Interview with Dr. Ilene Claudius, Part 1 of 2: Managing Suspected Child Abuse in an Emergency Setting


Dr. Ilene Claudius

Ilene Claudius, MD, is an associate professor of clinical emergency medicine and chief of pediatric emergency medicine at the Keck School of Medicine of the University of Southern California. She is editor-in-chief of Pediatric Emergency Medicine Practice, and her clinical research interests include pediatric mass casualty, non-accidental trauma (child abuse), and pediatric mental health.

Dr. Claudius has also studied apparent life-threatening events (ALTE) and brief resolved unexplained events (BRUE). She has authored or co-authored dozens of studies in peer-reviewed journals, and is an active contributor to the EM:RAP podcast. We talked to Dr. Claudius about her research and clinical expertise in pediatric emergency medicine.

One-third of children who die from abuse were seen by a doctor at some point for injuries not recognized as abuse at the time. Medical students and residents are taught that findings like spiral fractures are obvious red flags. How should those be approached?

Certainly, spiral fractures are concerning, except for the tibia—spiral fractures of the tibia in toddlers tend to be relatively common and rarely associated with abuse. If someone has a sternal fracture, a rib fracture, a scapular fracture, a metaphyseal fracture (a tiny fracture at the very distal metaphysis, often a tiny little chip fracture that you can barely see unless you blow up the image as much as possible)—those are much more likely to be associated with abuse than your standard long bone fracture.

Also, fractures in different stages of healing, maybe a fracture that supposedly happened yesterday but has callous formation on the radiograph that suggests it is more than a week old, should be a red flag for concern. Kids under two years don’t tend to sustain a lot of fractures, and definitely kids under the age of 1 [year] very rarely sustain accidental fractures outside of major trauma, so a low threshold should be maintained for reporting abuse in this group.

But with fractures, it’s not just recognizing those that are high-risk; it’s looking at the entire situation. If somebody is coming in with a tiny child who’s barely cruising and can’t even walk unless they’re holding onto furniture, there’s no major trauma mechanism, and all of a sudden this kid has a broken humerus: that’s concerning to me whether or not it’s a spiral fracture. Because this kid isn’t more than a couple of inches from the ground at any given time, not moving very quickly, and now without a great history has a fracture of a long bone. If the severity or the type of fracture doesn’t match the story you are being told, that is cause for concern.

On the other hand, one of my colleagues once had a kid who somehow managed to get his arm caught in the washing machine, and he had a spiral fracture, but it made sense. It was one of those situations where I understand that spiral fractures are taught to be due to abuse, but here’s the perfect mechanism to get a spiral fracture that isn’t abusive. So it’s not just the type of fracture as much as it is looking at the kid and the reported mechanism and making sure the fracture matches that, in terms of timing and severity and what the cause looks like. You don’t typically fall on your arm and get a spiral fracture. But if you had a twisting mechanism, then that might be the cause.

What are some of the lesser-known or often-missed signs for clinicians to look out for?

The most common presenting feature of child abuse is a skin finding. Bruising is often the first injury that a child who has been abused presents with, but as we all know, bruising in general is very common, even in absentia of child abuse. So how do you figure out which bruises are concerning when you see toddlers with bruises left, right, and center, just from normal toddler stuff?

It is the same thought process: does the bruise match what we’re being told? Say somebody tells me that a 2-month old lying in the middle of a king-size bed rolled off. But 2-month olds don’t usually roll. So you have to give me more to the story. If you tell me that there’s another kid on the bed, that you left the the baby near the side of the bed with their 2-year old sibling, that’s a different story than if you’re trying to convince me that a 2-month old managed to roll several times and and roll off the bed completely on his own. Is the child developmentally able to do what the parent says happened? Does it make sense that the kid was moving around enough to incur the kind of bruise they’re presenting with?

We have an adage: “Kids who don’t cruise, don’t bruise.” Obviously, anyone can bruise if something traumatic happens to them; anyone can be in a car accident and have bruises. But if you’re not at least mobile enough to be cruising around, it’s very unlikely that you’re going to sustain an accidental bruise with no additional history. Because you’re just not moving with that kind of velocity, and you just don’t have the kind of force. Even if you fall down or crawl into a wall, you won’t bruise from that most of the time. So if someone isn’t particularly mobile, and there’s not a good story, then how do we know that there’s actually an accidental reason for that bruising? That’s when you have to look into abuse a little more.

