What is shaken baby syndrome, the controversial diagnosis for which Robert Roberson is set to die?

Experts are divided on the use of shaken baby syndrome diagnoses in criminal cases. (Credit: Natalia Ojewska for The Texas Tribune)

---

On Thursday, Robert Roberson is set to become the first person in the United States to be executed on the basis of a shaken baby syndrome diagnosis.

This controversial medical diagnosis and disputed legal theory has divided courts, doctors, lawyers and law enforcement, with some calling it “junk science” and others continuing to embrace it as key to identifying and stopping child abuse.

The Texas Court of Criminal Appeals and the Texas Board of Pardons and Parole have rejected Roberson’s pleas for clemency. A bipartisan group of lawmakers, legal experts and even the detective who originally investigated Roberson’s case are making a series of last-minute Hail Mary efforts to try to save his life by challenging his conviction under Texas’ junk science law.

[Texas Supreme Court temporarily stops Robert Roberson’s execution]

The 2013 law allows courts to overturn a conviction when scientific evidence at the crux of a case has changed or been discredited. Roberson has tried repeatedly and unsuccessfully to use this law to prove his innocence.

While it refused Roberson’s claims last week, the Texas Court of Criminal Appeals just days prior threw out another shaken baby case, saying if “newly evolved scientific evidence were presented” at the original trial, “it is more likely than not that he would not have been convicted.”

At least 34 people convicted based on a shaken baby syndrome diagnosis have been exonerated, according to the National Registry of Exonerations.

What is shaken baby syndrome?

Starting in the 1980s, shaken baby syndrome was commonly diagnosed when doctors detected what’s known as “the triad” of symptoms — unexplained bleeding on the brain, bleeding behind the retinas and brain swelling.

Dr. Norman Guthkelch, a British pediatric neurosurgeon, is widely credited with first connecting these symptoms with shaking a baby in a 1971 paper. Researchers built on his theory as awareness around child abuse grew, and in 1993, the American Academy of Pediatrics declared it a “clearly definable medical condition.”

Doctors, wary of missing cases of extreme child abuse, latched onto the diagnosis as a way to explain otherwise inexplicable symptoms. Over time, a consensus emerged that shaking could be the only way this triad of symptoms could emerge. In 2001, the American Academy of Pediatrics published a position paper that said these three symptoms created a “presumption of child abuse.”

The next year, Roberson took his 2-year-old daughter Nikki to the emergency room in East Texas, saying she had fallen off the bed. Nikki had been sick in the days leading up to the incident, with a fever of 104.5, and had been prescribed medications that are no longer given to children. She was unable to be revived.

A snapshot of Robert Roberson and his daughter, Nikki Curtis. (Credit: Courtesy of Roberson's attorneys)

Doctors and nurses suspected child abuse and didn’t believe such a low fall could have caused her injuries. At trial, doctors testified that Nikki exhibited the triad of symptoms associated with shaken baby syndrome and Roberson was convicted.

What does the science say? 

Since Roberson was convicted in 2002, a lot more science has emerged about the diagnosis of shaken baby syndrome.

Starting around that time, research began to show that many of the same symptoms associated with shaken baby syndrome could also occur from short falls, as well as a wide range of naturally occurring medical conditions and accidental traumas. Studies have found that shaking a baby hard enough to generate those symptoms would likely also result in a broken neck.

Guthkelch himself disavowed the way his hypothesis had been used and, before his death, became an outspoken advocate against the primacy of the shaken baby diagnosis. He told NPR in 2011 that there was no way to say those three symptoms were always caused by shaking and nothing else.

"In a case of measles, if you get the diagnosis wrong, in seven days' time it really doesn't matter because it's cleared up anyhow,” he told NPR. “If you get the diagnosis of fatal shaken baby syndrome wrong, potentially someone's life will be terminated."

In 2009, the American Academy of Pediatrics widened the shaken baby syndrome term to “abusive head trauma,” and acknowledged that symptoms can be caused by incidents other than shaking, including short falls. In the policy memo, the group noted that “restraint is required” when making a diagnosis of abusive head trauma, and doctors should consider other possible causes for the symptoms.

In 2020, as more shaken baby syndrome convictions were being challenged, the group said these changes were “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.” The academy “continues to embrace” shaken baby syndrome as a diagnosis, they said in the position paper.

What have the courts said? 

There have recently been a flurry of shaken baby cases rejected by courts across the country.

In July of this year, the Michigan Supreme Court overturned a 2006 shaken baby murder conviction. Two Ohio cases have been rejected by higher courts, one after the medical examiner changed their assessment that a 3-year-old died of shaken baby syndrome to their belief that the child died from a short fall.

Last year, a judge in New Jersey likened the shaken baby syndrome to “junk science,” and noted that “no study has ever validated the hypothesis that shaking a child can cause the triad of symptoms” associated with shaken baby syndrome. An appellate court upheld that ruling.

In 2016, the Texas Court of Criminal Appeals halted Roberson’s execution under the junk science statute, but in 2023, decided the doubt surrounding the cause of Nikki’s death wasn’t enough to overturn his death sentence.

But last week, the same court threw out a Dallas County case in which Andrew Wayne Roark was convicted of shaking a 1-year-old, causing irreversible brain damage. If called to testify today, the state’s experts “would be confronted with twenty years of reputable scientific studies and publications that, if graphed, continually point away from their stated positions,” the court ruled. As a result, “it is more likely than not he would not have been convicted.”

The same child abuse specialist testified at Roark and Roberson’s original trials, according to Roberson’s attorney, Gretchen Sween. Based on this ruling, Roberson “too should, quite logically, be awarded a new trial,” she said in a statement.

And yet, the Court of Criminal Appeals rejected Roberson’s latest request without offering a written opinion. Roberson is still set to be executed, and there are people across the country still being convicted of abusive head trauma based on this same science, said Randy Papetti, a lawyer, expert and author of “The Forensic Unreliability of The Shaken Baby Syndrome.”

“There's been a recognition, at least in certain courts, that the science has changed so much that they're actually vacating convictions,” he said. “At the same time, the science is deemed good enough to be putting people away in other matters and playing an enormous role in executing a man.”

The current state of the legal field around this diagnosis, he said, is “completely contradictory.”