|
Activity Number:
|
158
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 7, 2006 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #306570 |
|
Title:
|
Novel Methods in the Visualization of Transitional Phenomena
|
|
Author(s):
|
Bruce Swihart*+ and Brian Caffo and Matthew Strand and Naresh Punjabi
|
|
Companies:
|
University of Colorado at Denver and Health Sciences Center and Johns Hopkins University and University of Colorado at Denver and Health Sciences Center and Johns Hopkins University
|
|
Address:
|
622 Panorama Drive, Grand Junction, CO, 81503,
|
|
Keywords:
|
history matrix ; history matrix visualization ; longitudinal categorical data ; Markov process ; multi-state survival analysis ; sleep apnea
|
|
Abstract:
|
When studying a group of individuals over time there exists an inclination to explore the data and report results with composite summary measures that collapse the time component. An example of this would be "Those with sleep apnea spent 18% of their total sleep time in REM, whereas those without sleep apnea spent 21%." Although these composite summary measures are convenient, they do obscure any time-related effects. History Matrix Visualization (HMV) is a technique that can allow the exploration and guide the analysis of longitudinal categorical data, and may provide insights into transitional processes that may have been otherwise obscured when implementing traditional composite summary measures. The HMV method is demonstrated using data of 59 matched apneic and non-apneic individuals from the Sleep Heart Health Study.
|
