Statistical models are aimed at statistical analyses and modeling of PC data. To develop and implement statistical models, we will utilize all of the available information from the Pancreatic Cancer Collaborative Registry, as well as data from relevant peer-reviewed studies. The information generated by our statistical modeling efforts will be used to build several Bayesian networks (BN) for PC Risk Factors, Diagnostics, Treatments, and Outcomes.

The choice of a concrete statistical model will depend on the research question of interest. Currently, given the lack of effective treatment for PC, the issues of the highest importance in treating PC include (1) estimation of the probability of developing PC within a certain time period for a high risk individual, and (2) prediction of the survival time for a patient with certain characteristics after being diagnosed with PC.