Framework for integration of regional transportation model (RTM) with collision prediction model (CPM)

Traditional road safety planning is largely reactive, while existing Collision Prediction Models (CPMs) often function as disconnected post-processing tools. However, a growing body of research has highlighted the need for proactive safety assessment at the transportation planning stage. Building on this research, the present study demonstrates a large-scale, automated framework that integrates CPMs with a calibrated, tour-based Regional Transportation Model (RTM), enabling travel demand and network-wide collision outcomes to be evaluated within a unified modelling workflow. Using the Calgary metropolitan region as a case study, Negative Binomial models were developed to estimate Fatal/Injury (FI) and Property-Damage-Only (PDO) collisions at mid-block segments and intersections. To ensure seamless operational integration, all explanatory variables used by the CPMs were derived exclusively from the RTM's multimodal outputs, thereby eliminating the need for additional external data collection. Validation against independent historical data from 2017 to 2019 demonstrated robust predictive accuracy at both the aggregate network level and the individual facility level. The framework was then applied to a 2028 planning horizon to evaluate twelve demand-side policy scenarios involving changes in fuel prices, parking costs, and transit fares. The results revealed a consistent, monotonic dose-response relationship across all interventions. Increases in fuel prices produced the largest network-wide reductions in collision, while lower transit fare and higher parking costs generated more targeted safety benefits. Intersections were particularly sensitive, especially with respect to FI collisions under the transit fare scenarios. By generating severity-stratified safety estimates alongside conventional demand outputs, without requiring additional data inputs or manual post-processing, the framework enables planners to evaluate mobility and safety trade-offs for proactive, data-informed long-range transportation planning.