This paper investigates road accident severity and the possibility of an accident occurring in an urban area by analyzing real-time traffic data. Accident data from a main highway in Athens, Greece were collected for the period 2006 through 2010. The traffic data were measured in real-time from the main Traffic Management Center in Athens. Logistic regression models were developed and the results suggest that the road accident severity is influenced by traffic density, the type of vehicle and the type of accident. When the traffic data are separated in two groups of peak and off-peak hour accidents, the parameter of traffic density is the only one that appears to be significant. Therefore, traffic volume is the only parameter that has an impact on the likliehood of an accident occurring.