Dr Gyamfi-Bannerman and her colleagues report data that illustrate the problem of late preterm birth and point the way to an answer. We in obstetrics have a problem with the indications for choosing preterm birth over continuing complicated pregnancies >34 weeks. Our neonatal colleagues wonder why we didn’t continue pregnancies complicated by placenta previa, oligohydramnios, or preterm premature rupture of membranes for another week when the nonstress test was reassuring. Largely unaware of the problem of stillbirth, they ask, “Why not wait ‘til 36 weeks? Why now?” We can rightfully point to the declining rate of stillbirth as evidence that we are saving fetal lives by choosing iatrogenic birth over prolongation of pregnancy, but with rare exceptions, we lack data to support decisions for indicated late preterm birth for specific indications and gestational ages.

Gyamfi-Bannerman et al present carefully analyzed data from 2 well-maintained databases that clearly state the problem. Among almost 2700 late preterm births, one-third were iatrogenic. Of these, more than half lacked an evidence-based indication. Reasons listed for indicated late preterm birth in stable patients without supporting literature were chronic hypertension, mild preeclampsia, gestational hypertension, oligohydramnios, fetal growth restriction with normal testing, and prior uterine incisions other than for low transverse cesarean birth. Infants born to women with these indications were more likely to suffer morbidity related to their age at birth, morbidity that might in some cases be acceptable, but without more data, we cannot quantify and compare the risks.

Will studies like this one give us the answer? I don’t think so. We need better data than what these authors describe and a new attitude about how to interpret these data. Specifically, instead of retrospective ascertainment from the obstetrical record of the indications for iatrogenic late preterm birth, we need a prospectively planned collection of information about the reasons for choosing delivery, with data fields prospectively defined and confirmed by direct query when memories are fresh. As an American Journal of Obstetrics and Gynecology editor, I’ve become attuned to 2 kinds of “prospectively collected” datasets, both of which are really after-the-fact and thus retrospective. In the most common form, researchers collect prospectively defined data with the intent to address multiple hypotheses as their service volume grows. Done well, as reported by these authors, this approach can produce very good information about the nature and extent of many problems in perinatal care, but often includes incomplete data. There is another, less commonly reported method of prospectively designed, retrospective data collection that is lacking in the discussions of late preterm birth. This type occurs when someone plans and funds a team of study-specific data abstractors who gather data within hours of the events described for entry into a dataset designed with predetermined hypotheses in mind. There are no empty cells in these databases; the abstractors chase the clinicians down within 24 hours of the event to fill in the blanks, and then follow the babies through their neonatal course with the same attention to detail. It’s “retrospective,” but it’s immediately retrospective and focused on specific hypotheses. A similarly sized dataset of this kind for late preterm birth would yield a new level of information about the rationale for physician-initiated late preterm birth. Until we see these kinds of data, we’ll continue to argue with one another and with our pediatric colleagues, payers, and regulators about what really drives obstetrical decision-making in these patients.

The second thing we need is a new attitude that places fetal functional maturity ahead of gestational age in the minds of obstetricians when we make these decisions. We’ve allowed gestational age to define fetal maturity. Do you expect your teenager to make mature decisions when s/he turns 16 years old, or whatever age you’ve determined to be “full term” for driving? Of course not. Why do we accept that logic when a fetus “turns” 34 or 35 or 37 weeks? More and better data about what makes a fetus functionally mature, together with more and better data about what drives decisions for iatrogenic late preterm birth, are what we need to choose the safest course for the fetus at risk >34 weeks. My own preference would be to bring the word “premature” back into perinatal thinking; “is this baby mature enough?” is a question worth asking when weighing the risks and benefits of delivery in a complicated pregnancy.