Date: Mon, 28 Sep 1998 15:46:46 -0400
Reply-To: HERMANS1 <HERMANS1@WESTAT.COM>
Sender: "SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU>
From: HERMANS1 <HERMANS1@WESTAT.COM>
Subject: Re[2]: How to choose closest date in longitudinal set?
Content-Type: text/plain; charset=US-ASCII
In this context the criterion "closest" as opposed to "latest prior
to" generally implies minimum absolute time distance from a point on a
time line. Whether that provides a good approximation to a last
observation prior to onset depends critically on what you are trying
to measure.
Content questions of this type actually do have a lot to do with
statistical programming. Many scientific studies require close
collaboration between content specialists and programmers.
A correct "closest date" solution may prove more complicated than a
"latest date prior to" solution. A content specialist has to
determine which to use. Sig
______________________________ Question _________________________________
Subject: Re: How to choose closest date in longitudinal set?
Author: JOSE FERNANDO DIAZ <jdiaz@welchlink.welch.jhu.edu> at Internet-E-Mail
Date: 9/28/98 3:14 PM
Wait, wait, wait. If you use the difference (fecha-fechax) you will get a
date before or after the onset. This may or may not have biological
significance. For example, the date of immunization should be earlier than
the date of antibody testing. Clinical trials are more or less
straightforward. Observational studies are more difficult. Is diarrhea a
cause or an effect of malnutrition???
Sorry this doesn't have to do with SAS.
Jose F. Diaz, MD
Division of Disease Control
Department of International Health
The Johns Hopkins School of Public Health
Phone (home) : 410 685 2155
