So the machine-learning algorithm is learning from the decisions that the providers have been making during this time period and predicting the probability that the mother needs an appropriate, or a medical necessary, C-section.

So that’s what we do when we say that we do control for observable medical risk factor for the appropriateness of having a C-section, so even after adjusting for differences among Black and white in their observable medical risk factors, even after adjusting for the hospital that they delivered and this—their socioeconomic characteristic, what we find is from this 25 percent racial gap, controlling for these factors only reduce the gap to 20 percent.

And this remaining 20 percent is explained—or it’s (likely) driven by provider discretion.

So basically if patients who were essentially identical in terms of their medical history and needs, but one was a white patient and one was Black, went to the same hospital, you found that the Black patient is more likely to be given a C-section.

Yeah, and even if they see the same physician.

Wow, yeah. And what are some other ways that pregnant people of color might be receiving different care than their white counterparts?

We don’t have the data to answer exactly what’s driving the doctor decision.

What we can say is that it’s not likely due to unobserved risk factors.

We do not completely show the full picture that the doctor is seeing at the moment of making that decision.

But what we show is that if we see what happened with the racial gap when the operating room is busy with a scheduled C-section versus when it’s empty, what you find is that this racial gap, it’s only present when the operating room is empty.

And what we think that this is showing is that if Black mothers were truly better candidates—and we are not able to capture this through the medical discharge records—then we should see that there is a gap regardless, if there is empty or if busy, but what we are seeing is that it’s not likely due to unobserved risk factors.

Now, if this is coming from a lack of patience from the part of the doctor, a difference in communication styles and cultural difference or if this coming from doctor’s perception of risk, that it differs between Black or white mothers, but we don’t have enough data to—or the right data to answer that part.

So when we talk about patience, are we essentially saying that physicians are perhaps quicker to say, you know, “This labor is going too slowly; I would recommend a C-section,” for a Black patient?

That could be one possibility. Again, we are not able to say from the data what is exactly going on.

Another possibility could be we are all reading the (Centers for Disease Control and Prevention) reports on infant—or Black infant mortality ...

And we have also evidence from previous papers that maternity wards do vary along their perception of risk and how they act on it;

so there are some maternity wards that are more reactive—like, they wait ’til there is a complication, and they act on it—and these type of maternity wards, they have a lower C-section rate versus maternity wards that are more, like, proactive—like, “I want to avoid a complication”—and this type of behavior tends to result in higher C-section.