Date: Sun, 5 Jan 2003 13:07:06 -0800
Reply-To: Sam <archana_gs@MAILCITY.COM>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Sam <archana_gs@MAILCITY.COM>
Subject: Re: Sas data validation
Content-Type: text/plain; charset=ISO-8859-1
Thanks replying my question.I known sas programming ,but not in pharma
area .I know all the terms,I need what other responsibility for
reviewing the CRF's .Can u Please tell me what proc statements should
we use for data validation.I thought data validation is to check for
data consistency ,check for duplicaate records ,unique key var. I read
an artical that IntrNet can used for Edit Check.Can u please explain
about edit check.
"Roland" <email@example.com> wrote in message news:<firstname.lastname@example.org>...
> "Sam" <email@example.com> wrote in message
> > Hi,
> > I need known what is sas Data validation? According to my
> > knowledgeit is used to detect errors while entering the data.It can be
> > done thro sas datasets and Proc sql.what is edit check for data
> > validation .My another quetion is about CRF's, what is reviewing
> > protocols ,CRF and statistical plans and Annotate CRF.I would really
> > appreciate if any one answer my questions.
> It is to do with Clinical trials. Data validation will not be just checking
> data on entry but will mean cross-checks and feasibility checks as well,
> most of which will be done post-data entry. Proc sql probably won't come
> into it. Reviewing protocols means reading and perhaps criticising the
> set-up of a clinical trial. The protocol are the processes written down that
> everybody agrees to follow before the clinical trial starts. CRF's are case
> report forms. That is the document (either paper or electronic) in which
> data for the subjects is entered. You have to make sure this will clearly
> capture the information you need and is understandable and clear.
> Statistical plans are the proposed method for analysing the data that has to
> be agreed before a clinical trial starts. An annnotated CRF is a blank copy
> of the CRF with the field names written again the entry boxes and other
> identifying information so that the programmers can match the data back to
> the box that was filled in.
> If you are not a statistician then don't bother applying for a job involving