|
Activity Number:
|
158
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Monday, August 3, 2009 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Health Policy Statistics
|
| Abstract - #304700 |
|
Title:
|
Surrogate Screening Models for Determining Low Physical Activity in the Cardiovascular Health Study
|
|
Author(s):
|
Sandrah P. Eckel*+ and Karen Bandeen-Roche and Paulo H. Chaves and Linda P. Fried and Thomas A. Louis
|
|
Companies:
|
Johns Hopkins Bloomberg School of Public Health and Johns Hopkins Bloomberg School of Public Health and Johns Hopkins Bloomberg School of Public Health and Columbia University and Johns Hopkins Bloomberg School of Public Health
|
|
Address:
|
615 N Wolfe Street, E3527, Baltimore, MD, 21218,
|
|
Keywords:
|
Boosting ; Gerontologic Frailty ; Logistic regression ; Prediction model evaluation ; Relative influence
|
|
Abstract:
|
Frailty is a geriatric syndrome of major public health importance. Low physical activity is a key item of the widely used frailty phenotype originally created in the Cardiovascular Health Study (CHS). In CHS, physical activity was assessed using the long Minnesota Leisure Time Physical Activity Questionnaire at only 3 of the first 10 annual visits. We develop and apply methods to identify and evaluate surrogate screening models for low physical activity with the dual goals of (1) streamlining screening and of (2) determining low physical activity when the questionnaire information is missing. We identify potential surrogate models using Generalized Boosted Models and logistic regression at baseline and evaluate the models at baseline and follow-up using complementary criteria. Results may inform future frailty screening tools.
|
