bonferroni: friend or foe? multiple testing in cardiovascular medicine
DESCRIPTION
Bonferroni: Friend or Foe? Multiple Testing in Cardiovascular Medicine. Dhruv S. Kazi, MD, MSc AHA Cardiovascular Outcomes Research Fellow Stanford University [email protected]. “Off hand, I’d say you’re suffering from an arrow through your head, but just to play it safe, let’s get an echo.”. - PowerPoint PPT PresentationTRANSCRIPT
Bonferroni: Friend or Foe?Multiple Testing in Cardiovascular Medicine
Dhruv S. Kazi, MD, MScAHA Cardiovascular Outcomes Research FellowStanford [email protected]
“Off hand, I’d say you’re suffering from an arrow through your head, but just to play it safe, let’s get an echo.”
Deat
h fro
m C
ardi
ovas
cula
r Cau
ses,
No
nfat
al M
yoca
rdia
l Inf
arct
ion,
or S
troke
= 9 billion dollars
Yusuf, S, et al. N Engl J Med 2001;345:494-502
CURE
Liver
Clopidogrel
Cytochrome P450-dependent oxidation
Binds to P2Y12 Receptor on Platelets
Ticagrelor Binds to P2Y12 Receptor on Platelets
CYP2C19
Potential Strategies
Clopidogrel Ticagrelor
Which one would you want?
Cannon, CP, et al. Lancet 2010; 375: 283-93.
Primary Efficacy Endpoint in the PLATO-Invasive RCT
Cannon, CP, et al. Lancet 2010; 375: 283-93.
Methods
Cohort: 100,000 patients who present with ACS and undergo PCI, age at entry – 65 years
Analytic Horizon: Lifetime
Perspective: “Ideal Insurer”
Interventions– DAPT 12 months from last ACS or PCI, whichever is
later– Aspirin monotherapy for life thereafter
Possible Explanations?
Possible Explanations?
True Difference Chance Finding
Fraud?
The Multiple-Look Problem
Number of analyses
Cum
ulati
ve p
rob
of
a po
sitive
ass
ocia
tion
0 10 20 30 40 50 60 70 80 90 1000
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
So how do you get around this?
Traditionally, “don’t run multiple subgroups” unless:- The analyses are pre-specified- The analyses are biologically plausible
And if you must, conduct rigorous statistical adjustment!
Bonferroni Adjustment
Conservative Assumes independence 1-(1-α)1/n ~ α/n
But does this make sense?BMJ. 1998 April 18; 316(7139): 1236–1238.
How Do We Proceed? (Do you still want the drug?) Multiple testing is problematic (even if pre-specified) The challenges of a priori hypotheses
Conclusions
Multiple testing is a complicated question: with real clinical consequences
Statistical adjustment is a necessary but imperfect solution
Trial and Error. Kaul S, et al. J Am Coll Cardiol 2010;55:415–27
Conclusion
The p value is no substitute for a brain.
Thank You!