Since the inception of Bridge Management Systems (BMS), transportation departments have used optimization and planning algorithms to anticipate near-term and long-term resources needed for maintenance, repair, and rehabilitation of highway structures. An integral component of the optimization framework is the family of the deterioration models used to forecast the condition of critical components of highway bridges over the duration of the planning cycle. While the earliest approaches used for developing these deterioration models relied on relatively simple statistical methods, as well as practitioner heuristics, to forecast the condition states, a sufficient database of historical condition ratings now exists in the United States to permit revisiting the strategies for deterioration modeling with novel techniques for big data. In the current study, proportional hazards regression has been performed on a statewide bridge database consisting of 35 years of historical inspection ratings to develop probabilistic deterioration models that account for the effects of significant factors, including design, functional, and geographic features, on the deterioration rate. A novel aspect of the developed framework includes the ability to analyze the time-dependent effects of explanatory factors on deterioration rates over the life-cycle of the structural components. The development and implementation of an automated software framework for this regression analysis will be presented and typical results obtained from application to state inventory consisting of over 17,000 bridges will be discussed. The results provide interesting insight on the extent that design, functional, and geographic factors influence deterioration rates of different bridge components and may have applications beyond deterioration modeling, including quantifying the value of preventative design measures and preservation strategies.