Pot, Teens and 40 Percent More Psychosis?

Maia Szalavitz

The “Drug Czar’s” office released a new study today and according to USA Today and the Washington Post’s reading of it, teen marijuana use raises the risk of mental illness by 40 percent.

But what the report and the media failed to examine is the complexity of this connection. While the report admits that many teen marijuana users are trying to self-medicate depression, rather than becoming depressed because they smoke pot, it continues to insist that marijuana smoking itself has been shown to cause depression and psychosis.

It cites one review of the research as saying “there is now sufficient evidence to warn young people that using cannabis could increase their risk of developing a psychotic illness later in life”; but it fails to mention that the same review also concluded, “evidence for affective outcome [depression, bipolar] is less strong, undermining the report’s claim about marijuana causing depression.

In fact, the data on marijuana’s link to psychosis is limited, and it would seem that there needs to be some predisposing factors for there to be cause and effect. The UK’s Advisory Committee on the Misuse of Drugs – under tremendous political and media pressure to declare a causal connection to support Prime Minister Gordon Brown’s call for tougher marijuana laws – recently reviewed the data again and concluded that

the magnitude of the effect of cannabis use on the subsequent development of schizophrenia does not appear to be substantial, in the population as a whole, with the cannabis preparations used during the late 1990s.

The problem is this: Marijuana use skyrocketed in the second half of the 20^th century, reaching a peak in the late 70’s and early 80’s in the U.S. where nearly two-thirds of teenagers tried it. But schizophrenia and psychotic disorders have stayed stable, and possibly even declined, according to some experts – even as marijuana potency increased.

If marijuana had a significant causal effect on schizophrenia, rates of the disease should have increased by now.

It’s not a good idea to self-medicate mental disorders with illegal drugs; but lying about the science isn’t going to stop teens from doing so.

The Risks of Formula and Formulaic Science Reporting

“Studies have shown,” says the New York Times Gardiner Harris, “that children who are fed formula have increased risks of ear and respiratory infections, obesity, diabetes and even cancer.”

But instead of the link text taking you to the actually studies which purport to show these connections, there are just links to Times query forms, which means you just have to take the reporter’s word for it that studies have indeed shown that children fed formula have increased risks for these diseases.

Unfortunately, in the increasingly overly-simplified, context-free world of reporting on health, the phrase “studies have shown” is often a formula for telling the reader what the reporter assumes has actually been shown.

Consider the studies that the Department of Health and Human Services (HHS) claim have shown a link between formula feeding and diabetes: Yes, the HHS found from a meta analysis that there is a 39 percent decrease in diabetes among those who nursed. But that’s comparing all and some nursing to no nursing at all, so that immediately complicates the picture - especially as one of the key narrative elements in the Times story is on getting mothers to nurse for longer than six months.

Then there is the problem that the groups showing the greatest association between breast-feeding and reduced diabetes are not representative of the general population (Prima Indians, U.S. and Dutch famine survivors); moreover, six of the seven studies in the HHS analysis were on adults, most of whom were drinking formula before the Infant Formula Act of 1980 gave the Food and Drug Administration the authority over the contents of formula. Also, in 2002, DHA (docosahexaenoic acid) and ARA (arachidonic acid) were introduced into some formula; these are both fatty acids that have been found in breast milk - and babies drinking that kind of formula were not included in any of these studies. (Read more by STATS on nursing and diabetes for other confounding factors which limit these studies).

What about serious respiratory infections? The most recent research does not support the contention that formula carries a higher risk, which is again driven by much earlier, and for various reasons, impeachable studies.

Cancer? One study finds that childhood leukemia is reduced by as much as 19 percent for breastfed babies compared to non-breastfed babies. But given that there are approximately 30 leukemia cases per million children, a 20 percent reduction due to breastfeeding avoids a risk of 1 in 150,000 that your child will develop leukemia; of these, 50 to 80 percent survive, depending on the type of leukemia. In other words, insisting that all women breast feed (and, by-the-way for more than six months) would save less than one life in 300,000.

