To tell if a baby has been injured or killed by being shaken, the courts use three hallmark symptoms: Bleeding and swelling in the brain and retinal bleeding in the eyes. Along with other evidence, those standards are used to convict caregivers of abusive head trauma, both intentional and unintentional, that can result in blindness, seizures, severe brain damage or death.
But in recent years a small cadre of experts testifying for the defense in cases across the country has called into question whether those symptoms actually indicate abuse. Though they are in the minority – disputing the consensus of child abuse experts, pediatricians and an extensive evidence base – they have gained traction in the media and in courtrooms by suggesting that shaking a child cannot cause these injuries. Instead, they argue that undiagnosed medical conditions, falls or other accidents are the cause.
So researchers have developed and validated a tool doctors can use to distinguish between head injuries resulting from abuse and those from accidents or medical conditions. The method, described in the journal Pediatrics Monday, asks doctors to check for six other injuries, each of which increases the likelihood that a head injury resulted from severe shaking, blunt force or both.
"It is vitally important that abuse head trauma is diagnosed accurately so that the team looking after the child can ensure that they receive appropriate support and are protected from further harm," lead study author Laura Elizabeth Cowley, a PhD student at the Cardiff University School of Medicine in the U.K., said in an email.
"However, it is also important that accidental head injury cases are not wrongly diagnosed as abusive," she continues, "because this can have devastating consequences for the families involved."
The tool itself is simple: listing rib fracture, seizures, long-bone fractures, bruises on the head or neck, periods of not breathing called apnea and bleeding in the retina of the eyes. The more of those a child suffers, the more likely the case resulted from abuse.
The researchers used medical records and child protection files of 198 babies with head trauma to see how well the tool identified child abuse. Extensive investigations, often including confessions or independent witnesses, had already established whether these children, mostly under 1 year old, had been abused.
By using a standard that at least three of the six symptoms were needed to presume abuse, the method correctly identified 82 percent of 133 cases as not abusive. It also correctly found that two thirds of the other 66 children had suffered abusive head trauma, though it missed 13 cases of abuse in which the child had few of the physical symptoms used by the tool. Twelve cases mistaken for abuse involved substantial injuries from a car accident or other situations in which a pediatrician would probably not use this tool.
"Part of the challenge is that there is no gold standard for making a diagnosis of child abuse – you can't do a blood test," said Cindy Christian, chair of both Child Abuse and Neglect Prevention at The Children's Hospital of Philadelphia and the American Academy of Pediatrics' Committee on Child Abuse and Neglect. "It's always going to be a clinical diagnosis with social investigation. A bruise is a bruise. A broken bone is a broken bone. A subdural hemorrhage is a subdural hemorrhage. There are certain injuries that are much more highly correlated with abuse, and there aren't very many things that cause these together."
More than a half million U.S. children suffer abuse each year, including approximately 30 cases of abusive head trauma among every 100,000 infants, according to the American Academy of Pediatrics. Each year about 80 children die from abusive head trauma. They are among the 1,500 children who die each year from abuse or neglect in the U.S.
In 2009, the AAP reclassified "shaken baby syndrome" as abusive head trauma to be more inclusive of all the ways a child's head can be injured through abuse, including but not limited to violent shaking.
Doubters of shaken baby syndrome exploited this shift in terminology to claim that shaking alone could not cause the triad of injuries often used in court cases – brain swelling, subdural hemorrhaging (bleeding between the brain and its outer membrane) and retinal bleeding. One of those doubters is Jan Leestma, a neuropathologist from Chicago who has testified for the defense in dozens of court cases on abusive head trauma.
"The idea that you can shake a baby and cause subdural brain hemorrhage – you can't without an impact," Leestma said. "I don't mean to say there's no child abuse – of course there is – but shaking as abuse is a straw man. It probably can't produce the kinds of things that people say it can."
Yet a strong body of research shows that shaking can injure infants and toddlers and cause these symptoms. A study in which perpetrators confessed to shaking the child showed that violent shaking, on average about 10 times until the baby stopped crying, could cause all these symptoms – even when three quarters of them involved nothing more than shaking.
The real straw man argument is the idea that diagnosing abusive head trauma relies solely on those three injuries and can lead to false accusations of abuse, said Bob Sege, division director of Family and Child Advocacy at Boston Medical Center and a member of the AAP Committee on Child Abuse and Neglect.
"At a personal level, our first reaction is really hoping it's not child abuse and to go through all the other possibilities to make it be something else, but sometimes you can't," Sege said.
When a child comes in with those symptoms, pediatricians conduct a thorough physical exam to look for other injuries, get a full history from the parents and look for inconsistencies in caregivers' stories. They also often enlist other experts, such as radiologists, ophthalmologists and child abuse specialists, to look at the evidence.
"There are certain bleeding problems and very rare metabolic diseases that might cause some of these symptoms," Sege said. "They're unfortunately very uncommon compared to child abuse, which is not as uncommon as I wish it was."
A common pattern in these cases arises when good but overwhelmed parents become frustrated with a baby's crying and reach their breaking point, so they shake the baby to stop the crying, Sege said.
"If you're a parent and you feel that [frustration], that's the time you put the baby down screaming and crying, close a door, make a cup a tea, call a friend, whatever you need to."
The method described in the study quantifies what pediatricians see in their clinical work, Christian said. She hopes it will make it more difficult to deny abuse when it occurs.
"All of our clinical experience and evidence shows that shaking can be extraordinarily harmful to a baby," she said.