The Ecstasy research scandal

The Chronicle of Higher Education has a thorough review of the MDMA (“ecstasy”) research done, under funding from the National Institute on Drug Abuse, by George Ricraurte and his collaborators at Hopkins.

The story doesn’t quite make it clear just how outrageous some of the research misconduct involved actually was, party because the author seems not to understand the details. (The famous “hole-in-the-brain” images involved doing odd data transformations and turning up the gain on a false-color-imaging program so that reductions in brain metabolic activity well within the normal range showed up as black areas; the primate deaths should have been queried because the experiment was supposed to replicate the experience of human MDMA takers, who die at rates of less than 1 per million as opposed to 20%.) The story also fails to discuss Ricaurte’s active role in making human MDMA studies virtually impossible.

The fact that Ricaurte is still defending the publication of that study without a prior careful autopsy of the dead primates suggests that he’s close to incorrigible. And if he really believes what he says — that, because 10 is a small sample size, 2 in 10 in the sample is actually consistent with a base rate of 1 in a million — he ought to consider retaking elementary statistics.

I hope someone at Hopkins is looking hard at Ricaurte’s lab, and that the Human Subjects folks at HHS have started to ask hard questions about coaching subjects to lie about whether they meet the exclusion criteria.

I hope that the editors of Science will take this as a wake-up call: that study never should have been published. (The quote from Donald Kennedy is anything but reassuring on that point. The reviewers couldn’t have known about the drug mix-up, but they should have known that the fatality rate was inconsistent with the claim that the experiment modeled ordinary human MDMA use.)

In addition, I have less plausible hopes: that the people in Congress and at the Sentencing Commission who made decisions about MDMA based in large part on Ricaurte’s studies are now rethinking those decisions and have made a mental note to be less credulous in the future, and that the people at AAAS who decided that Alan Leshner would made a good president of the organization feels as stupid as they now look.

The story is well worth reading, and is not without its encouraging moments. As soon as Leshner — who had invested his personal prestige and that of NIDA in an anti-MDMA crusade based largely on Ricaurte’s work — had left, NIDA quickly backed off, even before the retractions. And the current NIDA Director, Nora Volkow, is allowed a last, sensible word:

“The question that comes to light is, why has this attracted so much attention?” she says. “And I think perhaps it’s because some people are exaggerating the adverse effects of drugs.”

That’s true, of course, but it’s quite amazing to hear a NIDA Director — the head of an agency which in the past has done more than its share of such exaggeration — say so on the record.

Full text of the Chronicle article

Victory through redefinition:
    hoking the drug budget

John Walsh reports in the latest Drug Policy Analysis Bulletin that the Bush Administration has satisfied the long-expressed desire to rebalance the drug budget between supply-control spending (enforcement) and demand-control spending (prevention and treatment) the old-fashioned way: by lying about it. Suddenly the costs of prosecuting and incarcerating drug dealers have disappeared from the budget, bring prevention and treatment into parity with enforcement.

In a larger sense, the whole issue is a silly one. as Sally Satel and I explained some time ago, but that doesn’t excuse trying to fix it with looking-glass logic.

At last, some real progress on medical cannabis

A British company called GW Pharmaceuticals has developed a sublingual spray called Sativex which contains all the psychoactive chemicals in natural cannabis, and that medicine is likely to be approved in Britain for the treatment of MS within months. The rest of Europe and Canada will probably follow quickly, and it’s quite possible that the US won’t be too far behind.

Sativex, an extract of the whole plant rather than a blend of synthetics, contains — unlike the whole plant material itself — a constant ratio of the many active cannabinoids. The first version to be approved will have a 1:1 ratio of cannabidiol (CBD) to delta-9-tetrahydrocannabinol (THC). That’s somewhat more cannabidiol than most illicit-market cannabis, which should make using the spray somewhat more anxiolytic and somewhat less intoxicating than smoking a joint. (One of the reasons the only currently marketed cannabinoid medication, Marinol, has not been widely used is that its pure-THC formulation leads to a relatively high incidence of panic and dysphoria.)

The method of administration means that the effects will also be somewhat slower to come on. The manufacturer claims that most MS patients can get relief from spasticity and the related pain without becoming subjectively stoned.

A na├»ve observer would expect to find the proponents of medical marijuana dancing in the streets, and its opponents mourning and grumbling. After all, approval of Sativex would amount to a concession that cannabis has in fact had therapeutic value all along, and that by stubbornly refusing to approve it the government has been denying relief to large numbers of patients, some of them suffering very badly. (Yes, there are advantages to a standardized — and perhaps optimized — cannabinoid profile, and to using a sublingual spray to avoid the throat and lung insult of smoking, but those are clearly second-order questions.)