Yes, there is robust evidence that nursing reduces ear infections and diarrhea, and in an ideal world, and with everything else being equal, it would be preferable if all women could nurse. But not all women can breast-feed or can breast-feed all the time.

Nevertheless, the government has decided that women should do so as a matter of public health policy, and the way to get them to do that, it would seem, is to scare them into nursing by over-stating the risks and playing up limited research on the risks of formula. Unfortunately too, the Times has continued to give uncritical support to this campaign by using a journalistic formula of its own.

Mixing Terminology on Drink-Driving Survey Causes Media to Pile On Midwest

Pity the upper Midwestern motorist – or, rather, the poor pedestrian that crosses his bleary-eyed path. According to a story on FOX News:

The upper Midwest has the worst drunken driving rates in the country, according to a government report that says 15 percent of adult drivers nationally report driving under the influence of alcohol in the previous year.

Wisconsin leads the way. The federal government estimates more than a quarter of the state’s adult drivers had driven under the influence. Rounding out the worst five are North Dakota, Minnesota, Nebraska and South Dakota.”

From Minnesota (MPR) to Florida (the Palm Beach Post), and along with the national news services like the Associated Press (“Study: Midwest has the worst drunken driving rates”), Lake Wobegon turned into Lake Woebegone. Naturally, this required a sociological explanation. As the AP reported:

Eric Goplerud, research professor at George Washington University Medical Center, said cultural and demographic issues probably have a role in the higher rates of driving under the influence in certain states. He said that religious affiliations in the Southeast often strongly discourage drinking, but that doesn’t occur so much in the upper Midwest. ‘A good part of the social life is around drinking,’”

The Milwaukee Sentinel Journal, on the other hand, just needed someone to explain why the findings required no explanation:

‘I’m not shocked, I’m not surprised,’ Nina J. Emerson, director of the Resource Center on Impaired Driving at the University of Wisconsin Law School, said of the latest report.

Nor did the finding startle Paul Moberg, senior scientist in the Population Health Institute at UW-Madison and co-author of a 2007 study on Wisconsin’s alcohol and drug use patterns.

‘Not really,’ he said. ‘I think that’s been what we’ve seen historically.’”

But did the government study really find what all the media accounts claimed it did?

The Substance Abuse and Mental Health Services Administration (SAMHSA), which is part of the Department of Health and Human Services, surveyed 127,283 drivers nationwide over the course of 2004, 2005, and 2006. And it describes the results as a measure of “driving under the influence of alcohol.”

Most people would assume (as many journalists did) that this refers to the proportion of drivers who were legally DUI, e.g., had blood alcohol levels (BAC) of .08 or more, the level at which they are legally intoxicated. But SAMHSA uses the phrase “driving under the influence” to mean having any measurable level of alcohol in the blood.

The problem with this approach is that it conflates legal with illegal drink driving – the person who has a drink after work and who has a BAC lower than 0.8 and the person who shouldn’t be in the driving seat.

Does it make sense to do this? If you assume that any amount of alcohol consumption should prohibit driving, then yes. This would mean that the presumptive level for intoxication should be set at .01 BAC. But that assumption runs counter to the research on BAC levels.

The National Highway Traffic Safety Administration (NHTSA), in its July 2001 report “Legislative History of .08 Per Se Laws” found that while the risk of being in a crash gradually increases at each BAC level, there was little difference in relative risk for drivers with BAC levels between zero and .05 This risk, however, rises very rapidly after a driver reaches or exceeds .08. This report also states that:

A BAC of .08 is a reasonable level at which to set the illegal limit. A .08 BAC is not typically reached with a couple of beers after work, or a glass or two of wine with dinner. The average 170 pound male would have to consume more than four 12 oz. cans of beer within 1 hour on an empty stomach to reach .08 BAC. The average 137 pound female would need at least three cans of beer in one hour on an empty stomach to reach that level.”

SAMHSA’s survey method, on the other hand, suggests that people who have a glass of wine at dinner are drunk driving if they get behind the wheel afterwards. (It should be noted that a BAC of .05 – or very roughly, the level produced by drinking two standard measures if sober – is favored by many other countries as the legal cut-off point).

