The ongoing scientific and legal debate over the diagnosis of abusive head trauma (AHT) exemplifies the difficulty and importance of accuracy when medicine crosses into the courtroom. AHT is defined by the American Academy of Pediatrics as “an injury to the skull or intracranial contents of an infant or young child…due to inflicted blunt impact and/or violent shaking.” There are 25-35 AHT cases per 100,000 children less than 1 year old diagnosed per year; AHT is most common in this age group. Abusive head trauma is fatal for 10-20% of victims, and many survivors suffer long-term disabilities. Patients are unable to speak for themselves due to their young age, and independent witnesses are rare, so doctors must usually rely on medical exam results and caregiver histories. The stakes are also high: Diagnose AHT where there is none, and an innocent parent has their life uprooted, facing jail time, losing custody of any other children they may have, and intensifying their distress while they are still grieving the loss or injury of their child. Miss a true case of AHT, and justice is not served; the perpetrator could go on to endanger more children.
AHT was formerly known as shaken baby syndrome; its name and definition were updated in 2008 to shift emphasis from the mechanism to the clinical effect. However, this name change may also reflect uncertainty over whether shaking alone can cause the symptoms typically associated with AHT.
As the lengthy citations of the AAP report would suggest, there is a robust body of research establishing the validity of AHT as a diagnosis, and providers should be aware of AHT as a possibility when a child presents with characteristic symptoms. However, the diagnosis of AHT is complex; especially under ambiguous circumstances, providers must carefully consider potential implicit biases to avoid perpetrating injustice in either direction.
According to the American Academy of Pediatrics’ (AAP) 2025 technical report on AHT, which seeks to summarize the current body of evidence, researchers investigating this question typically inflict shaking injury on animal models or surrogates (mechanical dolls intended to replicate human biomechanics). While this provides some means of studying AHT, animal models are not fully analogous to human infants due to differences in maturity, brain mass, and overall brain response to trauma. Surrogates cannot perfectly mimic humans’ flexible spines and necks, often leading to inaccurate and contradictory data.
Those skeptical of AHT argue that medical professionals only consider the “triad” of symptoms that typically characterize AHT: subdural hematoma (SDH), retinal hemorrhage (RH), and cerebral edema, which is brain swelling. Sole reliance on the triad would be problematic if each of these symptoms can also be caused by underlying conditions or accidents, a claim asserted by Norwegian professor emeritus Knut Wester and colleagues, and contested by Colorado clinical professor Daniel Lindberg and colleagues.
Proponents of the current AHT diagnosis argue that the diagnostic process is far more nuanced and involved than what critics of the triad might suggest. The AAP technical report lays out a detailed diagnostic process extending far beyond the triad, considering a changing or inconsistent medical history as an AHT indicator and listing several differential diagnoses. While the outlined process is extensive, it is not always fully applied in everyday life, whether due to a lack of specialized personnel or mere oversight. Additionally, a critical review of the AAP report argues that its reliance on expert opinion allows too much room for subjectivity, especially as expert opinion is typically ranked as the lowest level of evidence in evidence-based medicine hierarchies due to the high risk of bias (in contrast to a randomized controlled trial, which would be ranked at the highest level). It also argues that diagnosing by exclusion (ruling out other conditions) introduces further subjectivity and uncertainty, considering that doctors must make judgment calls about the likelihood of each condition, and about half of infant deaths are caused by rare conditions.
In one case of AHT misdiagnosis, Kristina Kerlus was falsely accused of fatally shaking her infant son, Jocai Davis, based on a coroner’s autopsy report and a photo of Jocai in a supposedly “awkward position” the night before he entered the hospital. Kerlus’ children were taken from her and placed in her aunt’s custody.
Jocai was born with Sickle Cell Trait (SCT); while most people with this trait (predominantly African American) show no symptoms, if Sickle Cell Disease manifests, it can lead to serious bleeding complications. The coroner mentioned that Jocai had SCT, but he cited AHT as the cause of death due to the presence of hemorrhages in the brain, spinal cord, and behind the eyes.
