Doubly Robust Estimation of Causal Effects

Jonsson Funk, et al. have developed a SAS® macro for doubly robust estimation of the effect of a dichotomous treatment (or exposure) on a continuous or dichotomous outcome with bootstrapped standard errors.  Base SAS v8 or later is required.

DR Macro (v1.1) – Latest version – release date: May 25 2010
New: bootstrapped standard errors and confidence intervals
Related materials: simulated data, sample SAS code

Related Publications & Presentations

Doubly Robust Estimation of Treatment Effects by M Jonsson Funk, D Westreich, M Davidian & C Wiesen.
Originally published in Analysis of Observational Health Care Data using SAS. D Faries, A Leon, J Haro, and R Obenchain (eds.), SAS Press, Jan 2010.
Related materials: simulated study data, sample SAS code

Introducing a SAS® Macro for Doubly Robust Estimation
Jonsson Funk M, Westreich D, Davidian M, & Wiesen C. SAS Global Forum, April 2007, Orlando, Florida, USA.

Introducing Doubly Robust Estimation of Treatment Effects
Jonsson Funk M. 23^rd ICPE Meeting, August 19-22, 2007, Quebec City, Canada.

Contributors
Michele Jonsson Funk, PhD, Project Principal Investigator
Daniel Westreich, PhD, Collaborator
Marie Davidian, PhD, Collaborator
Chris Wiesen, PhD, Collaborator

Financial Support
Funding for the development of the macro was initially provided by the Agency for Healthcare Research & Quality (AHRQ) through a supplemental grant to the UNC Centers for Education & Research on Therapeutics (CERTs) (3U18HS010397-07S1). Ongoing work is supported by AHRQ through a career development award to Dr Jonsson Funk (K02HS017950).

Acknowledgements
We would like to acknowledge the valuable contributions of Harry Guess, MD, PhD (deceased) to this project. We would also like to thank our colleagues at the Duke Clinical Research Institute and students in the Epidemiology Department at the University of North Carolina who provided feedback on early versions of the macro.

Questions?
If you would like to discuss doubly robust estimation and/or the macro with others, we invite you to join the DR macro listserv by sending an email message to listserv@unc.edu with the following subject line: 'subscribe drmacro <your first  name> <your last name>'. To post a message to the group after you have joined, email drmacro@listserv.unc.edu. The listserv is moderated to prevent others from using it to send spam.

If you have questions about the macro, you may also contact Dr. Jonsson Funk directly.

This page last updated on 27 May 2010.