|
Activity Number:
|
540
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 2, 2007 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
ENAR
|
| Abstract - #309712 |
|
Title:
|
A Locally Weighted Regression Approach to Evaluating the Effects of Smoking Reduction on Birth Weight
|
|
Author(s):
|
Jeff M. Szychowski*+ and J. Michael Hardin and Michael D. Conerly and Lesa L. Woody and Wendy S. Horn
|
|
Companies:
|
The University of Alabama at Birmingham and The University of Alabama and The University of Alabama and The University of Alabama at Birmingham and The University of Alabama at Birmingham
|
|
Address:
|
RPHB 327, Birmingham, AL, 35294-0022,
|
|
Keywords:
|
smoothing ; variable selection ; smoking cessation
|
|
Abstract:
|
While smoking cessation is known to improve neonatal outcomes in pregnant smokers, reducing tobacco exposure has been more controversial. Windsor et al. (1999) showed that reducing tobacco exposure by 50% or more during pregnancy leads to beneficial effects on birth weight in neonatal outcomes. However, others have questioned the benefit of reduction in the absence of cessation. Some have argued that these conflicting reports are due to a nonlinear relationship between tobacco exposure and birth weight, (e.g., England, et al. (2001)). We employ a locally weighted regression (LWR) approach to evaluate the relationship between tobacco exposure and birth weight using data from the Smoking Cessation or Reduction in Pregnancy Trial (SCRIPT). By using an improved Cp statistic based on LWR, we verify this nonlinear relationship and identify a parsimonious predictive model for birth weight.
|