Many of the news reports on the SAMHSA study also linked the results to the rate of alcohol-related traffic crashes and fatalities. Yet data from the NHTSA shows that 85% of all alcohol-related traffic fatalities in 2006 involved drivers with a BAC of .08 and higher, and 58% with a BAC of .15% or higher – the so-called incorrigible group of “hardcore drunk drivers” who cause havoc on our roads.

When providing statistics to guide public safety, making such distinctions are important - in this case, it’s the difference between doing something legal and doing something illegal. The upper Midwest may not be off the hook on drink driving for historical and sociological reasons, but it can’t be judged guilty on the basis of SAMHSA’s stats.

Al Gore Trusted More Than the Media on Global Warming - by Climate Experts!

We’ve just released the results of our STATS/Harris Interactive survey of climate experts and among the interesting findings (no surprise that there’s little disagreement about the basic fact that yes, the earth is warming and we’re at least partly to blame) is that Al Gore’s movie, An Inconvenient Truth, is rated as more reliable than any media source:

Only 1% of climate scientists rate either broadcast or cable television news about climate change as “very reliable.” Another 31% say broadcast news is “somewhat reliable,” compared to 25% for cable news. (The remainder rate TV news as “not very” or “not at all” reliable.)  Local newspapers are rated as very reliable by 3% and somewhat reliable by 33% of scientists. Even the national press (New York Times, Wall St. Journal etc) is rated as very reliable by only 11%, although another 56% say it is at least somewhat reliable.

Former Vice President Al Gore’s documentary film “An Inconvenient Truth” rates better than any traditional news source, with 26% finding it “very reliable” and 38% as somewhat reliable. Other non-traditional information sources fare poorly: No more than 1% of climate experts rate the doomsday movie “The Day After Tomorrow” or Michael Crichton’s novel “State of Fear” as very reliable.

For more detail, including the results of who’s been pressured to say what by the forces of government, check out the full release here.

We’ve also published a long look into climate modeling: just how reliable are climate models?

Obama Gets Pushed by a Poll

Rebecca Goldin Ph.D

Political Intelligence (field reports by reporters from the Boston Globe) stated this morning that people are disagreeing with Obama – at alarming rates. The story draws on a survey conducted by Rasmussen, which found that 56 percent of Americans disagree with Obama about his comments on small-town America. Even among Democrats, the rate of disagreement is 43 percent. On the other hand, among a “plurality of politically liberal voters,” only 33 percent disagree with 46 percent agreeing.

But the poll’s question pushed the answer. It was phrased as follows:

Obama said that in small towns in Pennsylvania, people ‘cling to guns or religion or antipathy to people who aren’t like them or anti-immigrant sentiment or anti-trade sentiment as a way to explain their frustrations.’ Do you agree or disagree?

What respondents didn’t hear was what Obama said before making this statement. The full quote was:

You go into these small towns in Pennsylvania, and like a lot of small towns in the Midwest, the jobs have been gone now for 25 years and nothing’s replaced them. And they fell through the Clinton Administration and the Bush Administration, and each successive administration has said that somehow these communities are going to regenerate and they have not. And it’s not surprising, then, they get bitter, they cling to guns or religion or antipathy to people who aren’t like them or anti-immigrant sentiment or anti-trade sentiment as a way to explain their frustrations.

Consider the difference: in the first context, the person polled is asked whether he/she agrees that people in small towns cling to guns, religion, antipathy to others, or anti-immigration or anti-trade sentiments. In the actual quote, the context is regarding those people in small towns where the “jobs have been gone for 25 years and nothing’s replaced them” and where the communities have not regenerated over the course of successive administrations.

Asking Americans whether they agree or not with the whole quote may well have had a different effect. According to the Rasmussen poll, only about 55 percent of those polled were following the issue “very closely” or “somewhat closely” and a full 15 percent were not following the issue at all. It may well be that almost half probably hadn’t seen the speech or even heard of the controversy, making the wording of Rasmussen’s question even more influential. How would a typical Clinton or McCain supporter have answered this question, given the sparse context? It makes the 56 percent disagreement seem rather modest.

