James Schneider  

Deception-Based Medicine

Two cheers for the tasteless... AMA Highlights...

Emily Oster stirred a lot of controversy when she suggested that pregnant women could enjoy a drink of alcohol if they so chose. Why did she arrive at a view that clashed with much of the American medical establishment? She named two reasons. The first is that she did a better job of sifting through the evidence than the medical experts had done. More importantly, she was actually trying to communicate the facts. Many others are practicing deception-based medicine.

She explained in Expecting Better:

But the bigger thing, I think, is the concern (which was expressed to me over and over by doctors) that if you tell people they can have a glass of wine, they'll have 3 (or one giant "bowl-o-wine"). Even if one isn't a problem, three are. Better to say you can't have any, as that rule is easy to understand.
Oster did a tremendous amount of research before concluding that a little wine is okay. But it doesn't take skills like Oster's to verify that mothers are oversold on the dangers of moderate drinking. The National Institute on Alcohol Abuse and Alcoholism (NIAAA) has a website where they answer questions in an easy to understand manner. The first question under "Drinking and Your Pregnancy" is "Can I drink alcohol if I am pregnant?" The response:
No. Do not drink alcohol when you are pregnant. Why? Because when you drink alcohol, so does your baby. Think about it. Everything you drink, your baby also drinks.
This response is a rejection of scientific thinking. How could pregnant women possibly know whether mild amounts of alcohol were beneficial or detrimental to the human fetus without reviewing medical research? Given that moderate alcohol consumption might be beneficial for adults, there is no obvious reason mothers should know that it will be bad for the fetus.

Despite their overly simplistic advice to mothers, the NIAAA is not a mindless critic of drinking. Other parts of their website contain dispassionate science:

Women who give birth to FAS ... children generally practice frequent and protracted binge drinking, which produces high blood alcohol concentrations. Although a majority of women in many studies reduce their alcohol consumption during pregnancy, mothers of FAS children generally do not ... Most women who give birth to FAS children live within a social and cultural milieu that tolerates, condones, or is ineffective in dealing with problem drinking.
In another section of the NIAAA website:
The minimum quantity of alcohol required to produce adverse fetal consequences is unknown ... Clinically significant deficits are not common in children whose mothers drank less than approximately five drinks per occasion once per week ...
So why is the NIAAA so categorical when they advise expectant mothers not to drink? Interestingly, they provide the answer in their scientific summary:
Despite the many gains in knowledge, we still do not know if there is a "safe" dose of alcohol that can be consumed by pregnant women without risking damage to their unborn children. Until such a safe dose, if it exists, can be determined, the only responsible advice to women who wish to become pregnant and to those who are pregnant is to avoid alcohol use entirely.

This might be plausible. However, the case for zero alcohol can be made without urging women to rely on their biological intuitions.

In any event, the NIAAA clearly has two messages tailored to two different audiences. On the one hand, they dispassionately explain the state of science using a small font, dense text, and the language of the specialist. Here they admit that much ignorance and confusion exists. On the other hand, they unambiguously advise expectant mothers to abstain from drinking alcohol as if the science was a settled matter. This is written in easy-to-understand language with lots of white space. This example illustrates a common phenomenon. Ordinary people do not receive medical information in an unbiased manner. Instead, medical research is frequently passed through a filter so that it provides only what is prudent to believe. Seemingly, the unspoken belief is that people can't handle the truth.

Completely exonerating moderate drinking with scientific data would be a daunting task. Making such an exoneration less likely is the fact that scientists may not be looking very hard. As Robin Room asserts in The Lancet:

But there has been little research on what a safe level might be, or just what are the risks of occasional drinking (with no binge spikes). The scientific agenda seems to be set by an implicit social policy judgment that the issue is better left unsettled: any answer other than zero would undercut the justification.

Comments and Sharing

COMMENTS (26 to date)
Daniel Artz writes:

So the Medical Establishment treats us all as unthinking boobs, complete morons, children who need to be protected. Why not? The Government has been doing that for decades.

Silas Barta writes:

I remember someone (I think Bryan Caplan?) going through something similar with his doctor. Caplan asked his doctor a) if he was aware of the research about the benefits of moderate alcohol, and if so, b) does he ever suggest that non-drinking patients start drinking for these benefits?

