Date:         Wed, 15 Oct 2008 12:33:28 -0400
Reply-To:     Peter Flom <peterflomconsulting@mindspring.com>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Peter Flom <peterflomconsulting@MINDSPRING.COM>
Subject:      Re: Clinical programming without programmers?
Comments: To: RolandRB <rolandberry@HOTMAIL.COM>
Content-Type: text/plain; charset=UTF-8

RolandRB <rolandberry@HOTMAIL.COM> wrote>And another thing. When the statistician is preparing the table

>templates document then what if there were a tool to help them that
>not only allowed them to select a suitable template and allow them to
>add the table number, titles and footnotes but also to say what
>variable to select on and how and through that generate the code to
>produce the output? It's not so far fetched to write such a push-
>button reporting system.

I don't know much about clinical work, but it seems to me that there's a division of work:
1) Stuff only a programmer could do
2) Stuff only a statistician could do
3) Stuff anyone who can use a computer could do.

This is certainly true in much of the work I do in other areas.

The little interaction I have had with statisticians doing clinical trials (and correct me if I am wrong, this is hearsay) is that you learn everything you need to know in 6 months, then boredom until you quit.  The impression I get is that FDA requirements are so strict that there is little if any room for a statistician to use his or her brains.

OTOH, I do like the saying that "there are no routine statistical questions, only questionable statistical routines".  I imagine that something like that is also true for programming.  Certainly when I ask a programming question here on SAS-L, I usually get multiple replies, often quite different from each other.  Yet the impression one gets from 'outside' programming is that there is a WAY to do it RIGHT and the other ways are WRONG.  From my clients, I can tell that they get the same impression about statistics.

I am leery of automated "push button" reporting systems for the same reason that I am leery of menu-driven or gui driven statistics; they encourage the attitude that "the computer said it, I believe it, that settles it"



Peter

Peter L. Flom, PhD
Statistical Consultant
www DOT peterflom DOT com