But in fact the drug czar’s office is cautiously welcoming the new development, while the only criticism of Sativex is coming from prominent advocates of medical marijuana such as Lester Grinspoon and the Marijuana Policy Project.

Medical marijuana has been one of the very few drug-policy issues where the public sided with the “reformers” rather than the “drug warriors,” and Sativex may represent a way for the warriors to get out of an argument they can’t win without taking the (to them) unthinkable step of admitting that pot-smoking can actually be therapeutic.

LA Times story on Sativex

NY Times story on Sativex

Information from GW Pharmaceuticals website about Sativex and planned additional products with higher and lower THC:CBD ratios.

[More here in this post on The American Street.]

Obituary for a useful data series

Fox Butterfield has the details on the cancellation of the Arrestee Drug Abuse Monitoring (ADAM) program in today’s New York Times. Other than paying entirely too much attention to the views of a well-known loudmouth from UCLA, it seems to be a very competent story.

This is bad news for those of us who think about drugs and crime, but there’s also a larger lesson here. The indifference of the Bush Administration to the actual facts about the world is among its most salient characteristics, and the country will be paying for that indifference for a long, long time.

Previous post here.

Update My old friend John Coleman, who used to be the #3 at DEA as Assistant Administrator for Operations, has some thoughts:

The Justice Department has decided to end the ADAM program (see NY Times article below). The importance of ADAM always has been its stark statistics showing the large percentage of criminals high on drugs and alcohol at the time of their crimes. ADAM surveyed arrested felons and then drug-tested them to confirm their statements about drug use. It was all voluntary but showed, nonetheless, extraordinary levels in some cases of drug use by criminals. I recall several years ago reading that more than half the juveniles arrested for homicide in Washington DC tested positive for pot. ADAM was effectively the only data system designed to test felons and have the info confirmed by urine tox screens. Without ADAM, we can only speculate about the connection between drug use and crime. Chalk up a victory for the pro-drugs crowd on this one and shame on Justice and the White House for going along with their eyes and purses closed. Back in the days when I debated druggies on legalization issues, ADAM was the only data source they couldn’t rebut because it was factual and scientific and validated with testing. The best that they could argue was that the screens didn’t tell us how much dope the felons took or when. Recent developments in urine tox screening techniques have overcome some of these earlier limitations and we can now determine ranges for how much and when drugs were taken. But it won’t matter. Good-bye, ADAM.

Perhaps even more surprisingly, my old pal Mark Kleiman sees this the way I do! We are usually on opposite sides of these issues. And, he is right about the National Survey on Drug Use and Health (the former National Household Survey on Drug Use). It’s overloaded with demographic and social info and rests much of its analysis on a factor it defines as “lifetime use,” which means — quite contrary to what it sounds like it should mean — a person who has ever used an illicit drug even once in their lifetime. I wish someone could tell me the significance of this and why it’s worth $50 million a year to know this perfectly worthless “fact.” The pols these days are criticizing the intelligence community for not knowing what was happening in the streets in Iraq and other places in the Middle East prior to last year’s war. They may have a point. Given the billions spent on drug control here in the US, it would make sense that we would be improving, not removing, our essential drug intelligence collection systems, like ADAM, so that we know what’s happening in our streets and we don’t continue to waste money and time measuring the fact that someone out there smoked a joint 25 years ago. Wouldn’t it make more sense to know whether robbers, rapists, and other felons are high on drugs when they commit their crimes against us? I can’t do very much about the guy who smoked a joint 25 years ago and I’m not sure I need to do anything at all about him. But, I can and should do something about crime by drug users or drug use by criminals. Good-bye, ADAM!

John and I agree this is terrible news, but disagree about its political valence. The true hard-core “drug warriors” have always been more concerned with middle-class drug use, and in particular juvenile pot-smoking, than with the heavy chronic use or hard drugs by criminally active uses that constitutes the bulk of the actual drug problem as measured by illicit-market dollars, health damage to users, infectious disease, and crime. That’s what all those “use a drug, sponsor a terrorist” ads were about. Concentrating on the kids is partly a way to pander to the fears of middle-class parents, partly a wedge issue designed to make opponents look “soft on drugs.”

