The Scientific Controversy That’s Tearing Families Apart

Subscribe here: Apple Podcasts | Spotify | YouTube | Overcast | Pocket Casts

In 1971, a British doctor was trying to puzzle out a mystery: How can a child with no signs of external trauma or injury present with bleeding between the skull and brain? That doctor, A. Norman Guthkelch was part of a wave of physicians and researchers newly concerned that an epidemic of severe child abuse had been passing, undetected, beneath doctors’ noses.

As one law-review article recounts, “Prior to the 1960s, medical schools provided little or no training on child abuse, and medical texts were largely silent on the issue.” A turning point was the publication of the 1962 article “The Battered-Child Syndrome,” which urged physicians to consider that severe child abuse may be at play when children came in with injuries such as bone fractures, subdural hematomas, or bruising.

The article goes beyond offering medical advice to prescribing an ethical framework that would take hold: “The bias should be in favor of the child’s safety; everything should be done to prevent repeated trauma, and the physician should not be satisfied to return the child to an environment where even a moderate risk of repetition exists.”

Armed with these new insights, Guthkelch hypothesized that the children showing up to his hospital were being abusively shaken. Although they did not show up with the usual fractures or visible forms of physical trauma, the presence of a subdural hematoma could indicate what would come to be widely known as “shaken baby syndrome.”

Decades later, Guthkelch would publicly worry that his hypothesis had been taken too far. After reviewing the trial record and medical reports from one case in Arizona, NPR reported that he was “troubled” that the conclusion was abusive shaking when there were other potential causes. “I wouldn’t hang a cat on the evidence of shaking, as presented,” Guthkelch quipped.

The narrow claim that shaking a baby abusively can result in certain internal injuries morphed into the claim that if a set of internal injuries were present, then shaking must be the cause. On today’s episode of Good on Paper, I talk with a neuroscientist who found himself personally embroiled in this scientific and legal controversy when a caretaker was accused of shaking his child.

Cyrille Rossant is a researcher and software engineer at the International Brain Laboratory and University College London whose Ph.D. in neuroscience came in handy when he delved into the research behind shaken baby syndrome and published a textbook with Cambridge University Press on the scientific controversy that embroiled his family.

The following is a transcript of the episode:

Jerusalem Demsas: Many forms of scientific expertise in criminal-justice proceedings have been debunked or come under scrutiny in recent years. Things like bite-mark analysis and blood-spatter analysis used to be commonly understood as rigorous empirical analysis. But these questionable theories often fall apart on closer inspection.

This is how science is supposed to work. Experts observe, they hypothesize, they test, and they revise their previous understandings of the world. And in academia and in scientific journals, that’s all well and good—but what happens when evolving science is brought into the courtroom? In a courtroom, no one is well positioned to rigorously evaluate a scientific debate: not judges, not jurors, and not even the people calling expert witnesses.

[Music]

Demsas: My name’s Jerusalem Demsas. I’m a staff writer at The Atlantic, and this is Good on Paper. Today’s episode is about abusive head trauma, but you probably know it by its older name: shaken baby syndrome.

Babies cannot speak for themselves. As a result, when doctors or prosecutors accuse a parent or caregiver of having violently and abusively shaken their baby, they are often relying on something that has come to be known as “the triad.”

The triad refers to three medical findings: subdural hemorrhage, or bleeding in the area between the brain and the skull; retinal hemorrhage, or bleeding in the retina; and brain swelling. If these findings are present, according to shaken-baby-syndrome adherents, that would mean a baby has been abusively shaken. Shaken baby syndrome also indicates that since these symptoms arise rather quickly, the child must have been shaken by the last person he or she was with.

It’s important to understand that for many years, the presence of all three of these medical events was not indicative of child abuse; it was dispositive. In 2015, Kentucky’s former chief medical examiner, who had personally diagnosed SBS, told The Washington Post that “doctors, myself included, have accepted as true an unproven theory about a potential cause of brain injury in children.”

