Statistics repeatedly show that the vastly disproportionate majority of alcohol-caused injuries and deaths are caused by a few "problem drinkers" (for want of a better term). Thus, the first objective in any solution is to identify these individuals. In my experience, they can usually be identified by a combination of factors:

1. Their blood-alcohol level is not just high -- it is very high, say .16% to .30% or more.

2. This is probably not the first DUI -- and prior incidents are likely to be relatively recent.

3. There is a genetic flag: the individual is likely to have one or two "alcoholic" parents.

All right, we've identified some markers for who the problem is , but what do we do with them? To begin, let's understand what we don't do: we don't hit them with stiff jail sentences. If we do, we simply remove the person from society for a few days or months -- and on the day he gets out, he gets in his car and drives directly to a bar. What has been accomplished? Is society being protected -- or are we simply punishing people for drinking too much?

Since the punishment model clearly doesn't work for the problem drinker, we must consider the other criminal justice models: isolation, deterrence and rehabilitation.

1. Isolation. Yes, we can put the problem drinker in jail for a few months or even a few years, and we are safe from him for that period. But can we really afford to house tens of thousands more inmates? For how long? And what happens when they get out? For that matter, given the evidence, aren't we punishing them for a genetic condition?

2. Deterrence. How do you deter an "alcoholic"?

3. Rehabilitation. Once the favored approach in the criminal justice system, rehabilitation fell into widespread disfavor many years ago. Yet....Yet, this would appear to be the only logical approach with problem drinkers.

Ok, but what about the driver who is not a problem drinker but who is simply impaired from drinking too much? Answer: Treat him like any other misdemeanant. Statistically, we know he is unlikely to cause serious injury or death, but there is undeniably some risk there. Can this individual be deterred from such future conduct? Unlike with the "alcoholic", statistics show he can. Thus, it may be fair and productive to impose a fine on the typical first-offender, perhaps even suspend his driver's license for a short period; if a high blood-alcohol level is involved, say .15%, the punishment may include a 2-day jail term. But certainly not the punishments so destructive to families and careers that are now being administered to all caught up in the dragnet.

While we're at it, a refreshing approach -- and a healthy one for society -- would be to reinstate constitutional rights in DUI cases: due process, presumptions of guilt, denial of right to counsel, double jeopardy, the 5th Amendment right against self-incrimination, the right to confront witnesses, 4th Amendment roadblock violations, ad nauseum. (See "The DUI Exception to the Constitution".)

Does all of this finally solve the drunk driving problem? No: people will always drink and drive. But it will focus on the real threat -- the truly dangerous driver -- rather than on drinking and driving per se. And, in the process, reinstate the essential fairness and due process that has been slowly removed from the criminal justice system.