EBM for Dummies Flashcards
Statistically significant RCT
Meta-analysis or systematic review of RCT
Level 1
Lesser quality RCT
Meta-analysis or systematic review of RCT
Prospective cohort
Level 2
Retrospective cohort
Case-control study
Cross-sectional
Meta-analysis or systematic review
Level 3
Case study, case series
Level 4
Expert opinion
Animal studies
Cadaveric studies
Level 5
Ideal study for a PROSPECTIVE study?
RCT
Ideal study for RETROSPECTIVE study
Case-control
5 problems with EBM
No time Difficult to interpret Not useful Defined differently Statisticians have concerns
Determines how many patients to treat with selective modality until ONE patient shows benefit
NNT
Reciprocal of %success in tx group minus the %success in control group
NNT = 1 / (EER-CER)
What NNT value determines if a treatment is good?
NNT less than 4
Determines how many patients are needed to treat before harmful effects are seen
NNH
Reciprocal of the risk of bad outcome in tx group minus risk of bad outcome in control group
NNH = 1/ (EER-CER)
Negative NNH means?
The control condition is more likely to harm than the treatment condition
Higher the NNH the better or worse?
BETTER
Statistic of choice for case-controlled study that shows the odds of having the adverse outcome in the treatment group relative to having the adverse outcome in the control group
Odds Ratio
Odds ratios greater than, equal to, and less than 1 indicate?
>1 = bad is caused (NNH) =1 = neither <1 = bad is prevented (NNT)
Know how to use a 2x2 box for EBM
Divide the top 2 numbers, then divide the bottom two numbers. Finally, divide the 2 numbers you have and that is the odds ratio
What is PEER?
The rate of adverse events we would expect in a patient who received no tx or received conventional tx
How to calculate PEER
(# of control with adverse) / (# of control with and without adverse)
Utilizing PEER and the OR will do what?
Tell you NNT or NNH