The answers were:

a) Yes, it's sound, and
b) Heck no!

Unfortunately, I can't find that post. Does anyone remember where that came from?

James Schneider writes:

@Silas Yes, this is definitely something Bryan has commented on.

Dan King writes:

I think it's likely that a glass of wine improves ones driving, too. A person is more relaxed, less likely to road rage, and in less of a hurry. So as they used to say: Have one for the road. Maybe it's good advice.

James Alcontana writes:

This reminds me of the medical community's willingness to essentially fabricate an entire body of science: secondhand smoke. This "science" provided the underpinning for a wayward invasion of property rights and personal liberties. Of course, the ends (reduction of smoking) were justified by the means in the view of the medical obfuscators.

I think the medical community's binary approach to matters of personal pleasure breeds distrust among those of us who realize that matters of personal consumption involve tradeoffs. I applaud Ms. Oster for analyzing those tradeoffs on her own in the absence of reasonable guidance from the condescending prohibitionists.

Charley Hooper writes:

Think about the incentives. The authorities don't enjoy the benefits of that glass of wine; they only see the problems that drinking might cause. Plus, they are supposed to be the responsible party. So from the authorities' perspective, there's a big downside ("You were wrong and my baby was harmed!") and virtually no upside. The authorities are simply acting rationally.

This is a problem with all situations involving authorities. Be thankful they are just providing advice and not dishing out prison sentences.

David Friedman writes:

I don't know if Bryan has discussed the issue, but I have in an old blog post.

Lauren (Econlib Ed.) writes:

There is a related EconTalk interview at Oster on Pregnancy, Causation, and Expecting Better, for those who are interested.

ZC writes:

Question for you James: Would you let the mother of your child consume alcohol while she's pregnant?

This part of your post seems to sum up the fundamental issue you want to raise in the post:

"the NIAAA clearly has two messages tailored to two different audiences...the language of the specialist...easy-to-understand language. Ordinary people do not receive medical information in an unbiased manner. Instead, medical research is frequently passed through a filter so that it provides only what is prudent to believe. Seemingly, the unspoken belief is that people can't handle the truth."

Of course there are the 'simple' and the 'expert' versions when discussing any bit of medical advice/decision making. There must be some sort of 'translation' from the findings of various studies (which may utilize complex statistical analysis, contain various caveats as only certain populations were studied, under specific circumstances) as understood by the medical community to advice that can be given to the patient. As you (would hopefully) understand, the ability to process and act upon that information varies greatly from patient to patient. Therefore, and individual physician may readily give greater 'expert' level detail to a patient capable of using such to inform their decision making. However, other patients need a much simpler version. In many practice settings, a large majority of the patient population may sign their name with an 'X', read at less than an 8th grade level, or have other impediments to understanding the complexity of something along the lines of "we don't know the bound below which the risk of such an adverse event becomes zero, please look at these stochastic risk models in chart 1".

So, if you're a national organization, it makes sense to offer up broad advice catering to the lowest common denominator, with footnoted explanations of the complex (and often incomplete) underlying science.

And Charley Hooper's reply is great.

Krishnan writes:

A related issue is one of "nicotine"/other cigarette ingredients ... Even as Governments rake in money through selling what Governments consider "evil" - all we hear is alarm ...

Yes, the epidemiological connection between smoking and many cancers are striking - and we know the risks in smoking ... THAT has now morphed into the dangers of "second hand smoke" (and there is SIMPLY NO evidence) ...

Worse, we do not know why certain very heavy smokers do not develop all those nasty cancers/etc ...

Imagine a scientist applying for a grant to study the "Advantages of Smoking for People with certain Genotypes" - He/She would be laughed at OR thrown out of the building - for EVEN DARING to ask a "scientific" question ...

"Nicotine" (like ethanol, caffeine) is a drug - and like all drugs, have positive and negative effects - in some cases, (cigarettes) decisions seem to have been made that "CIGARETTES ARE EVIL and SO WE MUST DO THIS/THAT")

David W writes:

I'm thinking more about the "scientists might not be looking very hard" angle. I imagine solving this question requires just as much work as any other in medicine. But the typical problem leads to huge results, lives saved, and sometimes even lots of cash. New treatment for cancer? Awesome! You've saved lives, and people will pay a lot for that. New treatment for vision problems? Great, now people won't go blind, and they'll pay for that too, happily.

