Using the Trend Odds Model in Seroepidemiological Studies

• Activity Number: 83
• Type: Contributed
• Date/Time: Sunday, August 1, 2010 : 4:00 PM to 5:50 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Ana W. Capuano*+ and Gregory C. Gray and Jeffrey D. Dawson and R. William Field
• Companies: The University of Iowa and University of Florida and The University of Iowa and University of Iowa
• Keywords: epidemiology ; proportional odds ; constrained models ; ordinal logistic regression

Abstract

Infectious diseases studies often use ordinal outcomes based on collecting data from antibody titers. Using standard ordinal logistic regression is generally more powerful than dichotomizing the outcome into high versus low levels. When proportional odds assumption is not met, it may be possible to capture the lack of proportionality using a constrained structural relationship between the odds and the cut-points (Peterson and Harrell, 1990). We consider a trend odds version of this constrained model, where the odds parameter increases or decreases in a monotonic manner across the titer cut-points. In this paper, we fit the trend odds model to data from swine influenza studies using a linear constraint, and present simulation results that illustrate its empirical properties. We show that this model can provide insight to departures from the proportional odds assumption.