My guest today is Cyrille Rossant. He’s a researcher with a Ph.D. in neuroscience who plunged into the world of SBS when a caregiver was accused of shaking his child, an allegation she denied.

This is a very serious topic, and I want to be clear—child abuse is very real, and our public and private tools for addressing and helping children at risk are distressingly insufficient. But in their zeal to help children, many doctors, prosecutors, and scientists have allowed what one New Jersey appellate court has called “junk science” to tear apart the lives of thousands.

Let’s dive in. Cyrille, welcome to the show.

Cyrille Rossant: Thank you for having me.

Demsas: So you have a very personal connection to this issue. Can you tell us about how you first learned about shaken baby syndrome?

Rossant: Yeah, sure. So I actually lived a situation myself. So nine years ago, I had a baby who was being cared for by a caregiver. And when my baby was, like, five months old, he was sick. He was vomiting, and his head was getting bigger and bigger. So we brought him to the hospital, and they did a CT scan. And they found so-called subdural hematoma, which is blood around the brain.

And from that, they told me that it was shaken baby syndrome. It could be nothing else. So that meant that my baby had been violently shaken. So obviously, it was a very difficult thing to hear, and I was really distressed by the health of my baby. So he was taken care of. He had surgery, and fortunately he was fine after that.

So now he’s a healthy 9-year-old boy. But at the time, it was very hard. And obviously, since it was a situation of suspicion of child abuse, the hospital had to report the case to authorities and to call the police. And that’s how it all started.

Demsas: When that determination was made about your son, they didn’t leave you any doubt. When they saw the subdural hematoma, they said, Without a doubt, this is shaken baby syndrome?

Rossant: Yeah, exactly. So at the time at the hospital, most doctors were really sure that it was shaken baby syndrome, that it could be nothing else. So the thing is that we had a nanny. And my son had symptoms when he was being cared for by the nanny. So she was, like, the prime suspect. But still, it was very hard for us to believe that it was possible at the time.

That being said, there was one doctor who was less certain about the diagnosis. It was actually the only doctor who was a specialist of child neurology. And he was telling us that it could be shaken baby syndrome, but it could also be something else, namely a medical condition where there’s an excess of fluid around the brain. He told us it could be a risk factor for subdural hematoma and that he was not really sure that my son had been violently shaken.

So it was a bit confusing for me, to have, like, most doctors who are really 100 percent sure that it was shaken baby syndrome and another one who is supposed to know more about these issues to be less certain. So I was really confused, and I couldn’t really live with this uncertainty, and I needed to know what had happened to my baby.

Demsas: And what happened with the caregiver? Did they arrest her?

Rossant: So yeah, basically, it took them maybe six months or something. But yeah, after six months, she was put in custody because, in the meantime, there was a medical expert who looked at the case and said, Well, yes. It’s a shaken baby. So it happened when my baby was being cared for by the caregiver, so it had to be her. So she was put under custody. She was interrogated by the police. And then she was being prosecuted for four years.

Demsas: Oh my God.

Rossant: And in this longer legal process, another expert looked at the case, and he did not really agree with the first one.

He said, like the child-neurology specialist at the hospital, that it could be a medical condition and that maybe it was not shaken baby syndrome. So there were two different opinions in terms of medical experts. And on this uncertainty, the judge decided to drop all charges, and the nanny was cleared after four years.

Demsas: This, I think, really underscores how serious of an issue this is. I mean, when a doctor or a scientific expert tells a court or a public-policy official or a policy maker that they’re certain about what something means, it sounds like they’re talking about a natural law or they’re talking about physics. And there’s often not the ability for public-policy makers or lawyers or judges to independently evaluate the research.

So as a result of your experience, you dove headfirst into the research here and have even written a book about the finding of SBS. I’m gonna just refer to it as shaken baby syndrome, even though there’s some controversy about whether it should be called that or abusive head trauma, just because I think most people know the term shaken baby syndrome, but I want to note for listeners that there’s some controversy over the use of that term.