Gathering data showing that the real drug problem doesn’t fit that Bill Bennett image doesn’t help that political program at all. Worse, (as John points out to me in an email I quote with his permission) information about heavy drug use by people who are nominally under criminal justice supervision cries out to have something done about it. (That’s what my pet testing-and-sanctions-for-probationers program is all about.)

So the ADAM data were, as John puts it, like a red dashboard warning light pointing to the need for an expensive repair. It’s cheaper — in the short run — just to take out the warning light.

Opting for ignorance:
    ADAM program killed

As the National Academy of Sciences pointed out a couple of years ago, one fundamental problem with our approach to drug abuse is that we don’t know nearly as much as we need to know about what’s going on. And Peter Reuter has put his finger on one of the causes of that ignorance: While the overwhelming bulk of the activity in drug abuse control consists of law enforcement, almost all of the reasearch money comes from the health side.

The two big national surveys on drug abuse — what used to be called the National Household Survey on Drug Abuse (NHSDA) and is now called the National Survey on Drug Use and Health, which I suppose must be the NS-Duh, and the Monitoring the Future survey of high-school students — between them cost more than $100 million per year. They give us an excellent picture of casual, non-problem drug use and virtually no usuable data about the actual drug problem.

The Household Survey, for example, can be used to estimate the quantity of cocaine consumed each year, and the answer comes back somewhere between 25 and 30 metric tons. The actual quantity is about 270 tons. So the Household Survey misses about 90% of the action. (It also finds, absurdly, that there are fewer than half a million more-than-weekly cocaine users, and that fewer than 10% of them have ever been arrested.)

That’s no surprise, once you think about it. We know that 80% of the quantity consumed is used by 20% of the users. And those heavy users are disproportionately drawn from the non-household population (homeless or institutionalized) or from the 20% of the people on the Household Survey target list who either can’t be found or refuse to answer.

So where can we find the heavy users? Why, in the jails, of course. About 75% of heavy cocaine users get arrested in the course of any given year, either for drug possession or (even more frequently) for crimes committed in order to get money to buy cocaine: theft, dealing, and prostitution. The number of arrestee heavy users is about three times the number of non-arrestee heavy users.

We know that due to a data collection program once called Drug Use Forecasting (DUF) and now called Arrestee Drug Abuse Monitoring (ADAM). The National Institute of Justice calls ADAM the centerpiece of its drugs and crime research program. It costs a few percent of what the two big data-collection programs cost, which makes it by two orders of magnitude the most cost-effective data collection program in the entire world of drug abuse.

But that money comes out of the tiny budget of the National Institute of Justice rather than the huge budget of the National Institute on Drug Abuse, and NIJ’s budget is being squeezed.

So today, the day after the President’s State of the Union Address proposed spending $23 million a year on drug-testing programs in schools (an approach which has never been demonstrated to have any useful effect) the National Institute of Justice issued a stop-work order shutting down the ADAM program.

This step was taken without any consultation with any of the non-government experts on drug abuse, or I would surely have heard about it before Fox Butterfield of the New York Times called today to tell me that the announcement was on the NIJ website and ask me what I thought about it.

I suppose if you’re running an administration where facts are never allowed to interefere with decisions, it’s not necessary to gather any actual data.

Update A reader tells me that there is a backstory here. Apparently the ADAM contract was recently rebid and taken by firm that grossly lowballed its bid to undercut the incumbent, and then tried to push its price back up after the award. (There are several tricks an unscrupulous contractor can use, all of which more or less boil down to pretending that the activities promised in the bid were less than what actually needs to be done, and insisting that the agency issue “change orders” — triggering additional payments — to get to a reasonable performance level. “Oh, you didn’t want your reports printed on toilet paper? That will be extra, then.”

Apparently, the NIJ brass decided to strike back by cancelling the contract entirely.

This does indeed make the story more comprensible, but it doesn’t change the basic fact that it’s inexcusable. Apparently the decision was made over the protest of the drug czar’s office, which gives you some idea how weak that operation now is.

Second update Wrong! Everything in the update above is nonsense. Not sure what the crossed wire was with my source, but NORC, which has an impeccable record of competence and integrity, has had the big ADAM contract for several years now. So whether what I heard was NIJ spin or Abt spin (Abt having beeen the previous contractor) or merely random hot air, it certainly wasn’t the case.

Sorry to be the source of disinformation. I wish I could say “Won’t happen again,” but of course it will. All I can say is that the next time it happens I’ll let you know as soon as I do.

