|
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 - #303905 |
|
Title:
|
Sensitivity Analyses for Omitted Variable Bias in Multiple Regression in a Study of Right Heart Catheterization
|
|
Author(s):
|
Carrie Hosman*+ and Ben B. Hansen
|
|
Companies:
|
University of Michigan and University of Michigan
|
|
Address:
|
439 West Hall, Statistics Department, Ann Arbor, MI, 48109,
|
|
Keywords:
|
sensitivity analysis ; omitted variable bias ; observational studies
|
|
Abstract:
|
Omitted variable bias can affect treatment effect estimates obtained from observational data due to the lack of random assignment to treatment groups. Sensitivity analyses quantify the impact of potential omitted variables. This paper presents methods of sensitivity analysis to adjust interval estimates of treatment effect---both the point estimate and standard error---obtained from linear regression. This adjustment is of importance for inferences based upon the regression; the narrowing or widening of a confidence interval can alter inferences made about treatment in a different manner than a shift in the point estimate. Methods are demonstrated on data from the Connors et al. (1996) study of right heart catheterization. Consideration of various potential omitted variables shows the impact of omitted variables on interval estimates.
|