The location of the bruise is also important. Let’s say you’re riding your bike and you fall on your hip. There are areas of your body that will obviously be scraped up and bruised. And there are other areas that are much less likely to get bruised. Ears don’t tend to get bruised. Concavities tend not to get bruised, like the concavities of your cheek or neck. Kids tend to not get a lot of truncal bruising, or genitalia bruising, particularly kids who are in a diaper most of the time. If someone has a bruise in those areas, it’s not that I’m picking up the phone and calling social services; I’m just asking the parent in a non-leading, non-judgmental way, “Hey, what happened? You told me this kid fell. I need some more details.” And if someone in the family provides an interesting detail about a protuberant object that they fell on, that may make sense. But sometimes that doesn’t happen, and that’s when you need to be concerned and ask social services to investigate why this particular child from that mechanism has a bruise in that location.

How does this tie into your research?

Something that we were hoping to look at, and the reason why we started studying bruises in non-accidental trauma, is that unlike fractures, the timing of bruises is very difficult to understand just by looking at them. We used to say there’s an evolution of color: they start out dark purple, then they’re green and yellow. But that doesn’t really make sense. Because if you have a big deep bruise, it’s obviously going to go through a different color evolution over a different time frame than a small superficial bruise. Having a bruise on a densely muscular part of the body is different than having a bruise on the ear. So we don’t have a great way of timing bruises.

Also, we don’t have a great way of seeing how deep it is. If you tell me that a kid was cruising and he fell, I would expect that not to be a huge bruise, because it’s not a huge fall. But when I’m looking superficially at the bruise, I don’t know how deep it goes or how old it us.

So we were looking to see if ultrasound would help us flesh out some of that information, so we could add that to our armamentarium of things we use to decide if the bruise that we’re seeing matches the mechanism. I don’t think it’s ready for primetime yet, though. We have to ultrasound a lot more kids before we say that. But it does look like it might be somewhat promising, in terms of a helpful adjunct to our evaluation.

In general, a lot of these kids present to emergency departments, and one thing that emergency providers do really well, even if we don’t know a lot about a topic, is to apply common sense to the situation. And that’s really helpful with child abuse, so I don’t think people should underestimate their ability to diagnose it, or question their decision to report it.

At the very root of things, seeing a red flag for child abuse is just having good common sense. The parents are telling me this injury happened, so does what I’m seeing on physical exam match (a) that this injury could have happened, (b) that this is from that type of injury, and (c) that it happened at the time the parents are telling me? If the answer is no to any of those questions, you owe it to the family to ask more questions to get a sense if there’s a decent explanation that you’re just not hearing, and if they don’t provide one, then you owe it to the kid to report that to social services or law enforcement.

Can you talk more about how you actually approach these families and the type of language you use in asking these questions?

It’s tough. Sometimes, an abusive family can be obstructive or try to leave the department. In those cases, obviously, you need to protect yourself, and turn it over to social services and law enforcement for your own protection and for the protection of the child.  But most of the time, when we identify potential abuse of a child, the first thing that I hear from other attendings, nurses, medical students, from everyone, is, “They seemed so nice! I had no idea!” One thing that’s good to remember is that the person that you’re talking to in the room is very possibly not the one abusing the kid, and might not even know that the child is being abused. So if you approach it like that, as in, “We’re on the same team, so let’s figure out what’s going on with your child,” that helps me to remember to be child-focused and non-judgmental.

It’s a medical problem, so I see it as that, and I take a very detailed history. Often I will take a developmental history. And if pediatricians are reading this, they’re probably saying “You don’t take a developmental history on everyone?! What’s wrong with you?”, but of course in the emergency department, we very rarely take a full developmental history. But it’s not unreasonable to take one, especially if there’s any question of whether or not the kid is capable of completing the activity they’re telling me happened.

In those cases I will take a developmental history up front, before we address any injuries, just so we know if that’s in the child’s armamentarium. When did the child start crawling? And when did they start pulling up? Are they able to cruise along the furniture? Have they taken their first steps yet? And I’ll just go through that along with medications and past medical history so I have a sense if developmentally the child’s where I think they are based on age.