Math with Monty

Rebecca Goldin Ph.D

There’s no way of better inspiring people to think about math than to give them a conundrum where the stakes are high. So it is in the Monty Hall Game, which became famous on the program Let’s Make a Deal in the 1980s. The game gets played as follows. There are three doors, numbered 1, 2, and 3. Behind one of them is a car, and behind the other two are goats. First, a player chooses one of three doors, say Door Number 1. The host then reveals one of the other two with a goat behind it, say Door Number 2. Then the player is offered to switch to Door Number 3 – should she switch?

Our intuition may not be our best friend here. Your gut may tell you that by opening a door that has no goat behind it, you have gained no information (you already knew that there was at least one goat behind Door 2 or Door 3). So your chance of winning is just 50/50. But here’s where it gets confusing. Your best strategy depends on how the game really played.

If the rules require the host to open one of the remaining doors with a goat behind it, then you should switch. When you picked Door Number 1, there was a 1/3 chance that your door has the car, and 2/3 chance that the car is behind one of the other two doors. After the host reveals that one of them has a goat, if you switch to the remaining door, you’ll have a 2/3 chance of winning.

On the other hand, if the host is not required to offer a switch, the odds may not be on your side to switch. And depending on what rule he’s following, your chance of winning or not could rise or fall considerably.

Take for example, the case where the host knows there’s a goat behind Door Number 2, and no matter what choice you make, he intends to open that door. So if you choose Door Number 2 right at the beginning, you lose. If you don’t choose it, then he opens Door Number 2 and offers you the chance to switch. Your chance of winning is now 1/2, regardless of whether you switch or not.

On the other hand, the host could be truly mean-spirited. If you choose a wrong door, he plans to open it and show you that you got yourself a goat. If you choose a correct door, you are offered to switch (and even offered money to take it). Now the chance of you winning if you switch has been reduced to zero!

This problem illustrates the importance in mathematics of having a very clearly stated problem. The problem of whether you should switch is not answerable without a clear statement of the host’s strategy.

Recently, the mathematics of this problem has been applied to psychological experiments involving choice rationalization. The New York Times’ John Tierney has done an outstanding job at explaining the mathematical mistake made by some researchers in establishing a baseline for preference while conducting these experiments. The graphics are extremely illustrative, both in establishing how the Monty Hall problem is related, and in how an excellent article about a confusing yet inspiring math brain teaser can be written.

Prosecutors Try to Gag Medical Testimony

STATS’ fellow Maia Szalavitz reports on the Huffington Post on a case where prosecutors are trying to get a gag order on an activist group for pain sufferers. It seems the testimony provided by the Pain Relief Network as to the legitimacy of prescribing opioids for chronic pain is proving difficult to deal with in the ongoing prosecution of doctors for drug trafficking.

Another Reason to Focus on the Risks of Cigarettes and Smoking

While activists, legislators and journalists spin themselves into a frenzy over unproven, unfounded, hypothetical, or, at worst, minor risks from chemicals in plastics and cosmetics, a new study reminds us that when it comes to infant health, exposure to the 4,000 chemicals in cigarette smoke should be a national priority. As Medical News Today reports, research published in the American Journal of Respiratory Cell and Molecular Biology shows that “exposure to cigarette smoke inhibits innate gene expression and impairs alveolar growth in neonatal mice.”

The results of this study indicate that exposure to CS [cigarette smoke] during the neonatal period inhibits expression of genes involved in innate immunity and mildly impairs postnatal lung growth. These findings may in part explain the increased incidence of respiratory symptoms in infants and children exposed to CS,” write the authors.

What’s remarkable, beyond further confirmation that a cigarette is the most powerful  delivery system for toxic chemicals among consumer products, is that the study is getting little attention from the media. Given the recent feeding frenzy in the press over the risks of Bisphenol A in plastic bottles,  based on one ridiculous study and contradicted by five massive, independent risk assessments spanning Japan, Europe and the U.S., one can only wonder how much attention the activist spin machine would have generated if the study had found similar results for plastic bottles.

