The path to stopping the country’s first execution based on the widely refuted theory known as “shaken baby syndrome” narrowed Tuesday after a district judge in Texas declined to vacate the death warrant for Robert Roberson, a 57-year-old man with autism who is scheduled to die by lethal injection this week.

Roberson’s case has galvanized a bipartisan coalition of 86 Texas lawmakers, scientists and even the former lead detective from his 2002 case to fight for a reprieve. Supporters cite Roberson’s case as a prime example of a conviction wrongfully secured by a decades-old theory many scientists and legal experts say is unreliable “junk science” in the vein of discredited forensics such as bite-mark and bloodstain-pattern analysis.

Lawyers for Roberson argue he faces execution over a nonexistent crime: His 2-year-old daughter Nikki Curtis died in 2002 from an undiagnosed case of double pneumonia, lawyers said; doctors incorrectly presumed symptoms such as brain swelling and bleeding were from abuse and did not investigate other possibilities. Suspicion of Roberson’s guilt was fueled by his seemingly unemotional response to Nikki’s dire condition — the result, lawyers say, of autism spectrum disorder, which he was not formally diagnosed with until 2018.

Advertisement

“It is terrifying that Robert, an innocent, disabled man with the most gracious heart, is scheduled to be executed under an invalid warrant issued by a seemingly biased judge in just two days’ time,” Gretchen Sween, an attorney for Roberson, said in a statement Tuesday following the hearing on his motion to vacate the death warrant.

Roberson filed another petition Tuesday to the Texas Court of Criminal Appeals, after it denied a motion last week to stay his execution or reexamine his case. The appeals court’s Oct. 11 decision came just two days after it overturned the conviction of a Dallas man who Roberson’s lawyers noted was prosecuted under the theory of shaken baby syndrome in “a markedly similar case” that used “the same expert opinions that this Court has now recognized are not grounded in science.”

Roberson’s attorneys could also file an appeal to the U.S. Supreme Court, which last year declined to review his case.

Advertisement

His best chance now lies with Texas Gov. Greg Abbott, a pro-death-penalty Republican who has only once used his executive powers to stop an execution. The Texas Board of Pardons and Paroles must recommend Roberson for a commutation or reprieve for Abbott to act. The governor can also independently grant a 30-day execution delay.

Abbott’s office did not respond to a request for comment on Roberson’s case Tuesday.

Roberson’s is the latest high-profile death-penalty case in recent weeks to come under scrutiny because of questions about the quality of evidence used to convict the defendant. Last month, South Carolina executed 46-year-old Freddie Owens for the 1997 fatal shooting of a convenience store clerk, despite an admission just days earlier that a key prosecution witness had lied on the stand. Less than a week later, Missouri executed 55-year-old Marcellus Williams, who for years maintained he was innocent in the 1998 killing of a former reporter for the St. Louis Post-Dispatch. Williams’s DNA did not match the forensic evidence recovered from the crime scene.

Advertisement

“The evidence is assembled and available to the Texas authorities, but no one with the power to stop Roberson’s execution is paying attention,” John Grisham, the best-selling novelist and retired lawyer who is among Roberson’s most high-profile supporters, wrote in a Washington Post editorial Tuesday. “The courts slammed all the doors on the basis of technicalities, and even politicians’ pleas have been ignored.”

In theory, post-conviction appeals can serve as a fail-safe that allows the court system to reckon with past errors and avoid wrongful execution. In practice, the courts are rarely so nimble — or willing, according to death-penalty researcher and Fordham law professor Deborah Denno.

For pragmatic reasons, courts are concerned that opening the door to one challenge could lead to a cascade of challenges to cases determined using old science that, at one point, was widely accepted.

Advertisement

“It’s why courts are so loath to start going down this path,” Denno said. “If you start with shaken baby syndrome evidence, next you can look at fingerprint evidence, or any array of scientific evidence that was accepted before.”

Then there’s the gap between a jury’s conviction and a defendant’s execution — an average of 18.9 years, according to 2020 figures from the nonprofit Death Penalty Information Center.

“By the time defendants come up for execution, we can see all the problems with these cases,” Denno said. Technology advances, scientific theories are discredited and the understanding of factors like neurodivergence, trauma and mental illness grow more sophisticated.

A scientific theory comes under scrutiny

On Wednesday, the Texas House Committee on Criminal Jurisprudence will hold a hearing to discuss the effectiveness of Article 11.073, a 2013 law colloquially known as the “junk science statute” that gives defendants a pathway to challenge their convictions if new scientific evidence undermines that from their original trial.

Advertisement

“The whole point is that we have this law in Texas, but it does not appear the courts are following that law,” said Republican state Rep. Lacey Hull, who learned of Roberson’s case in her role as vice chair of the bipartisan Criminal Justice Reform Caucus. “Robert should have been granted relief under the junk science statue.”

The committee late Tuesday filed a petition with the appeals court requesting it delay Roberson’s execution until after the 2025 legislative session so it could made amendments to the junk science statute that might affect his case.

