268522 Estimating the effects of two classes of drugs on hemoglobin with a doubly robust method

Tuesday, October 30, 2012 : 8:45 AM - 9:00 AM

Brian P. Johnson, MPH , Division of Research, Essentia Institute of Rural Health, Duluth, MN
Charles E. Gessert, MD, MPH , Division of Research, Essentia Institute of Rural Health, Duluth, MN
Colleen M. Renier, BS , Division of Research, Essentia Institute of Rural Health, Duluth, MN
Adnan Ajmal, MBBS , Department of Hospital Medicine, University of Massachusetts, Leominster, MA
With observational data, the effect of the exposure is often confounded by baseline covariates. Doubly robust estimation utilizes models for the exposure and for the outcome to estimate the causal effect of an exposure on an outcome. In a retrospective study designed to evaluate change in hemoglobin (Hgb) with use of angiotensin-converting enzyme inhibitors (ACEI) or angiotensin receptor blockers (ARB), in a primary care patient population, we found that receiving ARB vs. ACEI was associated with several of the baseline covariates that were chosen a priori and thought to be informative in estimating follow-up (F/U) Hgb. We therefore adopted the doubly robust semiparametric efficient estimator of Robins et al. (1994) in a causal analysis of the treatment effect of ARB vs. ACEI on F/U Hgb. The method produces an estimate of the treatment effect by simultaneously incorporating the propensity of a subject to receive ACEI or ARB, given their levels of covariates, and the effects of the covariates upon the response of interest, F/U Hgb. It is doubly robust in the sense that it produces an unbiased estimate of the treatment effect if either the outcome or propensity model is correct. Following a complete-case ANCOVA, we found the estimated F/U Hgb and bootstrap bias-corrected accelerated (BCa) 95% confidence interval (CI) of ACEI and ARB to be 14.31 (14.21, 14.42) gm/dL and 14.48 (14.33, 14.62) gm/dL, respectively. The causal effect of ARB relative to ACEI and associated BCa CI is estimated to be 0.17 (0.00, 0.31) gm/dL (p = 0.0310).

Learning Areas:
Biostatistics, economics
Chronic disease management and prevention
Clinical medicine applied in public health

Learning Objectives:
Demonstrate the use of an estimation approach when the effect of an exposure on an outcome is confounded by covariates as is often the case with observational data.

Presenting author's disclosure statement:

Qualified on the content I am responsible for because: I am a professional Biostatistician, currently working at a non-profit research institute, having previously worked in the clinical trials divisions of two medical device companies for over nine years. I have an MPH in Biostatistics from the University of Minnesota, earned in 2000.
Any relevant financial relationships? No

I agree to comply with the American Public Health Association Conflict of Interest and Commercial Support Guidelines, and to disclose to the participants any off-label or experimental uses of a commercial product or service discussed in my presentation.