But there are basically three areas of controversy I want us to explore: the mechanism of injury, the diagnostic reliability, and the evidence quality. Let’s start with the mechanism of injury. The fundamental question here is whether shaking your baby is the only way to cause the classic findings of SBS. Can you lay out the scientific debate here over that question?

Rossant: Yeah, so actually, I think you can say there are two different questions here. First is: Can shaking cause the injuries that are typically associated with SBS? And second: Are these medical findings always caused by shaking? So it’s kind of two inverse relationships—the causal link, and is it the only cause? If it’s a cause, is it the only one, right?

Demsas: Yeah. The way I’ve been thinking about it is: Can punching a wall create a hole? Versus: If we find a hole in the wall, does that prove that someone punched it?

Rossant: Yeah, exactly.

Demsas: I think we can start with, like, can punching a wall create a hole? Can shaking a baby lead to the injuries of the classic findings of SBS?

Rossant: So the short answer for this is: We don’t know. We cannot shake babies for science, right? It’s not working like that. So we don’t know today. I really looked a lot for that question in the literature, and I could not find good, reliable scientific evidence linking shaking without impact—and that’s important—to the medical findings associated with SBS: so typically, subdural hemorrhage, retinal hemorrhage, and brain swelling.

What’s for sure is that babies are shaken. It exists. There’s no doubt about it. We know that many babies who are shaken—they are shaken not in the most violent possible way. There’s a whole spectrum in the degree of violence you can inflict to a baby. And when it’s not so much violence, it’s still child abuse, obviously, but it might not be enough for the baby to be injured. Okay?

So we have some interesting data from a lot of countries who asked a lot of parents and caregivers, Well, do you shake your baby? Do you hit your baby? Do you slap your baby? A lot of types of child abuse. And so it’s really just, like, self-admitted abuse, right? And parents actually say yes. And a proportion of them admit doing these kinds of things to their babies, including shaking.

So it’s something that exists. And still, we can think that 2 percent is not that much, but it’s still, like, 100 times more than the number of shaken baby syndrome diagnoses. So it means that we miss most babies who are really shaken, but maybe they are not shaken that violently, because they don’t appear to be injured. They don’t go to the hospital, and they are not detected. That being said, sometimes the babies are shaken so violently that they are being injured, and it’s hard to imagine how a baby could be shaken in the most violent possible way without the brain being injured.

Now, the exact type of injury you are going to find around the brain, etcetera, we don’t really know. We don’t know. We can imagine that it’s going to harm the brain, but we don’t know how, exactly. There is some data that was obtained on animals. So there’s a whole literature on animal studies, like mice, rats, piglets, lambs, who are shaken for science—it’s horrific when you think about it, but these are things that are done by researchers. So animals are shaken, and they are injured. And we find some injuries in the brain, but they do not really look like what you find in shaken baby syndrome. It’s not exactly the same kind of findings. It doesn’t really match.

Demsas: What’s the difference?

Rossant: It’s really technical, but you will find some injuries in the brain itself, like traumatic lesions to the neurons and to the cables between the neurons. You might find some bleeding, but again, it’s not the type of bleeding you find in shaken baby syndrome, which is really specific.

Same for the retinal hemorrhage—you will find very severe retinal hemorrhaging in shaken baby syndrome. And this is not typically what you find in animals who are shaken. It doesn’t really match. So the way researchers are interpreting this typically is to say, Well, these animals are not good animal models for human babies, and human babies seem to have specificities for the kind of injuries we find. So far, we are not able to prove the causal link between violent shaking and the classic medical findings of shaken baby syndrome.

Demsas: So models have failed to show that shaking can generate enough force to cause those injuries, and studies that are on animals haven’t been able to reproduce the classic findings through shaking alone. But there’s still a lot of uncertainty, right? So it’s still possible that shaking your baby could result in the classic presentation of shaken baby syndrome. But it’s also possible that it might be something else.