However, after the coroner's diagnosis, pediatric forensic pathology expert Dr. Evan Matshes analyzed the same evidence. He discovered extensive sickling of Jocai’s blood cells, which clotted and killed a blood vessel while he was still alive, causing hemorrhaging. Matshes concluded that Jocai died of natural causes, and the state voluntarily dismissed the case against Kerlus. This process took 4 years, during which Kerlus and her children remained separated, showing the devastating ramifications of a false AHT diagnosis.
AHT is disproportionately diagnosed in Black, Native American, or male children, but rather than reflecting some kind of biological susceptibility, this is far more likely due to a complex interplay of social determinants of health (e.g. socioeconomic status) and providers’ implicit bias potentially leading to misdiagnosis; thus, we cannot say that race itself is a risk factor for AHT, but rather is more likely to reflect a racial health disparity.
Implicit bias may lead to AHT being overreported in marginalized groups and underreported in white and stable families. In the retrospective chart review of Brown University researcher Carole Jenny and colleagues, doctors were more likely to miss an AHT diagnosis in young white children and in those with ambiguous symptoms such as irritability or vomiting. Of all 173 cases reviewed, 54 (31.2%) of cases were initially misdiagnosed, and it took an average of 7 days to arrive at the AHT diagnosis. Of those misdiagnosed, 15 (27.8%) were later reinjured, 22 (40.7%) suffered medical complications, and 5 died; the researchers concluded that timely diagnosis could have prevented 4 of those deaths. New York-based clinical professors Vincent Palusci and Ann Botash recommend that doctors should critically examine their “gut feelings” to prevent implicit bias from taking over, and that hospitals should institute training in recognizing implicit bias as well as markers of abuse.
AHT has been criticized for relying on circular reasoning: Once a child is misdiagnosed with AHT, they become examples for doctors’ judgment of future patients, which can result in a cycle of misdiagnosis. Even when someone confesses to inflicting AHT outright, this should not be seen as incontrovertible evidence–confessions may be coerced or fabricated in the interest of a plea deal, or some other motive independent of any official or legal incentive.
Some studies attempt to avoid this concern by conducting research in areas where plea bargains are not an option for AHT. For example, Catherine Adamsbaum and colleagues, of France’s St. Vincent de Paul Hospital, studied caregiver confessions of AHT, and all described the shaking as “extremely violent,” and in most cases repeated, with no impact. This would show that shaking alone is capable of causing the injuries observed in AHT.
However, University College London researcher Cyrille Rossant and University of La Laguna researcher Chris Brook aimed to show that confessions like these may still be unreliable. They surveyed 118 French caregivers who had been suspected or accused of AHT but maintained their innocence. Of the 97 who had been interrogated by police, nearly 80% reported that police had “presented the diagnosis of child abuse as absolutely certain and irrefutable,” 46% reported that they had been told that confessing would lead to lenience, and 43% that confessing would lead to the child being returned to the other parent. 11% said that they made false confessions either for this latter reason or to protect the other parent. Although this study relies on self-reporting, it calls into question the validity of confessions even in the absence of legal incentives to be dishonest.
In a separate study by Brook comparing diagnosed AHT to non-AHT cases that had an “independent unbiased witness” in the pediBIRN dataset, he noted that 14 out of 100 witnessed non-AHT cases were previously misdiagnosed as AHT. There was no statistically significant difference in symptoms between the misdiagnosed non-AHT cases and the AHT cases, suggesting that more severe brain injuries mimic and can be misdiagnosed as AHT.
As this article has shown, AHT can present in a variety of ways, and there are a number of diagnoses that must be ruled out before concluding that a child has AHT. It is crucial to take steps to eliminate medical bias and exercise caution in diagnosis, particularly when treating patients of disadvantaged backgrounds, to avoid perpetuating current health disparities. We must also critically examine current and future AHT research so that we can protect children and families from injustice in either direction.