R help - Sep 2007 - Repeated tests against baseline

I came across a post by Karl Knoblick regarding the modeling of longitudinal
data (see https://stat.ethz.ch/pipermail/r-help/2007-May/132137.html).  I am
often asked by physicians to perform what Karl refers to in his post as option
1: to perform paired t-tests against baseline at each follow up time point (30
days, 90 days, 6 months, etc.).  Unlike Karl's example, however, many of the
trials I am involved in are one-armed (so there are no across-trial-arms
comparisons).

No matter how hard I try to explain to physicians why this approach is not the
best, it has typically been to no avail.  I am wondering if anyone knows of a
paper I can quote instead?  One (or more) from the cardiovascular literature
would be especially precious to me.

Best regards,
   -Cody Hamilton



	[[alternative HTML version deleted]]

Here are a couple to look at, they may be helpful and the references in
them may give you a specific example that you can use (read them through
yourself, then decide if you want your docs to read them).

Kenneth F Schulz and David A Grimes (2005), "Multiplicity in randomised
trials I: endpoints and treatments", The Lancet, 365: 1591-95

Kenneth F Schulz and David A Grimes (2005), "Multiplicity in randomised
trials II: subgroup and interim analyses", The Lancet, 365: 1657-61


Hope this helps,


-- 
Gregory (Greg) L. Snow Ph.D.
Statistical Data Center
Intermountain Healthcare
greg.snow at intermountainmail.org
(801) 408-8111
 
 
> -----Original Message-----
> From: r-help-bounces at r-project.org 
> [mailto:r-help-bounces at r-project.org] On Behalf Of Cody Hamilton
> Sent: Wednesday, September 26, 2007 3:15 PM
> To: r-help at stat.math.ethz.ch
> Subject: [R] Repeated tests against baseline
> 
> I came across a post by Karl Knoblick regarding the modeling 
> of longitudinal data (see 
> https://stat.ethz.ch/pipermail/r-help/2007-May/132137.html).  
> I am often asked by physicians to perform what Karl refers to 
> in his post as option 1: to perform paired t-tests against 
> baseline at each follow up time point (30 days, 90 days, 6 
> months, etc.).  Unlike Karl's example, however, many of the 
> trials I am involved in are one-armed (so there are no 
> across-trial-arms comparisons).
> 
> No matter how hard I try to explain to physicians why this 
> approach is not the best, it has typically been to no avail.  
> I am wondering if anyone knows of a paper I can quote 
> instead?  One (or more) from the cardiovascular literature 
> would be especially precious to me.
> 
> Best regards,
>    -Cody Hamilton
> 
> 
> 
> 	[[alternative HTML version deleted]]
> 
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide 
> http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>