The need for better exposure data on vulnerable road users is a general issue in road safety studies, especially in studies concerned with urban environments with higher share of non-motorized transportation. One of the ways to compensate for the lack of data on vulnerable users is to incorporate the data on multimodal transportation infrastructure as a proxy for non-motorized exposure. The other alternative is to use regression models and estimate the number of people walking and biking based on the availability of infrastructure and land use patterns. This paper explores the potential of accessibility measures to serve as surrogate measures for vulnerable users’ exposure in crash studies. As the relationship between accessibility and exposure is well established in travel demand research, indicators of accessibility, defined as the number of destinations that users are able to reach within the given time, may be able to fill in the gap that exists in vulnerable users crash studies due to data unavailability. Proposed framework is developed for areal crash studies, in order to be able to capture area-wide effects which influence road safety of vulnerable users in urban environments. Research is conducted using data from Chicago, including crash data, socio-economic, land use, multimodal infrastructure, and travel demand data, aggregated on the census tract level. Accessibility measures are developed for pedestrians and bicyclists, using cumulative and weighted accessibility measures. Statistical models are estimated for fatal and injurious crashes involving pedestrians and bicyclists, using full Bayes hierarchical models to account for spatial autocorrelation and heterogeneity present in crash data aggregated on the census tract level. The purpose of the analysis is to explore the relationship between the indicators of accessibility and crash outcomes for pedestrians and bicyclists in urban environments, and answer how safety changes as accessibility to destinations increases for vulnerable users. The analysis results show that pedestrian fatal and injurious crashes can be associated with pedestrian accessibility indicators, while bicyclist accessibility does not seem to have a significant impact on bicyclist severe crashes. These results could serve to inform the decision makers in the process of planning and programming for long term safety targets in urban transportation systems.