On Non-parametric Bayesian Regression in Cardiovascular Disease Risk Assessment

• Activity Number: 479
• Type: Contributed
• Date/Time: Wednesday, August 3, 2011 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Olli Saarela*+ and Elja Arjas
• Companies: McGill University and University of Helsinki
• Keywords: disease prediction ; risk assessment ; non-parametric Bayesian regression ; continuous model expansion ; monotonic regression ; cardiovascular diseases

Abstract

Assessing the absolute risk for a future disease event for presently healthy individuals has an important role in the primary prevention of cardiovascular diseases (CVD). In addition to utilizing novel risk factor information on top of well established risk factors, the risk assessments can potentially be improved by using better, more realistic, statistical models. Here we concentrate on this latter task, utilizing non-parametric Bayesian regression techniques and continuous model expansion to relax parametric modeling assumptions where this is most beneficial. While allowing for non-linear associations all around, to achieve parsimony in the model fit, we allow for multidimensional relationships within specified subsets of risk factors, determined either on a priori basis or as a part of the estimation procedure. The computational tool utilized is our previously published Bayesian multivariate monotonic regression procedure, which we generalize here to a survival analysis setting. We apply the proposed methods to study whether 10-year CVD risk assessment in Finnish population based cohorts can be improved by a more careful modeling of the effects of classic risk factors of CVD.