely realistic.
  • The more strict inclusion criteria are, the more difficult ithe extrapolation of results becomes
  • A RCT therefore is not suited for multimorbidity, eg the patient with deep endometriosis, AND diabetes AND hypertension AND …….
  • A RCT balances between evidence which is already so likely that it is close to unethical to perform the trial and a difference which is so little that the result is clinically not important
  • RCT are of limited value when blinding is impossible eg for medical treatment of endometriosis

Authority based expert opinion

• More than 90% of medicine is experience based expert opinion, which integrates what has been demonstrated with the characteristics of the individual patient ie age, weight  other diseases etc.
• Experience and rules are often introduced following (near) accidents seen in the past  e;g;  explosion risk with pure N2O  during laparoscopy.
• These authority based rules which are a consensus opinion of experts are undervalued and should be upgraded in the pyramid of evidence

Clinical decision making

For diagnosis= diagnostic accuracy of a test

• consider sensitivity and specificity and the many pittfalls  as
  • the low predictive value which is low when prevalence is low
  • precision of the test
• what does the test evaluate ? especially important for imaging in endometriosis

For treatment

• Treatment balances the expected benefits with the risks of treatment
• Freedom of treatment is limited to what is accepted at a certain point in time
• a new type of surgery should be performed as a trial with close evaluation
• treatment contract should be explained in informed consent

Therefore who gives information is important 

• Check publications 
• check presentation 

Do not trust personal opinions 

Prof P.R. Koninckx

25-9-2020