Well, we do. Evidence from several sources goes into setting per se levels. That's evidence from basic laboratory studies of the impact of these drugs on skills, evidence from effects on driving performance or simulated driving performance, and then epidemiologic studies on the impact of having a drug on board on collision risk.
It's important to recognize that the key piece is the epidemiologic studies showing what the drug does to collision risk. We all know that having a good reaction time is required for safe driving. You see that if the driver in front of you slows down or stops, you need to react fast to slow down too. If amphetamines actually improve your reaction time in a laboratory situation, does that mean we should not have a per se level for amphetamines and maybe encourage people to use amphetamines and drive? No, because when we look at the epidemiologic studies, we see that drivers with amphetamines on board are more likely to be involved in collisions.
We need that range of evidence, and that substantial evidence is out there.