Update here

The Dallas police fake-drug scandal

If I hadn’t lent my car to someone who switch from the CD player to the radio and left it tuned to NPR, I probably never would have heard about the Dallas police fake-drug scandal. Googling it, I find no mention in any of the national media, except for one column by Ruben Navarrette carried by the WaPo syndicate.

As I piece it together, it’s pretty hair-raising.

Dallas police paid their drug informants based on the quantity of drugs seized. So some informants decided to manufacture cases by planting fake “cocaine” — variously described as the powder used to chalk billiard cues and as ground-up gypsum wallboard — on about 80 Mexican immigrants.

The police did “field tests,” all of which mysteriously registered positive for cocaine, and testified to having witnessed transactions that never happened. (After a long investigation, the U.S. Department of Justice indicted only one cop, and he was acquitted after what seems to have been a fairly badly bungled prosecution.)

The defendants, charged with possession of massive amounts of cocaine, were held on high bail, and since they weren’t in fact drug dealers they sat in jail awaiting trial. The public defender’s office refused to pay for independent lab testing, and several of the defendants pleaded guilty to avoid 10- and 20- year mandatory sentences.

Eventually the truth came to light. One source credits that result to a new law passed after the Tulia scandal; NPR says it was because some of the defendants’ families paid for private lawyers, who arranged for their own chemical analysis. Eventually, the DA somewhat grudgingly dismissed all the cases. I can find no mention of compensation being paid to any of the victims. Of course, the defendants who were here illegally and have been deported aren’t around to sue.

So what do we learn from this?

1. Almost all defendants are guilty. But there’s a big difference between “almost all” and “all.”

2. Paying informants is always dangerous. Informants lie. So competent investigation and prosecution requires double-checking what they say.

3. Enforcement against transactional crimes is always a tricky business. As long as we have drug laws, we’re going to have problems like this one. I think we need those laws, but that means that we need to be vigilant about the problems.

4. No police officer ever got a promotion for showing that someone was innocent.

5. Forsensic laboratory work is the stepchild of law enforcement. That needs to be fixed.

6. In most of the country, public defenders are underfunded. In some parts of the country, and especially in the South, they are not only grossly underfunded but under substantial political pressure not to slow down the wheels of “justice.” At some point, the federal courts are going to have to step in and enforce the Sixth Amendment right to counsel. Justice shouldn’t be for sale.

7. Juries don’t mind convicting cops for stealing, but they really don’t like convicting them for lying in the course of their duties.

8. Every state ought to have a small but well-paid staff of lawyers and investigators, with the same powers as ordinary police and prosecutors, whose job is exonerating the innocent and punishing those whose false testimony led to their incarceration. The penalty for falsely implicating someone in a crime should be the same as the penalty for the underlying offense.

9. What liberal media?

Update: Jim Schutze of the Dallas Observer has more on the scandal and the new investigation. He doesn’t like what he sees.

On one point I have to differ with Schutze: If I were a cop who had faked evidence, the last guy I’d want on my case would be an ex-FBI agent. I have lots of bones to pick with the Bureau, but the people there are pretty much straight arrows, and by and large they take a very dim view of people who cut corners. They also think they’re a superior breed of human being to the ordinary cop, which is a problem when they have to work with the locals but means that they don’t really identify wit the potential defendants in this case.

Remember, this wasn’t about faking evidence against guilty people, or even people the cops thought in good faith were guilty (which happens more than you’d like to think, and which many people in law enforcement don’t really disapprove of): this was framing people to get the numbers up, and maybe for a share of the bounty money. The cops involved in this were dirty, dirty, dirty, and the average ex-Feeb will get as much satisfaction from sending them for long spells as he would from doing the same to routine bad guys. Maybe more.

MDMA neurotoxicity:
   “An ouchie for George”

Today’s New York Times has a devastating article on the research methods of George Ricaurte, whose studies purporting to show the neurotoxic effects of MDMA (“ecstasy”) were used to support the original prohibition of the drug and have since been used to support stiffer penalties, ancillary laws such as the RAVE Act, and suppression of human research into the drug’s potential benefits. (Full text, and a long list of errors in the story, under “keep reading.”)

As one physician commented on a listserv that follows this class of issues, the story is “what we in the pediatrics world call an ouchie for George.” And Ricaurte has another ouchie coming fairly soon, in the form of a Peter Jennings special that keeps getting delayed as the producers add more and more detail about his problems.

The new director of the National Institute on Drug Abuse, which has funded Ricaurte’s studies to the tune of $10 million over the past two decades, was clearly aware of problems with his work before the latest article, and before Ricaurte’s retractions of two of his papers, including one in Science.

