Conferencias
Dipankar Bandyopadhyay
University of Minnesota, USA
Dental studies often produce spatially-referenced time-to-event data, such as the time until tooth loss due to periodontal disease. These data are used to identify risk factors associated with tooth loss and to predict the outcomes for an individual patient. In this talk, we assume a proportional hazard model and account for dependence between nearby teeth using spatial frailties, which are modeled as linear combinations of positive stable random effects. This model permits predictions conditioned on spatial random effects that account for the survival status of nearby teeth, and simultaneously preserves the proportional hazards relationship marginally over the random effects allowing for interpretable estimates of the effects of risk factors on tooth loss. We apply this model to a dataset obtained from a private dental practice to illustrate how this model can be used to identify important risk factors for tooth loss and predict the remaining lifespan of a patient’s teeth. This is joint work with Drs. Brian J. Reich and Martha Nunn
