[R] Use of R in clinical trials

[Peter Dalgaard]

Bill.Venables at csiro.au wrote:
> I can't believe I'm saying this, but I think Peter is being a bit harsh on SAS.  
> 
> I prefer Greg Snow's analogy (in the fortune collection): If SPSS (or
> 
SAS) and R were vehicles, SPSS would be the bus, going on fixed routes
and efficiently carrying lots of people to standard places, whereas R is
the off-road 4WD SUV, complete with all sorts of kit including walking
boots, kayak on the top, &c. R will take you anywhere you want to go,
but it might take you longer to master it than the simple recipes for
data analysis typical of the 'bus' programs.


I wasn't really trying to be harsh. I think my friend actually said
something like "endearing mediocrity", and particularly the recruiting
aspect is not something to take lightly if you have a business to run.

Incidentally SAS is not quite as bus-like as SPSS - it does actually
allow you to take the driver's seat as long as you keep to the road. It
would be more like the pickup truck or delivery van: moves best in a
straight line, but will take you between most places you need to go.

I have

Stata as the mini SUV city car (like SAS but quicker round the corners)

Genstat as the Land Rover (a bit like R designed in the 60's, but still
running)

The F1 car would probably be Ox or Gauss or similar programs designed to
do one kind of thing very efficiently.

-pd


> 
> Bill Venables
> CSIRO/CMIS Cleveland Laboratories
> 
> 
> -----Original Message-----
> From: r-help-bounces at r-project.org [mailto:r-help-bounces at r-project.org] On Behalf Of Peter Dalgaard
> Sent: Thursday, 18 February 2010 5:55 PM
> To: Frank E Harrell Jr
> Cc: r-help at r-project.org; Cody Hamilton
> Subject: Re: [R] Use of R in clinical trials
> 
> Frank E Harrell Jr wrote:
>> Cody,
>>
>> How amazing that SAS is still used to produce reports that reviewers 
>> hate and that requires tedious low-level programming.  R + LaTeX has it 
>> all over that approach IMHO.  We have used that combination very 
>> successfully for several data and safety monitoring reporting tasks for 
>> clinical trials for the pharmaceutical industry.
>>
>> Frank
> 
> There is a point to it, though. One of my friends and colleagues in the 
> business put it in one word: Mediocrity.
> 
> SAS does a mediocre job at analysing and reporting and data handling 
> using a mediocre control language. But: It can be handled by mediocre 
> programmers writing and modifying mediocre programs, and those people 
> are more available and replaceable, maybe even cheaper. R/LaTeX may run 
> circles around SAS in terms of capapilities, flexibility, and elegance, 
> but it can also send a programmer who doesn't have the required skill 
> set running around in circles.
> 
> -pd
> 
>> Cody Hamilton wrote:
>>> Dear all,
>>>
>>> There have been a variety of discussions on the R list regarding the 
>>> use of R in clinical trials. The following post from the STATA list 
>>> provides an interesting opinion regarding why SAS remains so popular 
>>> in this arena: 
>>> http://www.stata.com/statalist/archive/2008-01/msg00098.html
>>>
>>> Regards,
>>>
>>> -Cody Hamilton
> 
> 


-- 
   O__  ---- Peter Dalgaard             Øster Farimagsgade 5, Entr.B
  c/ /'_ --- Dept. of Biostatistics     PO Box 2099, 1014 Cph. K
 (*) \(*) -- University of Copenhagen   Denmark      Ph:  (+45) 35327918
~~~~~~~~~~ - (p.dalgaard at biostat.ku.dk)              FAX: (+45) 35327907