So we’ve talked about, can punching a wall create a hole? Now I want to ask you about, if you find a hole in the wall, does that prove someone punched it? Because I think one of the core parts of this controversy is that it’s not just that when a baby presents with these injuries that doctors will say, It’s possible this baby was abused. There’s been a training of doctors to indicate certainty, that if you find this—it’s often called “the triad”—if you find this triad of injuries, then you should presume that the baby was shaken and that the baby was shaken abusively by the last person who was with the child when it began presenting with those symptoms. So why was that the medical consensus? Why did that training happen?

Rossant: Yeah. I think you’re right to say that it’s not just a cause; it’s the only cause. That’s the theory, right? Shaking is a possible explanation for the findings. It’s the only possible explanation, and it occurred just before the baby showed symptoms. So it’s a really, really strong theory that has been taught to doctors. Now, why has that been the case? I don’t know.

What I know is that historically, and you really need to dig down into the history of shaken baby syndrome to understand how it was born—it was born in the ’70s with this hypothesis that maybe shaking was one of the possible causes of subdural hematoma, but at the time it was just a hypothesis. And today it’s still a hypothesis. And that was in the beginning of the ’70s. And a few years later, doctors in the U.S. started to presume abuse whenever they found subdural and retinal hemorrhage in infants. Why that has been the case, why this mere hypothesis, Maybe it’s shaking, was transformed into, It is shaking, and we need to call the police, that I don’t know. But by the ’80s, you start to see prosecutions based on this theory.

So somehow—I don’t know—some doctors in the U.S. started to do this. It’s important, also, to say that there was the context of reporting any suspicion of child abuse to authorities. That started in the ’60s. There was a big subject here in pediatrics at the time, because before the ’60s, it was not really obvious for doctors to think about abuse. They didn’t really think about it. And it suddenly changed in the ’60s. And they really realized that it was important for the medical community, and especially pediatricians, who see children all the time, to think about abuse whenever they find suspicious findings, like fractures, bruises, and subdural hematoma. That came in the ’70s.

So there was this big push for doctors to really report as many children as possible to the authorities whenever they have the slightest doubt on abuse. And since there was this hypothesis that maybe subdural hematoma is one of the signs that should make doctors suspicious of abuse, well, they started to call the police and to report these cases to authorities.

That being said, I think there’s a difference between being just aware that maybe children are abused and calling the police, and going to court and saying, This is abuse, and this is nothing else, and this is certain. This is a big difference to me. And this is really what I don’t understand, because, to me, doctors should treat patients and report possible child-abuse cases to authorities. But going beyond that and saying to the courts, I’m a doctor, and I know that this child was abused, even though there’s no other piece of evidence apart from the medical findings, this is the thing that is going too far for me.

Demsas: What other things can happen that can cause these symptoms to present in babies? Are there other potential explanations that you found when researching this?

Rossant: Yeah. So actually, now we know a lot of possible causes of subdural and retinal hemorrhage. First, there’s everything involving accidental trauma, like short falls and domestic accidents. Whenever there’s an impact to the head, even what appears to be a small impact, it can really cause severe injuries, including what we see in shaken baby syndrome. There’s also biomechanical data about this. Impact is really dangerous for the skull and for the brain in a child.

Then you have many rare diseases, like a genetic, metabolic, neurological conditions that can all cause subdural hematoma or be a risk factor for the development of subdural hematoma after a minor impact, which really happens in most babies. Once they start to sit down, they can fall, and when they try to get up, to stand up, they can also fall. So it’s really, really common for babies to hit their head. Most of the time, it’s not going to cause anything, but if there is a medical condition, if there is a risk factor, then it might cause the findings of shaken baby syndrome.

There are infections that can cause blood-clotting disorders. There’s really a lot of things. It’s really complicated. It’s really not possible to say that only shaking is the only cause of the SBS findings. There are also risk factors—again, it can be medical conditions.

But it can be just a prematurity. It’s a big risk factor. These babies are much more fragile. Babies who have a low birth weight, babies who have a large head—there are many little things like this that can increase the risk of a subdural retinal hemorrhage after minor impact. So yeah, it’s really complicated.

Demsas: I want to talk about some of the pushback that people like you have received from other scientists who stand by the shaken-baby-syndrome diagnosis and say that it is perfectly reasonable for doctors to presume, and for the courts to presume, that when these injuries present in babies, that we should assume a caregiver has shaken the child. What they often point to is the fact that numerous perpetrators have confessed and admitted to shaking their babies, and that the confessions often will provide detailed accounts that match the medical findings of shaken baby syndromes, and that they have these consistent patterns.

Why is that not convincing to you that, perhaps, maybe it’s the case that science hasn’t figured out exactly how shaking will cause these symptoms, but if people are admitting to having shaken their baby and then their baby is presenting with these symptoms, that’s a reasonable cause and effect to presume?

Rossant: Yeah, it’s true that confessions—today, it’s the main piece of evidence for shaken baby syndrome. The question is not whether confessions exist or not; it’s how reliable they are, and what you can learn from the confessions. So confessions do exist. And I also want to stress that, obviously, some of these confessions are true, and that some parents do abuse the babies, and they end up confessing when they are being interrogated by the police.

So yeah, you cannot possibly say that all confessions are false. It’s not working like this. That being said, I’ve studied this question a lot in the scientific literature, and it’s true to say that they are really unreliable scientifically. They are not scientific evidence. And it’s not obvious to realize that, because when you don’t know the subject, you think, Well, if the person admits something that horrible, it must be true. There’s no reason for an innocent person to say they did it if they did not.

But it’s much more complicated than that. The topic of false confessions in general—not just for a shaken baby, but for any type of crime—has been known for decades. You know the Innocence Project that was able to exonerate a lot of people based on DNA evidence? Well, it turns out that between 25 and 30 percent of all the DNA exonerations had falsely confessed, and they were factually innocent, because that was proved by DNA, which could point to someone else. So it’s something that exists.

Now, you wonder why innocent people confess. So there are many reasons. There’s even a classification that was done by a psychologist and scientist. There’s a psychologist called Saul Kassin, who is an expert of this, and he has devised a classification of false confessions.

So for example, one of the reasons is just plea deals. So they plead guilty, and they might not go to jail, and they can walk free, but they have to say they did it. But even without that, even in the police interrogation room, innocent people can be led to confess what they did not do, in this context.

For example, the police can say, Well, if you confess, the child, who is in foster care, can go back to you. You can get back your child if you confess. Or maybe, If you confess, your child can go back to the other parent. There are many incentives that are given by the police to the persons to confess. And the reason for that is that the police are being taught, like medical doctors, that shaking is the only possible cause and that it has to be that last person with the baby at the time of the collapse, right?

There’s also the whole thing of internalized false confessions. It’s really mysterious. It’s really a psychological effect that can lead innocent people who deny having harmed their child while being interrogated to slowly, in the course of hours or even days of interrogation in a very stressful context, to believe that maybe they did something that they did not think was harmful but was actually harmful, or at least that’s what they are led to believe. And it really happens a lot in this particular type of case. The police can make a lot of scenarios: Okay, maybe you didn’t want to harm your child, but you took the child from the bed a bit quickly, or maybe the head hit something, or you were not careful enough.

There’s a lot of scenarios that are being fed to the suspect. And in this specific stressful context, innocent parents who are really—they have this guilt of maybe they think they did not do everything they could to save their children. Maybe they tried to resuscitate their child with cardiac massage or something, or maybe they tried to slowly shake the baby, but it’s a very mild type of shaking. It’s really not a violent shaking. So you can always find something.

And this really happens, and I’ve seen it a lot, and it’s been documented on videotaped police interrogation. You can see it in some transcripts. So when you really dig into the details of what has been confessed exactly and how it has been confessed, and when you look at all the context of the interrogation, you realize that it’s really not, Okay. Well, okay. I admit it. I just lost my nerves and violently shook my baby. This is very rare. This is not the type of confession you see in shaken baby syndrome. It’s: Oh my God. Okay. Maybe I did it. I didn’t realize my baby was collapsing. And yeah, I tried to revive my child. And maybe in the process, I harmed my child, and I—oh my God.

It’s a really complex phenomenon. But it’s been documented, and I think it happens a lot. So you cannot just say, Okay, confessions exist, so it must be true, right? That being said, most confessions are obtained after the diagnosis—so typically, after the police have been called, and it’s most often in the police interrogation room, right?

So there’s this big contamination, and it’s more than that. It’s really a pressure, because the police think they know that the baby was shaken and that this person is guilty, so there’s a whole bias, right? What would be interesting would be to see if confessions can be obtained before—before the police interrogation and even before the medical exam that will show the medical findings that are associated with SBS. Because if it’s obtained before, then there’s no way the diagnosis could influence the confession. Obviously, this kind of confession is very rare, but there is at least one study that was able to find something like 36 cases where the confessions were obtained before any kind of medical exam.

So there’s a lot of reason to believe that these kinds of confessions are genuine, true, right? And what’s interesting is that, in those cases, you are not going to find the medical findings of SBS. So yeah, there are many reasons to be skeptical of this theory.

[Music]

Demsas: After the break: Shaken baby syndrome goes by a different name now, but the same problems persist.

[Break]

Demsas: I’m hoping you can explain why the term shaken baby syndrome has fallen out of use. Now we’re more likely to hear the term abusive head trauma, and I’d assumed that was due to the criticisms levied by folks like you who’ve become skeptical of the SBS diagnosis. But in a policy statement about the diagnosis, the American Academy of Pediatrics said that the name change to abusive head trauma “was misinterpreted by some in the legal and medical communities as an indication of some doubt in or invalidation of the diagnosis and the mechanism of shaking as a cause of injury.” And then they say that the AAP, “continues to embrace the ‘shaken baby syndrome’ diagnosis as a valid subset of the AHT diagnosis.”

So what’s your interpretation of what’s going on there?

Rossant: Yeah, so it’s true that in 2009, the AAP published a position paper stating that from now on, “abusive head trauma” should be used instead of “shaken baby syndrome.” And the way they justify this was because it was to encompass a broader source of abuse inflicted to babies—not just shaking, but also impact to the head. That’s the justification.

And there was one big study in ’87 by [A. C.] Duhaime and a few biomechanicians who really showed, first, that there were very often signs of impact, which was not really compatible with the idea that these babies were just shaken, right? And also, they did a biomechanical study to show that the forces involved with shaking are much lower than when there’s any kind of impact to the head. So impact to the head is really, really severe and implies very big forces to the head and big deformation and big energy. So it’s much more dangerous.

And there was some controversy in scientific articles about this that partly led to, I think, this decision to change the name from shaken baby syndrome to abusive head trauma, because it was not just shaking very frequently; there’s also impact.

Now, it’s true that before 2009, especially since the Louise Woodward trial in Boston in 1997, there was a lot of controversy in the media and the scientific articles on specifically shaken baby syndrome, the hypothesis that you could infer abuse just with the triad, without any sign of impact. And yeah, some people believe that this change of name is a consequence of this controversy. Obviously, the AAP, the medical institutions do not really acknowledge this, because their position has always been to say that there is no controversy at all, right?

So here we are. I mean, we have this change of name that has been more or less accepted by everyone, even though the term shaken baby syndrome is still quite widely known in the public.

There’s something else that should be pointed out. It’s that the term shaken baby syndrome was criticized—the very term—in particular by Norman Guthkelch, who first identified shaken baby syndrome, or at least the link between shaking and subdural hematoma, in 1971. He criticized the fact that the same term is used to describe both an act—shaking—and a set of injuries. So it conflates a unique hypothetical cause to objective findings. And it’s a real problem because you can’t talk about what you see without accepting that there might be other causes than shaking.

So that’s why he recommended to use another term, which was, I think, retinal-dural hemorrhage of infancy. That really just describes the fact that you find subdural and retinal hemorrhage in an infant without presuming anything about the cause. Unfortunately, that was not the choice that was made, and now we have abusive head trauma, which is also problematic because it also implies that, well, it’s abuse. So it’s a medical diagnosis that comes after you discover specific medical findings in an infant, and you give the term abusive head trauma, which implies that it was caused by abuse.

So there’s this whole thing of intent that is really not the job of medicine. It’s for the police and the justice to determine what happened and what the intent was. So the previous term was a problem with this respect, and the new one is still problematic in this respect.

Demsas: Yeah. It was really interesting when I was reading about the controversy with Norman Guthkelch, who you just mentioned. I mean, he’s called, like, the father of shaken baby syndrome because of his 1971 paper. NPR reported that he reviewed a case in Arizona, and they wrote that “he was troubled to see that the medical examiner’s autopsy had concluded that the baby died of shaken baby syndrome while discounting other possible causes.”

So, you know, given that, why do you think it’s been difficult for the medical community to become more agnostic about whether these injuries that show up in children are necessarily the result of abuse or of some other thing going on? Like, why is the AAP still saying this? I mean, I know you can’t speak about them specifically, but why do you think there’s just been such reticence from the medical community?

Rossant: Oh, that’s a very good question, and I am wondering this. I mean, it’s been almost 10 years that I ask myself this very question, and I don’t really know.

I think there’s—I mean, it’s more general than that. It’s, you know, in the human psychology, the fact that it’s very hard to recognize that you were wrong before. It’s very hard to change one’s own mind, especially when doctors have made a lot of diagnoses with very severe consequences: with removal, going into foster care, and, you know, criminal prosecutions, etcetera.

It’s very hard to accept that, as a doctor, you were wrong and you maybe did some misanalysis and you were responsible for, you know, miscarriages of justice, etcetera. I think it’s really the No. 1 thing that is blocking everything, this psychological aspects of doctors, who are humans, like everyone else.

Some doctors do change their minds with the new science, the new articles, new data, their new experience, new cases. They realize that maybe it was not as easy as we thought before, and they start to change their minds. They start to work for the defense, and then they are being targeted and bullied, attacked, especially by the establishments, and it’s always the same. So yeah, it’s really in human nature.

I think it’s also the scale. I think it’s not just a few errors here and there. It’s really, really massive. I think there are many, many cases which are misdiagnosed and that the so-called shaken babies were not really shaken. I mean, obviously, it exists, and there are many cases where the diagnosis was true, but many where it was not the case. So that’s also why it’s so hard to accept one’s own mistakes, because it's a really massive mistake that was done.

Demsas: Something you just said about how much they react to public censure here: There was a quote from this Milwaukee prosecutor. I believe this is from a ProPublica article, but he’s a deputy district attorney in Milwaukee, and he said it was, “providing reasonable doubt for sale.”

Essentially, there’s some criticisms of people who will provide reasonable doubt or arguments that provide reasonable doubt to accused criminals, whether it’s forensic evidence, like fingerprint analysis or DNA analysis or things like the SBS. But there’s been a real backlash from within the law-enforcement community to scientific evidence being muddled in courtrooms. And I wonder: Have you come across people who react to your work and believe you’re giving cover to child abusers? What do you say to them?

Rossant: Yeah, I think I kind of understand because there’s always this tension between, you know, being too safe on the safe side—I mean, is it better to put an innocent person in jail, or to let a guilty person walk free and potentially harm children, right?

And personally as someone who believes in democracy and, you know, les lumières, which is a really French thing. I think it’s really important not to harm innocent people. So it’s really kind of philosophical attitude, I think. But yeah, I can understand why some people believe that it’s not possible to take this risk of letting potentially dangerous people walk free.

But you know, in the end, I think we should all try to do our best and try to be as accurate and scientific as possible. Try to look at all the evidence on the one side or the other side, and then let the criminal-justice system do its work. And that’s why we have this notion of “beyond a reasonable doubt,” you know, the burden of proof, all of these things. They were designed precisely to avoid, as much as possible, putting innocents behind bars, which is a very terrible thing to do as a democracy, I think.

Demsas: Do you have thoughts on how you would’ve preferred to have things play out in your case? When a doctor is concerned about SBS with your child, what do you wish had happened?

