Analysis of Observational Health Care Data Using SAS

Chapter 4

Doubly Robust Estimation of Treatment Effects

Michele Jonsson FunkDaniel WestreichChris WeisenMarie Davidian

Abstract

4.1 Introduction

4.2 Implementation with the DR Macro

4.3 Sample analysis

4.4 Summary

4.5 Conclusion

References

 

Abstract

Estimation of the effect of a treatment or exposure with a causal interpretation from studies where exposure is not randomized may be biased if confounding is not taken into appropriate account. Adjustment for confounding is often carried out through regression modeling of the relationships among treatment, confounders, and outcome. Doubly robust (DR) estimation produces a consistent effect estimator as long as one of two component regression models is correctly specified and assuming that there ...