But Hopkins, where the work was done, is apparently still in denial, or at least is trying to sound that way. Given the sensitivity there to human-subjects issues — a couple of years ago, the feds came close to shutting down medical research at the whole institution after the death of a healthy volunteer in an asthma study drew attention to various technical deficiencies in the Hopkins process for reviewing experimental protocols — I’d expect the dean of the medical school to be asking some very serious questions right now.

What doesn’t seem likely to happen, but should, is a review of all the legislative and regulatory decisons made, here and abroad, on the basis of what we now know to be terribly flawed studies.

We didn’t need Ricaurte’s brain imaging to tell us that MDMA does something lasting to the brains of the people who use it repeatedly: the drug’s atypical pattern of quasi-tolerance is strong enough evidence of that. (For many if not most users, the valued effects of the drug, other than its pure stimulant powers, diminish fairly rapidly, in a non-dose-reversible, non-time-reversible fashion, with cumulative lifetime dosage.)

However, open questions remain about how extensive and how damaging those changes are, and what if any steps users could take to prevent or limit them. But the exaggerated claims about MDMA’s neurotoxic effects made it virtually impossible to do the studies that might have demonstrated the efficacy of various precautions. Despite the laws, Americans currently consume tens of millions (perhaps hundreds of millions) of MDMA doses per year. It’s probable that various relatively simple steps could significantly reduce the resulting brain changes, but that possibility has never been the sort of dedicated research effort that has gone into “proving” how dangerous the drug is.

Moreover, the proposed therapeutic applications of MDMA would involve administering it on a very small number of occasions — in some cases, only once. The probability that a single MDMA dose of known purity and quantity administered under clinical conditions could do significant harm seems extremely remote. But here again, the neurotoxicity scare has prevented the relevant studies from being carried out.

With any luck, the latest scandal could lead to a more sensible set of approaches. But as Bill Bennett would no doubt say, don’t bet on it.

Full text of NYT story, with errata list

[Previous posts Sept. 9 and September 25.]

Drug free school zones

After a long hiatus that was almost entirely my fault, the Drug Policy Analysis Bulletin is back in operation, due almost entirely to the efforts of our new managing editor, Douglas Ross.

The latest issue features a study by Will Brownsberger and Susan Aromaa of Join Together, which shows how the Massachusetts “drug-free school zone” law has turned into a general sentence-enhancement law with roughly no relationship whatever to the problem of drug dealing in or around schools. (In older cities and towns in Massachusetts, virtually any location is within the statutory distance of one school or another.)

The next issue, with John Walsh’s analysis of the smoke and mirrors behind the national drug budget, should be out in January.

Why make “structuring” a crime?

Jane Galt thinks (1) that the drug laws are a bad ided and that (2) the money laundering laws are an even worse idea, and show how much too far the drug war has gone.

Neither of those propositions is transparently false. The money laundering laws certainly demonstrate the Hayekian point Jane makes: the more ambitious a law is, in terms of changing behavior, the more additional laws it will need for its support.

But I don’t think it’s quite right to say, as Jane does, that the conduct forbidden by the money-laundering laws, when it comes to money withdrawn from banks by users of illicit drugs to give to sellers of illicit drugs, is absurdly remote from any actual social harm. One of the better arguments against the drug laws is the illicit enterprises they help create. The laws about reporting currency transactions are designed to impede the functioning of those enterprises.

(Note: While Jane is right that it would technically be a crime to “structure” transactions to avoid currency reporting even if the purpose of the transactions were entirely licit, the reports themselves are not matters of public record. Therefore there’s no actual reason for someone to conceal the cash he spends on his yacht or stamp collection from the IRS, unless he’s doing so to conceal the fact that either he, or the recipient of the funds, evaded or intends to evade taxation.)

Just about the only damage Rush Limbaugh’s drug use did to anyone but Rush Limbaugh and his intimates was the financial support it gave to the illicit drug trade. So his efforts to conceal the money involved were actually more proximate to social harm than his actual possession and consumption of boatloads of synthetic narcotics.

Cash creates the capacity for undocumentable transactions. That is highly convenient not only for drug dealers but for tax evaders, bribe-payers and bribe-takers, and anyone else whose sources of income are illicit.

Making it harder to acquire and spend cash is a reasonable approach to dealing with large-scale transactional crime. Whether the risk to liberty created by the resulting increase in state power is not, justified is a reasonable topic for debate. But I don’t think it’s as silly as Jane makes it out to be.