Definitive answer to fetal alcohol processing? Um...some women will have a slightly more pleasant pregnancy. In a way that it's impossible to profit from, as a researcher. Plus, if pleasant pregnancy is your goal, I wouldn't even start here: focus on morning sickness or labor pains or probably a dozen other things that are more of an issue than changing an alcohol ban to an alcohol limit.

I mean, it's still knowledge, with the potential to lead elsewhere. And proving it would be useful in itself, if small. It wouldn't be a waste to research. But leaving it as a low priority seems perfectly rational to me.

RPLong writes:

I see this as running far deeper than merely the medical field. People who give professional opinions these days seem (to me) to speak with far more thunderous authority than they did, say, 50 years ago.

Are people less able to cope with nuance? Would an at-risk patient be able to make an informed judgement if her doctor told her something like, "Because we are unsure of the impact of multiple servings of alcohol on an unborn fetus, but very certain of the impact of excessive alcohol consumption on fetuses, we recommend complete avoidance of alcohol." ?

Your eyes probably glazed over as you read that. Mine very nearly glazed over as I typed it. It's a cultural change - for better or for worse, we seem to prefer being barked at. :/

Finch writes:

It's funny to contrast this with informed consent or use of tainted data as discussed in Caplan's Mengele post. It seems like the question of when it is okay to make decisions for other people is one for which we don't have clear answers.

For example, only the fringe rejects the idea that the atomic bombing of Japan, killing about a hundred thousand to save millions, was moral. In this post we see doctors perfectly comfortable lying to patients and helping propagate a lie. In medical research, informed consent rules prevent things from being funded, even when it's doubtful informed consent is really possible for more than a sliver of the population and the potential benefit may be high.

It's hard to see an organizing principle, other than maybe that you need to look at the quality of the decision and the effort that went into making it, the implications for the persons you're making it for, and the costs and benefits, when judging morality. There's a sliding scale and it probably changes with time.

Alex writes:

Great post. I recently read Oster's book, and put the question about drinking to a friend of mine, who is a midwife. Her position, which I think is common, is that "women can't handle the truth". In other words, if you concede that one drink is okay, some women may decide it's okay to drink more than that. I don't agree at all with that argument - but it may reflect the underlying views of the medical establishment.

Andrew writes:

My biological intuition suggests that an ounce of alcohol would be harmful when applied to a developing brain that weighs not much more than an ounce.

Is there reason to doubt that such exposure is harmful, or does the mother's metabolism protect the fetus from exposure?

Swimmy writes:

@Silas Barta, the post is "Should I Take to Drink?"

MingoV writes:

Physicians knew for decades that moderate consumption of alcohol was a net plus for most adults. The major medical societies continued to recommend abstinence. They were afraid of lawsuits from people who didn't understand 'moderate'.

The same is true for alcohol use during pregnancy.

The same is true for dietary cholesterol and fats.

They've known that caffeine use is harmless or beneficial for decades.

Most physicians believe they are the nannies in a nanny state.

Note: I've been a pathologist for 26 years.

James Schneider writes:

@David Friedman I remember your post as well. I'm a big fan of your work.

@ZC I'm super cautious so I am admittedly happy that my wife didn't drink. However, I think Oster is right about the facts and I think there is a cost to piling too many rules on mothers. Not only are mothers people too, but onerous rules might cause some mothers not to give birth in the first place.

James Schneider writes:

@Andrew there are a couple steps before the alcohol reaches the fetus and the mother has a chance to metabolize alcohol. this, of course, is a slow process, so in addition to limiting alcohol intake, it is also important to drink very slowly

Tracy W writes:

ZC: question for you - what makes you think that James could stop her?

As for not doing research on moderate drinking: there's been quite a bit of research done, which Emily Oster cites. It indicates that it's entirely possible there might be mild benefits for the children's future health.

What I find depressing is the willingness of so many medical types to lie based on such flimsy arguments.

