Multivariate Cumulative Incidence Models for Twin Data

• Activity Number: 304
• Type: Contributed
• Date/Time: Tuesday, August 6, 2013 : 8:30 AM to 10:20 AM
• Sponsor: Biometrics Section
• Author(s): Klaus Holst*+ and Klaus Holst
• Companies: University of Copenhagen and University of Copenhagen
• Keywords: twin studies ; cumulative incidence probability ; multivariate competing risks ; threshold model ; concordance ; random effects model

Abstract

There are considerable interests in deriving to what extent specific cancer occurrences are inherited. This is typically achieved via family and twin studies by comparing the degree of association for different pairs that share different amount of genes. However, failing to take right censoring in such studies into account can lead to substantial bias. Standard modeling approaches such as the Liability Threshold model does not readily account for this.

We present new parametric and non-parametric methods for analyzing twin data based on modeling of the bivariate cumulative incidence for cancer, i.e., the concordance probability that both twins experience cancer before some time t with death as a competing risk.

The methods are demonstrated on data from the Danish twin registry.