Crash modification factors for urban signalised intersections based on a new overall Sight Distance Index

Signalised intersections are critical points in urban road networks with high crash risks due to traffic density and complex manoeuvres. While traditional safety analyses are reactive, current research emphasises proactive methods. Tools such as Crash Modification Factors (CMFs) and Safety Performance Functions (SPFs) are essential for quantifying the impact of geometric and operational design on safety. However, CMFs tailored to specific geographical and traffic conditions remain limited. This study demonstrates that geometric factors significantly influence crash risk at urban signalised intersections. The developed Poisson regression model showed high predictive accuracy, and the derived CMFs and SPFs provided valid results for safety analyses under local conditions. The proposed ‘a’ index and the Overall Sight Distance Index (L), combined through a composite CMFₗ, serve as practical tools for quantifying geometric risk and prioritising intersection improvements. A key finding is the development of the ‘L’ index, a composite geometric parameter that integrates horizontal skew and vertical gradient to represent the overall sight distance condition at an intersection. The L index is incorporated into a composite crash modification factor (CMFₗ) by combining the individual CMFs for MHABIL and SNSL, enabling quantitative assessment of how combined geometric deficiencies affect crash frequency. Lower L values, reflecting more severe geometric constraints, are associated with higher CMFₗ values and thus greater crash risk, providing a practical tool for ranking intersections and prioritising safety interventions. In conclusion, the study’s results offer engineers and urban planners practical, data-driven tools for the proactive management of geometric risks, even under topographical and spatial constraints in urban environments. This approach supports safer intersection design and targeted safety interventions before crashes occur, moving beyond reactive measures.