Larger Debate

The modern era of child protection essentially started in 1962, with the publication of Dr. C. Henry Kempe’s call to action, “The Battered Child Syndrome,” in the Journal of the American Medical Association. Within a decade, most states had passed child-abuse statutes that established social service organizations charged with removing children from abusive homes and required doctors, teachers, and other “mandated reporters” to report all cases of suspected abuse to the authorities. Doctors began establishing guidelines for how to identify abused children.

From the beginning, the guidelines tended to emphasize the “discrepant history,” that is, an explanation offered by the parents that doesn’t adequately explain a child’s injuries. If a parent reports that an infant rolled off a changing table, for example, but the child arrives at the hospital with broken bones and bruises in unlikely locations, the treating doctors conclude that the baby was beaten.

Inevitably, these decisions are subjective. Objective scientific research in the arena is almost impossible to conduct, because there is no way to know what happened before the child’s first medical contact. Much of the child abuse literature is based on observational studies, which begin with the sorting of cases by the child abuse experts into groups of “abused” and “non-abused” patients. The researchers then analyze the differences between the groups—with the result that the original sorting algorithm  appears to be validated by the study. See, for example, this 2000 paper by some of the leading authors in the field:

Early on, doctors adopted a number of guidelines that were never tested against reality. Spiral fractures, for example, were once considered a convincing sign of abuse, even though doctors generally agree that adults can suffer a spiral fracture from a fall, or any other trauma capable of breaking a bone, and recent research has revealed that spiral fractures are no more common in abused children than transverse fractures (see, in which the authors note that “many practitioners think spiral fractures are pathognomonic of abuse”). Based on their own observations, child abuse doctors have also designated posterior rib fractures in infants as a sign of abuse, although recent research has shown that children can suffer posterior rib fractures from birth trauma (

Similarly, the “classic metaphyseal lesion”—a discontinuity at the edge of a bone’s growth plate—is often described in court as a sure sign of abuse, on the theory that the lesion is caused by twisting of the limbs during an assault, and is never produced by natural events. Recently, doctors looking at the bigger picture have noted that CMLs actually look a lot like some of the lesions seen in rickets, a condition that had become quite rare by the time the child-abuse doctors were establishing their guidelines (see And as various researchers have noted, rickets is now making a come-back, as North Americans tend to be deficient in vitamin D.

In the early 1970s, Dr. A. Norman Guthkelch in Britain and Dr. John Caffey in the U.S. proposed that children with subdural hematomas (bleeding in the lining of the brain) and retinal hemorrhages (bleeding at the backs of the eyes) but with no obvious signs of impact, might be the victims of abuse by shaking. This hypothesis was quickly adopted in the courtroom by way of expert opinion, without objective scientific support.

Though more rigorous procedures are now entering the equation, for several decades the child protection teams charged with evaluating abuse cases routinely concluded that the cluster of symptoms that defined shaken baby syndrome always proves abuse by shaking, without considering other causes. Some physicians, however, have looked further and found underlying conditions that can produce the same symptoms or predispositions to those symptoms. The list is long and growing, and there are likely others still unidentified.

Physicians also testified historically that children who were shaken collapsed immediately following the assault, so that the person last with the child could reliably be identified as the perpetrator. This position has been shown to be false: Children who are abused or suffer accidental head injury react in widely varying ways, and their condition is not always obvious, even to experienced care providers.

Finally, there is the problem of short falls. Child abuse pediatricians have historically claimed that household falls cannot cause the findings commonly attributed to shaking. Report of falls from high chairs, changing tables, the arms of parents—even falls down stairs—were historically rejected in favor of a diagnosis of abuse, even when the falls were witnessed by others. The American Academy of Pediatrics no longer takes this position, although some physicians still testify to it in court.

Unfortunately, all of this research and experience has taken place in the context of the courtroom. Child abuse doctors have become accustomed to defending their opinions in court, against the opinions of doctors for the defense. The result is that the two sides do not communicate as colleagues looking together for the truth, but as bitter enemies. