|
Activity Number:
|
222
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 3, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #304545 |
|
Title:
|
Summarizing Response: When 'Effect Over Time' Doesn't Actually Capture an Intervention's Effect in Longitudinal Data
|
|
Author(s):
|
Rochelle E. Tractenberg and Kang-Hsien Fan*+ and Anne-Michelle Noone and Kathryn Sandberg
|
|
Companies:
|
Georgetown University Medical Center and Georgetown University and Georgetown University and Georgetown University Medical Center
|
|
Address:
|
Biostatistics, Bionformatics &Biomathematics, Washington, DC, 20057,
|
|
Keywords:
|
repeated measures ; area under curve ; response trajectory
|
|
Abstract:
|
In a mouse model of blood pressure (BP), animals are monitored for 3 days prior to infusion with a substance that affects BP (angiotensin-II), and for 8 days after. The responses to infusion vary over time, and 'mean response over time' is one of the least informative summaries. Three types of summaries capture different features of the response trajectories. Time based: Hours from infusion to peak response & hours to clinically-meaningful % change reflect 'responsiveness'; hours from peak response to 5% below peak reflect 'resilience.' Relative responses: % increase from baseline reflects the 'deviation response'; % response (peak-lowest/lowest) reflects the 'strength of response'; % increase from baseline to end reflects the 'overall change.' Area under the curve (AUC) is the most response-specific outcome, summarizing 'response' within a clinically-meaningfully defined period.
|