I ask a lot of details about the trauma that happened, and if I really do feel like this is an abusive mechanism, I’ll frequently tell the parents, “I’m concerned, and this is not an injury we would typically see from what happened. I don’t know if it’s just really bad luck, but I know that your child is my patient and it’s my duty to protect them. And I’m sure that if anyone is hurting your child in any way whatsoever, you love your kid, a lot more than I do, to be honest, and you would want to find that out as well so that together we can work to protect your child.”

It’s hard to argue with that, to find an argument that, “I don’t want to protect my child, I don’t love them more than this random doctor that’s never met them before.“ So that’s how I approach it with parents, and typically I’m talking to the mom—it’s often not the perpetrator with whom I’m speaking. Usually they’re relatively cooperative. And sometimes people say, “Are you saying I’m hitting my child or hurting my child?” I’m not saying anything. All I’m saying is that this is a really bad injury, from what happened, and I want to make sure there’s nothing else going on. If nothing else, this is a bad enough injury that I’m wondering what other injuries the child might have sustained, because I never would have imagined they would have sustained this from falling off the bed, or whatever they told me happened. And so I’d like to do an x-ray of all the bones in the body, to make sure we’re not missing any other trauma, because this is a huge injury, and I don’t know why that happened. You can usually get by with a skeletal survey from that.

What other tips do you have for talking to families in cases of suspected abuse?

One thing to watch for is, I never give the patient ideas. I was very upset recently, when we were interviewing a family, very non-judgmental, and there were five of us who independently interviewed the family and were given no mechanism to explain a fairly major injury. Finally, the social services person came out and said that now Mom remembers something else that happened. I was very concerned that someone had gone in the room and said “Did this happen? Did that happen?” If you give somebody a list of reasonable choices of how their child could have been injured, chances are they will pick one that sounds reasonable if they’re actually abusing their kid.

Then you’re in that really difficult situation, because now they’re giving you a reasonable mechanism that they didn’t have for the first three hours of their stay. Do you believe it? So, avoid giving them options. You can say, “Did any trauma happen? Even if the trauma didn’t happen today, has any trauma happened in the last week or two weeks that you weren’t thinking of because it didn’t just happen before you came in that you might be able to tell me about?” But don’t ask, “Did they fall off a kitchen counter? Maybe you dropped them? Maybe somebody was walking and fell down the stairs?” Don’t give them a list of potentially reasonable options.

As you know, MDCalc is a resource for clinical decision tools. Is there a role for evidence-based clinical decision tools or pathways that can help emergency physicians and pediatricians to better identify and manage suspected child abuse?

If you could create that, that would be absolutely amazing! I don’t know how to do that for child abuse in general. Your tools are so great because they are simple— I click a few buttons and it tells me that the probability of the patient having this is X or Y, and it’s so amazing to be able to go through that as an educator. We did that with the Centor criteria the other day—somebody was explaining why they didn’t think a patient had Strep, and I was like, well, let’s pull this up on MDCalc. We went through the actual numbers in this and other potential scenarios, and it really hit that point home. So it’s a great tool.

I don’t know that for child abuse we’re ever going to be at a point that you can ask, “Does the injury mechanism match the injury that you’re seeing? Is there anything on the skeletal survey that you’re seeing?” and get a probability of abuse. We definitely do not have the literature to allow you to create something like that. If I were to create a tool, I would feel silly — it would be things like just what we talked about: if any of these fractures are seen on the skeletal survey, if there’s bruising in any of the following locations, does the injury mechanism match the injury. The problem is, some of those are mutually exclusive. If they have bruising on the ear, for me, that’s a high risk injury for abuse, but if they just got hit with a baseball bat on the ear, it completely erases that. It would be really hard. I don’t know how you would pare that down to a clinical decision rule. But I would love it if someone could!

What’s the difference between BRUE and ALTE, and what’s the correlation of either of them with non-accidental trauma?

BRUE is the new term. ALTE, or apparent life-threatening event, has been around for decades, since 1986, and there are a lot of problems with that definition. It’s so broad. And it’s so dependent on what the parents thought and witnessed at the time of the event, which can be unreliable if they were very nervous. So, a group got together to redefine the ALTE patient that looks well in the ED . If somebody comes in and the parent says, “My kid turned blue and floppy at home,” and the kid comes in and they’re still blue and floppy, that’s not really a difficult patient for me to see. They might require a great deal of resuscitation, but a blue floppy baby is a sick baby, and I’m going to work that up like a sick baby. If somebody comes in and says, “My baby was blue and floppy at home,” and the baby is giving me high-fives and blowing bubbles and has a perfectly normal exam, it’s really hard to know what to do, because there’s a scary history and a terrified parent with a kid who looks absolutely perfect- it’s hard to even fathom that there’s anything wrong with them.

