JSM (the Joint Statistical Meetings) is the largest gathering of statisticians held in North America. It is held jointly with the American Statistical Association, the International Biometric Society (ENAR and WNAR), the Institute of Mathematical Statistics, and the Statistical Society of Canada.
Session: 141 Monday, August 7, 2006 : 10:30 AM to 12:20 PM Sponsor: Biometrics Section
Title: Use of a Marginal Structural Model To Estimate the Causal Effect of Epoetin Use on Hemactocrit Value among Hemodialysis Patients
Author(s): Yi Zhang, Mae Thamer, Dennis J. Cotter, James Kaufman and Miguel A. Hern Medical Technology and Practice Patterns Institute, U.S. Department of Veterans Affairs and Harvard School of Public Health
Address: 4733 Bethesda Ave., #510, Bethesda, MD, 20814,
Keywords: causal inference ; marginal structural model ; inverse probability of treatment weighted estimation; erythropoietin; hematocrit
Abstract: Standard methods to adjust for time-varying confounders that are affected by previous treatment may lead to biased estimates. Marginal structural models (MSMs), introduced by Robins, can appropriately control for this type of time-dependent confounding. The parameters of MSMs are estimated by inverse probability weighting. We present an application of MSMs to estimate the causal effect of erythropoietin (EPO) on hematocrit (Hct) response among renal failure patients with anemia. The research question is motivated by the discrepancies between observational and clinical trial results examining EPO treatment to ameliorate anemia. Specifically, in contrast to clinical trials, observational studies have reported a significantly inverse association between EPO and Hct.