Power methods in adaptive randomization trials

[Thompson,Paul]

I have been looking for some thoughtful discussions of power methods in adaptive randomization trials. For instance, the I-Spy2 trial (Barker et al, 2009, Clin Pharm Ther, 86, 97-100) presents a large trial with nary a word on power issues.

 

I have been troubled for some time about these trials. Any discussions of power in adaptive trials, personalized medicine, etc would be helpful.

 

Paul A. Thompson, Ph.D.

----------------------------------------------------------------------- Confidentiality Notice: This e-mail message, including any attachments, is for the sole use of the intended recipient(s) and may contain privileged and confidential information. Any unauthorized review, use, disclosure or distribution is prohibited. If you are not the intended recipient, please contact the sender by reply e-mail and destroy all copies of the original message.

[SteveDrD]

www.fda.gov/downloads/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/UCM201790.pdf

 

The above link has the FDA's latest (2010) thoughts on adaptive design, including several issues with power.

 

Hope this helps.

 

Steve Denham

[Thompson,Paul]

Thanks – very pertinent.

--
--
To post a new thread to MedStats, send email to MedS...@googlegroups.com .
MedStats' home page is http://groups.google.com/group/MedStats .
Rules: http://groups.google.com/group/MedStats/web/medstats-rules
 
---
You received this message because you are subscribed to the Google Groups "MedStats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to medstats+u...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

[Steve Simon, P.Mean Consulting]

On 3/30/2013 3:07 PM, Thompson,Paul wrote:

> I have been looking for some thoughtful discussions of power methods
> in adaptive randomization trials. For instance, the I-Spy2 trial
> (Barker et al, 2009, Clin Pharm Ther, 86, 97-100) presents a large
> trial with nary a word on power issues.
>
> I have been troubled for some time about these trials. Any
> discussions of power in adaptive trials, personalized medicine, etc
> would be helpful.

For adaptive trials, the only way to assess power is through extensive
simulations. These simulations have to envision a wide range of
scenarios. For a dose finding adaptive trial, you have simulations that
assure that there is a high probability of selecting the correct dose
when the correct dose is the lowest dose, one of the middle doses, and
the highest dose. It has to look at the behavior of the design when two
doses are equally good and when all doses are equally good. If the
measure of whether the dose is "correct" depends on a efficacy/safety
trade-off, then you have to look at a range of different choices for the
shape of the efficacy curve and the safety curve.

I'm not surprised if an adaptive trial does not discuss power. It is not
a simple number anymore if you are using an adaptive trial. It can't be
explained in a single sentence or even a single paragraph. It might make
sense to publish the simulation results that establish the power of a
study in a separate publication or as a technical appendix.

Steve Simon, n...@pmean.com, Standard Disclaimer.
Sign up for the Monthly Mean, the newsletter that
dares to call itself average at www.pmean.com/news

[Belinda Dawson]

Hello everyone,
 
I am using "proc forecast" to do some projection. I have a dataset about percent(0~100%), but when I run "proc forecast" on it, it gave me some results above 100%. I donot know how to set a limit on the result,  can somebody help me?
 
Thanks,
 

[Munyaradzi Dimairo]

Hi Paul

That is a very good question indeed. Sample size is normally the maximum expected done through simulation to understand the operating characteristics of the study such as type I and II error, impact of the allocation probability, etc. Unfortunately, the procedure is not well described on the proposal maybe it's due to space limitation. However, there is no excuse whatsoever not to give the details even in brief. I suspect they used bayesian predictive power approach in this case (http://cctst.uc.edu/sites/default/files/files/Slides%20Berry%202011-06-17.pdf)

My main concern is on reproducibility of the approach especially with little information provided. It's a challenge for reviewers of such grants.

cheers

Munya

--

--
To post a new thread to MedStats, send email to MedS...@googlegroups.com .
MedStats' home page is http://groups.google.com/group/MedStats .
Rules: http://groups.google.com/group/MedStats/web/medstats-rules
 
---
You received this message because you are subscribed to the Google Groups "MedStats" group.
To unsubscribe from this group and stop receiving emails from it, send an email to medstats+u...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

--

"Learn from yesterday, live for today, hope for tomorrow. The important thing is not to stop questioning" ~ Albert Einstein

***********************************
Munyaradzi Dimairo
NIHR Research Fellow in Medical Statistics
University of Sheffield
School of Health and Related Research (ScHARR)
Clinical Trials Research Unit
Regent Court, 30 Regent Street
Sheffield
S1 4DA

Physical address:
Room 3.10, Innovation Centre

email: m.di...@sheffield.ac.uk 
          mdim...@gmail.com
Twitter: @mdimairo [views are my own, retweeting is not endorsement]
 fax: +44 (0) 114 222 0870
http://www.sheffield.ac.uk/scharr/sections/dts/ctru/  
***********************************