“It is beyond dispute that medical evidence presented at Mr. Roberson’s trial in 2003 is inconsistent with modern scientific principles,” the committee wrote in its Tuesday filing. “Article 11.073 was meant exactly for cases like this one.”

Advertisement

For decades, doctors were trained they could presume child abuse in cases in which a child presented with a certain array of symptoms, known as the “triad,” which became associated with shaken baby syndrome: bleeding on the surface of the brain, bleeding behind the eyes and brain swelling.

The hypothesis was first introduced by British pediatric neurosurgeon Norman Guthkelch in the early 1970s and expanded by prominent pediatric radiologist John Caffey. Caffey theorized that bleeding on the surface of the brain and in the back of the eyes was evidence that the victim had been violently shaken.

By the 1990s, the American Academy of Pediatrics declared shaken baby syndrome “a clearly definable medical condition,” citing Guthkelch and Caffey’s findings. The group encouraged physicians to immediately report any suspicion of head injury caused by shaking to the authorities.

Advertisement

For years, many doctors presumed the “triad” could only be explained by violent shaking. Since Roberson’s arrest in 2002, research has emerged showing diseases, genetic conditions and accidents — such as pneumonia, birth trauma and a short fall — can produce the same triad of symptoms.

Even Guthkelch came to doubt the theory he introduced, writing in a 2012 essay that one shouldn’t assume that bleeding over the brain and behind the eyes “are caused by trauma, rather than natural causes.” He stressed that shaken baby syndrome is a hypothesis, “not proven medical or scientific facts.”

“I am doing what I can so long as I have a breath to correct a grossly unjust situation,” Guthkelch told The Washington Post in 2015, about a year before his death. “I think they’ve gone much too far.”

Over the years, the reliability of the “shaken baby” diagnoses came under greater scrutiny in court cases. Since 1989, at least 32 people convicted under the shaken baby syndrome hypothesis have been exonerated, according to the National Registry of Exonerations.

A short lifetime of illness, overlooked

In 2002, Roberson was single father living in rural Palestine, Tex., with his young daughter, who had a history of chronic illness, including ear infections and unexplained “breathing apnea,” according to court documents. The week before the girl’s death, she was vomiting, coughing and running a fever, prompting Roberson to take her to the pediatrician.

Advertisement

Nikki was prescribed Phenergan, in cough syrup with codeine — a medication the Food and Drug Administration in 2018 restricted from children because of its potential for causing breathing difficulties and death. Two days before her death, Nikki had registered a fever of 104.5 degrees Fahrenheit.

The morning of Jan. 31, 2002, Roberson heard his daughter cry from falling out of bed, his lawyers said. Later that morning, he found the girl limp and unresponsive before rushing her to the ER.

When hospital staff at the emergency room noticed a lump on Nikki’s head, doctors ordered a CT scan that revealed a small bleed on the surface of Nikki’s brain — one of the symptoms of the “triad.” According to court records, the CT scan showed Nikki’s brain had swollen and shifted to one side, despite no other indications of abuse, such as fractures, abrasions or neck injuries.

Police arrested Roberson the next day, relying “solely” on an affidavit provided by a pediatrician at Children’s Medical Center in Dallas, where Nikki had been transferred, according to court records.

Despite claims by Nikki’s maternal grandparents that the child was “totally well” when they last saw her the night before she died, court records indicate that Nikki had gone to the doctor more than 45 times in her short life, including back-to-back visits in the three days before she died.

‘Forever haunted’ by prosecution

Roberson’s case has drawn an unlikely coalition of supporters across the political aisle, which experts like Denno said is especially notable in a deeply pro-death-penalty state such as Texas.

Hull, the Republican member of the Criminal Justice Reform Caucus, said she gets involved with cases when the defendant has a strong innocence claim, in part because she wants Texans to have faith that prosecutors and judges are getting cases right. Putting Roberson to death, she said, undermines that confidence.

“Executing Robert would be a stain on our criminal justice system in Texas,” Hull said. She was among Texas lawmakers who visited Roberson last month to pray with him.

But of Roberson’s defenders, few are more passionate than Brian Wharton, the former chief detective of the police department in Palestine, Tex., who helped convict Roberson 21 years ago.

“I will be forever haunted by my participation in his arrest and prosecution,” Wharton said during a press briefing last month. “He is an innocent man.”

Now a Methodist minister retired from law enforcement, Wharton told Roberson as part of a July video op-ed, “I’m convinced we did the wrong thing.”

Since Roberson’s conviction in 2003, Wharton now admits that investigators did not consider other possible causes of Nikki’s injuries. While Nikki’s CT scan did not show evidence she had been beaten, Wharton said he simply accepted the narrative that doctors and the medical examiner put forth.

“So many things have [gone] — and still are going — wrong in Robert’s case,” Wharton said during last month’s briefing. “I am firmly convinced Robert is an innocent man. There was no crime that was committed here.”