Emily Oster’s Expecting Better, Part II: Prenatal Diagnostic Testing, not new, outdated, & incomplete

ExpectingBetterIn her book, Expecting Better, Emily Oster, an economics professor at the University of Chicago, writes about lessons learned while pregnant that run counter to the conventional wisdom about prenatal care. One lesson concerns prenatal genetic testing, specifically invasive diagnostic testing. But Oster’s lesson is not news, is outdated, and is incomplete.

As shared in yesterday’s post, Oster crunched the numbers, accepted prenatal screening, and came to regret that choice because of the lingering concerns about false positives and false negatives. To know for certain, Oster concluded that if she became pregnant again, she would skip prenatal screening and have chorionic villus sampling, CVS, the invasive prenatal diagnostic test that is performed in the first trimester.

Oster bases her decision on her conclusion that CVS’ risk of miscarriage is likely comparable to that of amniocentesis (performed in the second trimester), and that that risk is likely closer to 1-in-800 versus the typically quoted 1% chance for miscarriage. This, though, is …

Not news

Some of Oster’s research includes the same studies the American Congress of Obstetricians and Gynecologists (ACOG) relied on in 2007 to change the standard of care.

Prior to 2007, only women over the age of 35 were recommended to be offered invasive diagnostic testing, because the miscarriage rate was believed to be between 1-2% for amnio and CVS, respectively. More recent studies, however, had shown the miscarriage rate for both procedures was lower when performed at experienced centers by experienced practitioners–as Oster states

The more practice, the better they get.

These recommendations were further informed by the American College for Medical Genetics and Genomics (ACMG) recommending in 2009 that all women be offered invasive diagnostic testing, first, and then, if they decline, offered prenatal screening.

Further, Oster’s chapter on prenatal testing is …


Oster’s chapter focuses on what are now considered the “old” screening tests: nuchal translucency and quad testing. This is because since October 2011, non-invasive prenatal screening (NIPS) has been available in the United States. And, NIPS is changing the way prenatal testing is done in America.

To be fair, Oster was pregnant during 2010 and delivered her daughter before NIPS entered the marketplace. But, the book was just published. So, she mentions NIPS at the conclusion of the chapter, predicting that “it is likely [NIPS] will replace the noninvasive screening.” But Oster does not take into account the impact NIPS is having on invasive testing.

As covered in other posts, studies are finding that women accepting NIPS often are not electing to have invasive diagnostic testing–regardless of the result. This is causing a plummeting in the number of invasive procedures performed after NIPS. This is why Oster’s conclusion that CVS is “clearly the better option” is …


Oster bases her book on her economics training for what is “the correct way” for making a decision: getting the data and weighing the pluses and minuses based on personal values. But her data on CVS to conclude it is “transparently a better option than amniocentesis” is incomplete.

Oster relies on one study that found over a 20-year period that the risk of miscarriage with CVS decreased to become equivalent to that of amnio. Then, she relies on another study that found the risk of miscarriage from CVS was actually less than compared to women who did not have a CVS. Therefore, Oster concludes that CVS’ risk of miscarriage is similar to amnio and it has a much-lower-than-reported miscarriage rate.

The only thing is is that the first study was from a single center in California, the second study was from a hospital in St. Louis, and where was Oster receiving her prenatal care? Chicago.

The only relevant miscarriage rate is that of the practitioner and the facility where the procedure is performed. It does not matter if over 20 years, a single center in California reduced its miscarriage rate if you’re having a CVS performed in Cincinnati. Similarly, the miscarriage rate for a hospital in St. Louis is not the same as that for a hospital in Seattle.

As Oster said, “with practice  the better they get.” But prior to ACOG changing its recommendations, another study found that only 2% of practitioners performed CVS. Couple that low number of practitioners with the plummeting numbers of invasive testing performed due to the impact of NIPS, and it is unknown how much practice the physician and facility has where any given patient is having a CVS performed. 

Lastly, Oster leaves out some data that others may think relevant about CVS.

CVS is a sample of cells from the placenta–not the fetus itself, like amnio. There are instances of placental mosaicism, where the placenta has an extra chromosome, but the fetus does not. Further, there are also instances where the CVS tested placental cells of the mother, not the developing fetus.

It would seem that all of this is relevant data that others may consider in weighing the pluses and minuses and may differ with Oster’s view that “CVS is clearly the better option.”

Oster’s advice is sound: patients should gather the relevant data and then weigh the pluses and minuses. But on the issue of invasive testing, which subjects an expectant mother to the still unknown risk of miscarriage specific to her practitioner, Oster’s choice to start with diagnostic testing rather than screening testing is not new, is outdated in the wake of NIPS, and is based on incomplete data.


  1. […] She will also be counseled about invasive testing, but be told what the specific risk of loss is based on the actual practitioner’s and hospital’s experience, not based on a national number based on outcomes at hospitals and providers she will never see. […]

Leave a Reply

%d bloggers like this: