Extension of the Peters-Belson Method for Estimating Health Disparities to Logistic Regression with Multiple Groups and Outcomes from Survey Data

• Activity Number: 43
• Type: Contributed
• Date/Time: Sunday, August 3, 2014 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Barry Graubard*+ and Yan Li and Pengyu Huang and Joseph Gastwirth
• Companies: National Cancer Institute and University of Maryland and University of Maryland and George Washington University
• Keywords: Health disparity ; proportional odds logistic regression ; multinomial logistic regression ; Survey data

Abstract

Determination of the extent of disparity between groups of people, e.g., race or gender, is of interest in many fields including public health for medical treatment and prevention of disease. An observed difference in the mean outcome between an advantaged group (AG) and disadvantaged groups (DG) indicates a potential disparity, which can be due to differences in the distribution of exogenous covariates. The Peters-Belson (PB) method fits a regression model with exogenous covariates to the AG to predict for each DG member their outcome measure as if they had been from the AG. The difference between the mean predicted minus mean observed outcomes of DG members is the (unexplained) disparity of interest. We focus on applying the PB method to inferences for disparity based on binary/multinomial and proportional odds logistic regression models estimated from complex surveys with more than one DG. Simulation results and analyses of disparity in body mass index and chronic condition data for hypertension and diabetes among racial groups in the National Health and Nutrition Examination Survey 1999-2004 is presented.