Modern studies increasingly enjoy the ability to examine a large number of exposures in a comprehensive manner. However, several risk factors often tend to be related in a non-trivial way, undermining efforts to identify the individual ones using standard analytic methods due to inflated type I errors and possible masking of effects. Epidemiologists often use data reduction techniques by grouping the prognostic factors using a thematic approach, with themes deriving from biological considerations. However, it is important to account for potential mis-specification of the themes to avoid false positive findings. To this end, we propose shrinkage type estimators based on Bayesian penalization methods, in the context of analyzing time-to-event data. The properties of the estimators are examined using simulations. The methodology is illustrated using the Agricultural Health Study cohort to examine the role of multiple types of pesticides on the etiology of breast cancer in farm women.