[Skip to Content]
Access to paid content on this site is currently suspended due to excessive activity being detected from your IP address 23.23.47.118. Please contact the publisher to request reinstatement.
[Skip to Content Landing]
Citations 0
Letters
May 16, 2007

Analytic Approaches to Observational Studies With Treatment Selection Bias

Author Affiliations
 

Letters Section Editor: Robert M. Golub, MD, Senior Editor.

JAMA. 2007;297(19):2077-2078. doi:10.1001/jama.297.19.2077-a

To the Editor: We have some concerns about the study by Dr Stukel and colleagues1 comparing propensity score and instrumental variable methods for removing the effects of selection bias in observational studies. The authors concluded that all standard risk-adjustment methods have the same limitations regarding removal of unmeasured treatment selection bias, and that compared with standard modeling, instrumental variable analysis may produce less biased estimates of treatment effect. The accompanying Editorial2 questioned why the instrumental variable “regional catheterization rate” was not included in the list of variables defining the propensity score, and the effect of its potential inclusion. We would like to add that not including an instrumental variable in the propensity score analysis also violates the main assumption of the propensity score method (ie, that all information on the factors that can affect the choice of the treatment is included). This study only shows that adding an additional important variable to the analysis may substantially improve the results.

First Page Preview View Large
First page PDF preview
First page PDF preview
×