Increases in crash risk and in proportion of "responsible"crashes that have been identified with increases in blood alcohol concentration are two inseparable aspects of a single epidemiological phenomenon. Identification of the relationship with alcohol per se has been possible only because studies controlling or accounting for time, place, and age have been carried out. In the absence of similiar study designs, it is not possible to attribute high proportions of "responsible" crashes involving drivers with other drugs to drug effect per se.