It is a very rare drug that will not kill someone.

The numbers involved with aceto are very very low.

Sure a rounding error, you’re a funny guy Balko, but what about the estimated 56,000 emergency room visits, 26,000 hospitalizations. What do those numbers do to the cost of insurance and health care in America?

Even in the absence of an actual threat from the DEA, so many doctors are sufficiently afraid of the perceived threat that they refuse to prescribe opioids properly.

You, Lee, seem to have hit the physician jackpot. I am sorry to hear you’re in need of pain meds.

The problem with drug dogs and evaluating their effectiveness is that they’re really good at smelling recently smoked pot (and other drugs) and drug users are really good at saying the dog is lying. What drug dogs are not very good at, even though it’s their stated purpose, is detecting 5 kilos of professionally packaged coke.

Also, I had a run-in awhile ago in Austin with a drug dog. I was one of the 48% who doesn’t carry weed while the car smells like weed so no harm, no foul. But that dog immediately identified where the unsmoked blunt was kept and the ashtray. It was quite impressive really. The APD K9 officer asked me (away from the others) when was the last time the truck had been smoked in and where any weed was stored. He said it was to track the dog’s performance and make sure he’s sharp and really had no reason to lie. I’ve had one other run-in with a K9 in the middle of nowhere on the interstate, and the thing went down exactly the same.

The cop who shot him was back on duty four days later, during which he was involved in a second fatal shooting. He’s now on paid desk duty.

If I were the coffee boy in his office, I’d be very careful when bringing him his order.

Doctors cannot prescribe straight hydrocodone for you to pick up at your friendly neighborhood pharmacy.

I have a couple of pill containers that would disagree with that statement.

Steve:
Yep, you’re right I screwd up the quick reading I gave the link. I thought WalMart was backing the a reform plan, not an employer mandate.

Technically you are on the right path, but just a tad bit off. Yes, we want to use the data we have in hand. However, the dog is allowed to “inspect” all 100 cars, so that is part of our data in hand so to speak. Using field data is tough since we wont know about those cases where the dog has didn’t find drugs or a bomb when indeed there was one.

I’d also add that dogs should go through not only re-tests, but also refresher training.

Are we going to get health care reform? Yes. Would Wal-Mart want to have some say in how it is forumalted? Yes. Why not do it so as to hurt your competitors. If Target has lower costs, then a federal mandate that has higher costing requirements but lower than Wal-Marts is a smart move. I’d also suggest you read up on how guys like J.D. Rockafeller made his money. A large part of his success depended on getting sweetheart deals from rail roads so that he could undercut his competitors. Another way to do that is to favor legislation that raises your competitors costs.

Regarding dogs:

Suppose we have 1000 trials, and that 10 of them are postive (bombs, drugs, etc.). What we want to know is:

Prob(Bomb|Alert)

We can find this by looking at

Prob(Alert|Bomb)Prob(Bomb)/Prob(Alert)

The probability of finding a bomb is thus proportional to the probablity that there is a bomb and the porbability that the dog alerts when there is a bomb.

Now if the probability that the dog alerts in general is 50%, then the probability that there is a bomb when the dog alerts is equal to the probability of there being a bomb. Using our numbers a 1% chance. In this case, the dog offeres no help in finding the bomb. If the probability of an alert is dropped to say 5% and the probability of the dog alerting when there is a bomb is 80% then we have the following

Prob(Bomb|Alert) = 0.16

In other words, even a dog that doesn’t alert that often, and alerts 80% of the time when there is bomb will still likely have a poor track record.

Unless I’ve messed up my math.

Based on this I’d say that dogs should be evaluated by an independent agency. Dogs should be checked both in terms of how often they alert when there is a bomb, how often they alert when there isn’t a bomb, and from this determine how often the dog alerts in general.

What also needs to be kept in mind is that a dog can alert even if the item is not there. That a dog alerts does not mean that there is something there. Also dogs should not be trained for cross purposes. Train a dog to detect bombs or drugs not both. If a bomb detection dog alerts but you find no bomb and drugs, I’d argue that the evidence should be inadmissable since the probably cause due to the dog’s “alert” was for explosives not pot or whatever.

Of course, it is a bit more problmatic when its the other way around, a drug dog alerting and finding a bomb. Do we really want a bomb builder/terrorist to be let lose? Hopefully this is sufficiently unlikely that it wont happen.

For chronic pain conditions, they are terrible, because most people develop significant tolerance to the effects of opioids, but no parallel tolerance to the hepatotoxic effects of NSAIDs. Consequently, people take more to get the same analgesic effect. So uncoupling NSAIDs and opioids is a GREAT idea, for chronic treatment. The problem, as has been pointed out, is that if you start prescribing straight opioids, the DEA will be crawling up your ass.

The real argument that needs to be made here is that a well-monitored opioid dependence is vastly preferable to both hepatotoxicity AND needless suffering due to puritanical drug policy.

Lets look at the #16 reference to #7 if of the 100 cars the dog alerts on two, but one is not correct he states that the percentage is 99% not 50%.

I guess that may be right? I’m not sure though, because it leaves 98 cars that the dog did not alert on. The police can’t search those cars, and they may all have marijuana in them. Now like I said my math is not great but I would put that at a 1% accuracy rate.

OK I understand that the chances are very slim that the remaining cars all contained drugs, but the real fact is we would never know.

If there is one thing we have learned especially lately (thanks to the Radley) is that criminal sciences, are not what CSI would have us believe.