Multiplicative effects and frailty intercepts on survival for fertile and subfertile men: a piecewise constant exponential survival model for sperm count data

Ronny Westerman, University of Marburg
Katharina Belting, University of Marburg
Hanna Seydel, University of Marburg
Walter Krause, University of Marburg
Ulrich O. Mueller, University of Marburg

Examing the effects of interaction simutaneously for group- and individual-level information and their impact on the outcome variable have gained more relevance for health data analysis. Such multi-level procedures can be used to understand the substantial connections between disease and risk factors by separating the "real" independent effect from the defective association caused by different qualitative levels of variation in data. Age-specific effects and within-group correlation in cohorts can be maintain as the major reason for disparity in survival for subfertile and fertile men and also be better predictors for population hazards. The quality of sperm count and motility is also age-specific decreasing with matured age could also correspond with higher risk of mortality more likely occuring in some geographical units. We use the piecewise exponential survival model to consider reciprocal terms of interaction for individual- and group-level variables. Applying the frailty-term for exploring the between-group unobserved heterogeneity will be specified with the gamma-distribution.

  See extended abstract

Presented in Poster Session 2