The paper addresses key issues in safety performance assessment of highway infrastructure. The state of research in safety performance assessment methodologies in the US, specifically the Highway Safety Manual crash prediction algorithm, is first presented, with an illustration of how the algorithm can be evaluated for application outside the US. Fundamental to the algorithm's performance are crash modification factors (CMFs) for assessing how safety is affected as a roadway feature is changed. Issues in the development of these CMFs are discussed and illustrated with recent application examples in the development of CMFs for countermeasures targeted at improving intersection safety. Finally, the paper discusses future research directions. The paper is a culmination of several recent research projects, some of which are related the Highway Safety Manual, which was released in 2010, and is already being used worldwide.

highway safety, crash prediction models, crash modification factors, Bayesian methods, safety countermeasures