From changym@email.uc.edu Thu Jul  8 14:30:32 1999
Received: from mxu2.u.washington.edu (mxu2.u.washington.edu [140.142.32.9])
	by lists.u.washington.edu (8.9.3+UW99.02/8.9.3+UW99.01) with ESMTP id OAA26340
	for <waphgis@lists.u.washington.edu>; Thu, 8 Jul 1999 14:30:31 -0700
From: changym@email.uc.edu
Received: from newman.bch.uc.edu (newman.bch.uc.edu [129.137.33.152])
	by mxu2.u.washington.edu (8.9.3+UW99.02/8.9.3+UW99.06) with ESMTP id OAA15877
	for <waphgis@u.washington.edu>; Thu, 8 Jul 1999 14:30:31 -0700
Received: from ihphsr17 (node124073.msb.uc.edu [129.137.124.73])
	by newman.bch.uc.edu (8.9.2/8.9.2) with SMTP id QAA10349
	for <waphgis@u.washington.edu>; Thu, 8 Jul 1999 16:19:39 -0400 (EDT)
Message-Id: <4.1.19990708173030.00a0a710@email.uc.edu>
X-Sender: changym@email.uc.edu
X-Mailer: QUALCOMM Windows Eudora Pro Version 4.1 
Date: Thu, 08 Jul 1999 17:32:28 -0400
To: waphgis@u.washington.edu
Subject: multiple entries per person
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"

Dear listers,

My research has encountered a challenging issue - “multiple entries per
person”. 

On the health data I am working on, considerable proportions of patients had
requested health services more than one time for the same disease/ symptom
(verified by ICD9).  Many of them have moved from location A to location B, and
possible C.   

Since geographic features are important parameters to be tested and later I
would need to geocode* them using GIS, it becomes the issue.  Although randomly
choose one visit per patient can be technically done through SAS programming
for patients without (address) moving, I am struggling with the best strategy
for the patients with multiple visits and relocations.  What would be the more
acceptable methods to deal with this situation without bringing in skewed
weighting (or missing the point) for modeling later?   Or I should keep every
entry for each patient when model is being tested.  

Your input would be greatly appreciated.  Thanks in advance.

Yu-mei Chang




.