So what do you do in that situation? Until now, it has been a decision-making conundrum. A group out of the AAP [American Academy of Pediatrics], led by Joel Tieder, used the existing literature on ALTE, after there was a tipping point of having enough literature to actually do that, and ferreted out who falls into that ALTE-esque category but actually is at low enough risk of a subsequent event or having something serious, that could possibly go home and not get a huge workup. BRUE stands for brief resolved unexplained event—and it also sounds a little bit better than apparent life-threatening event, let’s be honest, “I’m sending your child home with an apparent life-threatening event!” is kind of a weird thing to say to people. This group came up with a low risk group of kids who’d had one of these resolved events who could actually be sent home and not get a sizable workup because we know their prognosis is so good. It’s sort of under that ALTE category, but has largely replaced ALTE as an illness.

In terms of the relationship between ALTE or BRUE, and child abuse, we know that with the old standard ALTE definition from 1986 (any child who’s had an event concerning to the caregiver characterized by some combination of color change, change in tone, apnea, choking or gagging), somewhere around 2.3% of the patients without other signs of abuse would have inflicted head injury diagnosed. We also know that these kids are subject to a number of different variants of child abuse over and above abusive head injury. We worry about poisoning—there was one study that showed about an 8% incidence of positive tox screen when they did not just the regular urine tox screen but they did more advanced testing. Some of them were just pharmacologic, over the counter cough and cold medications, but some of them were things like benzos and narcotics that obviously someone had given to the baby, probably to quiet them down.

We also worry about inflicted suffocation. There are families who suffocate their own children as a sort of form of Munchausen by proxy, and they stop the suffocation just before the kid has a fatal outcome, and then run in saying, “My kid stopped breathing.” There was a legendary study many years ago out of England where people were videotaped, people whose kids had recurrent ALTE or siblings who had died of SIDS and were hospitalized, and they videotaped parents suffocating their child until the infant passed out, and then runing out into the hallway screaming, “It’s happening again! It’s happening again! My baby stopped breathing!” So there are victims of inflicted suffocation population, inflicted head injury, and poisoning.

When you look at the patients that this AAP working group has identified as low-risk BRUE, the rate of child abuse in that group is very low- much lower than with the broader ALTE definition. So assessment of non-accidental trauma really is something that should be incorporated into any history and any physical exam for an ALTE patient — looking for bruises, looking for areas in the long bones and the spine that might be tender and need an x-ray, making sure you don’t hear or see anything on exam that would make you suspicious for abuse, certainly. But something else that’s interesting that has come out in a study looking at long-term follow up of ALTE patients is, if you follow them up for many years after their ALTE, a fair number of them (11%) do end up becoming victims of abuse. And I don’t think that’s because we’re missing abuse at the first presentation—I feel like ALTE is one of those nebulous presentations that might actually be a cry for help on the parent’s part. I think it’s good to have a talk and make sure there’s no depression or tension in the family that you need to worry about.

So there is sort of a connection between the two, just not that strong of a connection at the time of diagnosis. The AAP group has created a very nice algorithm for identifying those low risk patients, who are low risk for subsequent events, as well as being low risk for concerning pathology like abuse.

This is a very difficult question to answer with evidence. For a surgeon, you can ask, “Did this patient lacerate their spleen? Let’s look at the CT—oh, there’s a splenic laceration!” There is a gold standard, which may not be 100%, but it’s still pretty good. But for abuse, if I’m a savvy abuser, and you never catch me, then my kid goes into the non-abuse category when abuse is researched. And if I’m not abusing my child, but I get convicted of it because I’m a disenfranchised homeless meth user and someone expected me to be an abusive parent, they go in the abuse category. So our gold standard for anything related to abuse is really relatively poor and heavily weighted by our current concept of abuse. And I think that’s a huge limitation on our ability to find some type of algorithm to go by.

Come back next week to read part 2 of our interview, in which Dr. Claudius tells us about her work studying pediatric mass casualty and disaster preparedness, as well as psychosocial components of pediatric trauma and mental health.