Unsafety: Making no mockery of honest ad-hockery
Post on 04-Jan-2016
48 Views
Preview:
DESCRIPTION
Transcript
Unsafety: Making no mockery of honest ad-hockery
Janet WittesStatistics Collaborative
ASA/FDA 2005
Topic du jour
Sleight of tongue
Game: Remove safety and add virtue in two easy steps
D S M C_ _ __ _ _ _
Sleight of tongue
Game: Remove safety and add virtue in two easy steps
D S M CD M C_ _ _ _
Sleight of tongue
Game: Remove safety and add virtue in two easy steps
D S M CD M CI D M C
Efficacy
Stop for “overwhelming” efficacy =0.05; power = 90%; four looks Probability of stopping early>70% Early stop: estimate pulls toward
null
Our current approaches
Prespecify Classify precisely Give lots of data Rely on mechanism
Divide and (un)conquer (e.g. Neuropathy)
Event T C
Neuropathic pain 1 0
Neuropathy 1 0
Neuropathy NOS 5 2
Neuropathy peripheral 2 0
…
…
…
e.g. Neuropathy
Event T C
Anosmia
…..
Autonomic neuropathy 1 0
…
Cranial neuropathy 2 1
…
…
e.g. Neuropathy
Event T C
…
Parathesia 3 2
Parathesia NOS 4 0
Parathesia other 0 1
…
Peripheral motor neuropathy 6 0
Peripheral sensory neuropathy 3 2
Other examples
Heart failure• Separate near synonyms• Allocate to heart and lung
Other examples
Heart failure• Separate near synonyms• Allocate to heart and lung
Bleeding: distribute over body systems
Consider mechanism
If you don’t get the drug, you can’t react to it• Eschew ITT: safety population• Modified Daley’s Rule: censor early and
often• Don’t collect extraneous information
Appeal to statistical conservatism
Populations
ITT mITT Safety: one dose of study med ATP Etc.
e.g., Vioxx- short follow-up
Through 36 months
With denominators
Bresalier et al., NEJM, Feb 2005
The denominators
Mo Rx Po0 1287 129912 1129 119518 1057 115624 938 104230 896 100136 727 835
No data dredging
We test hypotheses Too many type 1 errors if we
dredge
What we typically present
Current behavior• Same tables for interim & final
analyses• Long complete listings
What we should do• Interim data are different from final• Presenting too much dulls the mind
Sentinel events
Single event (e.g., death in a vaccine trial)
Several events (e.g., 3 retinal thromboses)
Sentinel event rate (e.g., WHI)
Lachenbruch and Wittes, SIM, to appear
How to handle
1. Identify sentinel event2. Establish stat’l method for future
events Need reasonable power Type I error rate > 1-sided 0.025
Methods
Individual• Number of non-events until the k’th
eventNegative binomial SPRT
• Time to the next (or k’th) event (-dist’n) Rate: event rate in future patients
NormalPoisson
Pitfalls
Time is subtle Power is low Censoring is tricky
Pull-up
Safety hypothesis: • E.g., Cox-2 leads to 2 fold increase in
MI etc.• Design: stop early if you reject• Estimated relative risk must be pulled
toward 2 Insight from Joe Heyse
The option
Respect PI: Adjudicate adverse events
Precision: Reclassify, reorganize Mechanism: Be an empiricist Dredging: Use sentinel events
top related