ZC writes:


"I think there is a cost to piling too many rules on mothers. Not only are mothers people too, but onerous rules might cause some mothers not to give birth in the first place."

Hahaha! You really think this is an issue of 'onerous rules might cause some mothers not to give birth in the first place?' That's rich. Sure, you can make the intellectual debate that at the margin some mother won't want to get pregnant because she won't be able to get wasted every night, but such an argument carries little weight in the real world.

First off, these aren't rules, they're recommendations. There is no authority to hold anyone responsible for not following them.

Secondly, all you need to look at is the number of mothers who give birth to multiple children over successive pregnancies who are affected by FAS, suffer the manifold consequences of maternal smoking during pregnancy, or have to go through withdrawal from crack/meth/cocaine or other drugs in their system to see that, if seeing the awful consequences of one child doesn't discourage a mother from making more choices during pregnancy, a recommendation from an 'authority' who has no real authority over her life certainly won't carry any weight.

The simplified advice isn't for the woman with a post-graduate education who can understand the data (or lack thereof) and make a decision based on her own understand of the risk-benefit analysis...it's for the millions of women a year in this country who are marginally literate, don't have the desire or background to weight the varied evidence for or against the question at hand, and just need a quick, easy to understand soundbyte that relates what is the safest possible generalizeable advice for their situation.

Caplan touted that you, "know the on-the-ground details of how the U.S. health industry actually works." So, 'Medical Deception' makes for an eye-grabbing title, but seems to be a bit deceptive itself. Your fundamental issue seems to be 'broad national recommendation medical advice is too broad and simplistic'. My question for you would be:

How would you best relate the risks of alcohol consumption to one of the 100's of thousands of unwed teen mothers who will get pregnant this year?

Oh, and FYI, fetal alcohol syndrome is the most common known non-genetic (non-inherited) cause of mental retardation in the United States. About 40,000 babies are born with FAS each year. This sort of flies in the face of @Daniel Artz snarky comment that 'the Medical Establishment treats us all as unthinking boobs, complete morons, children who need to be protected. Why not?'

Silas Barta writes:

@Swimmy: Thanks, that was it! Odd that it didn't come up in a search for "alcohol doctor".

Lee Kelly writes:

The English have a penchant for understatement, and they develop particular habits of speech around it. This can get them into trouble with foreigners, who are unaware of the implicit rules of English communication. They may not understand that when an English person describes something as 'a bit of a nuisance', they may actually mean that it's pretty much intolerable. However, English people know this about each other, and so they pick up on the understatement and aren't misled.

Now, it seems to me that, especially in the United States, it's a cultural presupposition, an implicit norm of communication, that any kind of authoritative organisation will exaggerate dangers and risks. Therefore, people have learned to interpret them with a grain of salt, to assume that the authority is leaving themselves a large buffer between what they recommend and what is actually hazardous by common standards.

Perhaps this situation evolved because of concerns about legal liability. Whatever the case, once a norm like this gets established, then doctors have to deliver their advice in such a way that it will be correctly interpreted, and that means they only appear to exaggerate the dangers. Given the norms governing the communication, they would be misunderstood if they didn't exaggerate, just like one English person might mislead another by not using understatement when it would be socially appropriate.

AS writes:

You can try to rationalize deception, but it's still deception. If you do rationalize deception, you probably understated the costs and overstated the benefits. This is where cost-benefit analysis fails and principled analysis is better. Here's a simple principle: Truth is better than deception. In the long run, people adjust their expectations in response to deception and the outcome is worse than if truth was told from the beginning.

James Schneider writes:

@ZC Either the costs or the benefits have led people to having fewer kids and it isn't that kids are less cute than they use to be. This rule will have a marginal impact on birthrates but I think telling the truth will have an even smaller impact on health outcomes.

Even if you disagree with Oster, telling the truth is easy. "We know that excessive alcohol is extremely harmful to the fetus. We don't know how much is too much, so we think it is better not to have any alcohol at all."

brenb writes:

There are new research methods that can identify the causal effects of even small amounts of alcohol in utero:


bottom line: don't drink when pregnant.

Comments for this entry have been closed
Return to top