""In 2008 we lost 431 people in America. Out of those 431, 43% were alcohol- or drug- impaired drivers," said Gregor Smith."

About 40% of US traffic fatalities have a relationship to alcohol or drugs, which means that a sober driver striking and killing a pedestrian with a mere measurable amount of alcohol or drugs in their system makes the NHTSA definition of a drug or alcohol related traffic fatality.

Therefore, Gregor Smith is speculating when she states that those were all impaired drivers that died.

Additionally, NHTSA uses computer models and not real data many times when fabricating its statistics on "alcohol related traffic fatalities". Therefore, it would be a lie to state any of those statistics, especially when the definition of alcohol relationship is so wide it covers just about anything legal to someone sober driving to a liquor store while doing an errand as a favor for their roommate.