Rossant: I think most parents, most families would accept something, some kind of measure that is not removing the child. So, you know, a follow-up with the psychologists, social workers, people that go to your place and that look at the room, how you handle your child, that follow you from a few months, maybe one year, I don’t know. It depends on the situation, but that is fine.

And we actually had that. Even though the nanny was being prosecuted after the first month, we did have a follow-up with a psychologist and social workers for, like, two years, I think. It was very light. It could have been, you know, more intrusive, and it would’ve been fine. I mean, there’s no problem as long as you have the freedom to have your child with you, right?

It’s really the fact that to remove a child from its environment, it’s a really big cost, so that should really be done in the most extreme cases. And typically, some judges do say that it only happens in the most extreme cases, and that they tried to do their best not to go that far and to find all possible solutions before resorting, as a last solution, to foster care.

But in practice, in the cases I know—and especially in the abusive-head-trauma cases—for very, very small children, babies who don’t speak, it’s very often removal into a foster family in a nursery. And that’s really, really hard for the child, and for the parents to know that the child is suffering from being suddenly put into a different place and without parents, without siblings, without the teddy bears and, you know.

Demsas: Yeah, I guess it depends on where you are and how the authorities function. But, you know, I could imagine that most parents would be okay with having an interview with Child Protective Services if, you know, it was respectful and there’s a clear protocol for what was being followed.

And I think there are a lot of civil servants who take their jobs seriously and want to make families better. But I think it’s very variable, especially in the United States, where this is not a federalized system. There’s very many different administrations of child protective services. And where you are can vary very differently, how you’re treated and how you interact with and how the state interacts with you.

And so, I agree—we would want to create a system where parents felt fine and open and welcome to that kind of surveillance and interaction, but I worry that we don’t actually have that in the entire world.

Rossant: Yeah. And it’s actually the same in France. Today there’s no centralized child-protection system. It’s each department, each region of France that has its own system. So there are great disparities between the different regions, and we do see very, very different treatments of similar situations, depending on where you are in the same country. So yeah, it’s a very difficult problem, and some families do not understand why it’s worse in their own case compared to other families.

Demsas: This was obviously a very serious episode, but I always like to end on a question that draws people to think about and reflect on a time when they themselves have believed something that didn’t turn out to be true. So in your life, is there a time where you believed something that ended up being only good on paper?

Rossant: Wow. Good question. I really never thought about this. But maybe I can say that initially—and sorry; it’s still about shaken baby syndrome, but initially—I believed that shaken baby syndrome was a thing.

Why? Because my own father was a pediatrician and an expert in court, and he told me many times about shaken baby syndrome, and he did testify for the prosecution in shaken-baby cases. So I knew it.

Demsas: Wow.

Rossant: Yeah, it’s really a crazy coincidence, right? Sadly, he passed away one month before the symptoms of my child. So just before that. He could have been very helpful, obviously. But that’s why I had this bias at the beginning, when I was starting to look into the literature. I was sure that shaken baby syndrome was entirely legitimate, and I thought there was no controversy.

So when I started looking into it, I really didn’t think that I would change my mind, but I actually changed my mind. It was really hard for me to do it, because I was so convinced that the scientific consensus was right. And even my own father was testifying in court that this theory was correct. So it was not easy for me to change my mind, but I had to, because that’s what the evidence was telling me.

Demsas: Well, that’s a model for what we’re trying to do on this show. Cyrille, thank you so much for coming on.

Rossant: Thank you. Thank you a lot.

[Music]

Demsas: Good on Paper is produced by Rosie Hughes. It was edited by Dave Shaw and fact-checked by Ena Alvarado. Erica Huang and Rob Smierciak engineered this episode. Rob Smierciak also composed our theme music. Claudine Ebeid is the executive producer of Atlantic audio. Andrea Valdez is our managing editor.

And hey, if you like what you’re hearing, please leave us a rating and review on Apple Podcasts.

I’m Jerusalem Demsas, and we’ll see you next week.