|
-
Observational studies tend to overestimate treatment
effects.
-
Hierarchy of evidence is not absolute.
-
If treatment
effect is sufficiently large and consistent, then sometimes RCTs
are appropriate. (Example: Pap smear screening virtually eliminates
100% of advanced cervical cancer in women who are screened appropriately.
The evidence for performing PAP smears was all derived from retrospective
cohort studies.)
-
A full description of the levels of evidence with
definitions can be found at The Center for Evidence-Based Medicine. http://www.cebm.net/levels_of_evidence.asp
|