When the Media Stitches Up Scientists: PBS Pulls a Fast One on Phthalates

One of the most depressing things about writing for STATS was the discovery that many scientists are deeply reluctant to talk to journalists. Most cite the fear of being misquoted or having their positions misinterpreted to fit a black and white narrative that, in scientific terms, is marked by numerous shades of gray. Just as depressing is the discovery that those who talk to journalists are often reluctant to correct mistakes because they fear jeopardizing their future access to the paper. Not all journalists are like this; we have been immensely impressed with how much effort some reporters put into trying to understand complex issues; but the reality is too many either don’t care or are presold on one side of a controversy.

We got a taste of that in the last few weeks, having been approached by PBS’ “NOW” for balance on a story about the supposed risks of phthalates. We could only wish that the media treated more pressing public health problems, such as tobacco and lung cancer in women – with the avidity they have gone after this topic. But nothing quite prepared us for the degree to which “NOW” edited out information that challenged the show’s narrative. It’s not that my colleague Rebecca Goldin, a woman recently honored by the Association of Women in Mathematics for outstanding work, was trying to make the case that phthalates were safe; she was simply saying that the evidence that they were dangerous was very weak.

Perhaps because a significant amount of the scientific evidence on phthalates contradicts the alarming message that Mark Shapiro, an investigative journalist from Berkeley, pushes in his new book, Exposed: The Toxic Chemistry of Everyday Products’ (which formed the basis for the NOW show), it was deemed confusing or inconvenient in the face of an urgent appeal by one scientist – Dr. Shanna Swan, the primary source in the NOW program – to respond to the possibility of risk.

But it’s hard to accept that as a possible explanation given that the arguments against Swan’s work (made by independent experts one might add, as well as the plastics industry) were ignored. And the degree of deception really hits home when you consider the full facts behind the strongest argument marshaled by Shapiro and NOW for phthalates being a risk: Europe’s ban on phthalates in toys for children under three.

The permanent ban in 2007 succeeded a series of temporary bans which were not supported by the scientific evidence presented by the European Union’s own risk assessments and which elicited some harsh criticism from the scientists involved in conducting these assessments (see the Royal Society of Chemistry’s Journal of Environmental Monitoring and EurActiv).

This was left, so to speak, on the editing room floor. And so the question of whether there was sufficient scientific evidence for precautionary legislation in Europe, and whether there is a need for such a ban here went unasked. Instead, the journalists at NOW made a decision: there ought to be a ban.

And here is a mathematician’s reply.

Time to Chill Out Over Drugs in the Water

Maia Szalavitz

A recent investigative report by the Associated Press has found traces of various medications – ranging from painkillers to sex hormones to antibiotics – in the drinking water of various locations around the country. The report has now sparked Senate hearings over what is to be done.

But what this intensive investigative report lacks is context and clarity. For example, the story said the drugs were found in concentrations ranging from parts per billion to parts per trillion. The distinction between the two is not trivial: Imagine telling someone that they either owed $100,000 $1,000 or $1,000,000. One might be affordable, the other, catastrophic.

And the AP’s reporting doesn’t tell us anything about whether drugs can have effects in such minute quantities. Homeopathic “medicine” is based on the idea that extremely diluted substances can cure illness, but research suggests that any results are limited to the placebo effect. If I drink water with 1 part per billion of codeine in it (say, 1 billionth of a gram or .000000001 grams) but the effective dose is 60 milligrams, I am highly unlikely to get any pain relief – or any side effects. The dose, as they say, is what makes the poison. If it wasn’t, the naturally-occurring chemicals in vegetables would kill us.

The story also doesn’t provide any means of comparison - how many parts per billion of oil or gasoline or other kinds of contaminants (some of this water comes from treated sewage, scary to think about what lurks in parts per trillion levels there!) are normally found in drinking water? The AP cites possibly dangerous effects on fish, but doesn’t note that these could also be due to other pollutants.

After all this, it’s worth pointing out that the New York Times covered this issue much more thoughtfully a while back, here.