mybetterdoctor

I know why the hoopla around Emily Oster‘s new book, Expecting Better, bothers me, it’s because she hasn’t let you in on obstetrics’ big bad secret… and here it is…. from a pulmonologist.

What Ms. Oster points out, and where she is absolutely correct, is that the data to support most pregnancy advice, indeed, most medical advice, are poor and sorely lacking.  By most estimates, modern medicine is supported by any kind of evidence, at best, 10-20% of the time.  And when you delve into the evidence, for the most part, it is of moderate or poor quality.  Most physicians are acutely aware of this – this is the space we work in.  When I was interviewing for medical school, one of the interviewers told me that “half of what we teach you in medical school is wrong, the problem is, we don’t know which half.”  Medicine, and science, is full of stories about how treatments that were previously standard were overturned.  You don’t have to be very old to have seen this evolution for yourself.  As recently as a decade ago, post-menopausal women were routinely prescribed estrogen for “cardiac protection,” and now hormone replacement therapy has become harder to come by than crack.  The observational studies supporting estrogen were so strong that many debated whether doing a randomized, controlled trial of the practice – the strongest statistical design to account for confounders – was even ethical.  Then we learned that hormone therapy increased the risk of strokes, heart attacks and blood clots, and we stopped prescribing it overnight.

We’re also aware that the “business” of medicine has corrupted much of what we practice as modern medicine.  Drugs that are heavily promoted resulting in huge profits for the pharmaceutical companies are withdrawn a few years later with new side effects, corruption scandals, off-label advertising, or unethical clinical studies.  The clinical studies that are done to approve the drugs are usually short-term and measure a surrogate marker – what does a 100ml improvement in a lung function test after 12-weeks of therapy mean for the patient? We want to know about mortality, or symptoms, or hospitalizations, or real outcomes that affect patients, not just their numbers on tests.  But these studies are expensive and take a long time, and so the initial approvals are often due to these “short cuts,” that provide incomplete data.  And we, too, read the headlines about scientists falsifying data – because everyone has ulterior motives.  Even doctors are not free from bias.  It’s not that we’re evil or trying to profit from your disease, it’s that we’re human.  The urologist sees people die of prostate cancer.  The primary care doctor sees healthy men become impotent with life long side effects who die of their heart attack and whose diagnosis of prostate cancer served to ruin their quality of life and increase health care costs.  At their hearts, both believe they are trying to help the patients when recommending for or against PSA testing, the screening test for prostate cancer, and until recently the recommendations were for each doctor to discuss the risks vs. benefits of the test with the individual patient.  Finally, the US Preventative Services Task Force recently recommended against screening for prostate cancer weighing the small potential benefit versus the large potential harm of screening.

The pregnancy literature, especially, is terrible.  Modern obstetrics, with all its problems and the natural movement push back, has dramatically reduced neonatal and maternal morbidity and mortality.  We may not be able to pinpoint which 50% of what we do in obstetrics is the good half, but the improvement in health with the whole package is obvious.  The effect of modern obstetrical care is so huge, it is “visible to the naked eye.”

Advice in pregnancy, however, like much other advice, is not based on strong data.  Nobody has ever done, or will do, a trial of randomly assigning pregnant women various amounts of alcohol to drink during their pregnancy and then following them and their children over years to see what the outcomes are.  In addition, the magnitude of the outcome matters.  Thalidomide had such severe teratogenicity that its terrible effect was visible without a randomized controlled trial.  On the other hand, just by the fact that fetal alcohol syndrome took decades to be described, we know the effect of alcohol on pregnancy is not as strong.  The adverse effect of alcohol is somewhere between non-zero and that of thalidomide.  For each individual pregnancy it probably depends on underlying genetic factors about the mother and baby as well as a myriad other environmental factors.  How much you drink daily, how often, what your baby’s gestational age is, other factors and other environmental exposures, tobacco, other drug use or medications.  “There’s no safe dose of alcohol,” is a fair statement in the absence of data.  In other words, we have not proven that any dose of alcohol, however small, is “safe,” but it’s also a statement of uncertainty.  It doesn’t mean an occasional glass of wine will certainly adversely affect the fetus.

And so, in alcohol, as in many other areas of uncertainty, we try to asses the likely risks vs. the likely benefits.  If you rarely drink alcohol, then pregnancy is probably not the right time to start – however small the absolute risk to your baby, it doesn’t seem worth it to increase it.  On the other hand, if you want to get a glass of champagne at your graduation or your sister’s wedding, then there’s probably not much harm that will come to your baby as a result.

Pregnancy restrictions tend to be cultural as well.  In France, women eat raw cheese with few untoward effects.  If you’ve eaten raw cheese for years, the small risk of Listeria to your baby may be worth it to avoid yet another lifestyle change.  In the US, however, where most of us consume pasteurized cheeses, trying the fancy raw cheese at the cocktail party while you’re pregnant is probably not worth it.  Japanese women eat sushi while pregnant, but for Americans, who view sushi as synonymous with maguro (tuna), switching up their fish selection at the local sushi joint may well be worth it.  (Our sushi chef in San Francisco’s amazing Ariake restaurant was the first to know that I was pregnant, I assume because I wasn’t the only sushi patron to alter her order upon finding out the news.)

Life is full of uncertainties, and our doctors are meant to help us navigate the landscape.  Babies are born and grow and I can see why we’re all frustrated with the myriad recommendations – they can feel like laying the blame on the mother for any potential problems.  “Autism?” Well, you must have had too many ultrasounds during pregnancy. Or ate too much tuna, or drank too much coffee, or exercised too much, or didn’t exercise at all.  Breastfeeding, too, was very frustrating.  When one of our kids was colicky, I heard a ton of advice about what foods I should avoid, which, if followed, would have left me with little food to eat.  The moms with formula had it no easier, it was the formula that was to blame, and then they spent weeks switching brands around to see what might work, until eventually the kids got better.  I get it.  I’m not saying we should all follow these guidelines to the letter.  Most professional societies recognize as much by not just issuing recommendations but also grading the recommendations based on the quality of evidence that they are based on.  A small, badly designed study, for example, that shows an improvement will result in a weak recommendation.  Yes, as far as we know you may be better off doing this, but the data behind it are so poor, don’t be surprised if we change our mind in a few years.

And so where does that leave us?  If it is certainty we’re seeking, it is, alas, not available for any of us in any area of our lives.  Medicine, like all human endeavors is just that, human.  And perhaps what irked Ms. Oster about her obstetrician is this lack of acknowledgement of its fallibility.  For when she fails just where she blames her obstetrician did – when she tries to issue edicts about pregnancy.  There are no randomized, controlled trials proving drinking alcohol during pregnancy is safe, any more than there are trials proving it is not.  We are all, on much surer footing criticizing the science on which the recommendations are based – it is poor, without a doubt.  It does not allow directives, just a best guess, and Ms. Oster would do well to remember her own lessons well.

When I found out I was pregnant, I was already several weeks into the pregnancy.  I called my obstetrician concerned about the wine I’d had at a recent event.  “Don’t worry about it,” my obstetrician said.  “Plenty of women drink alcohol before they know they’re pregnant.”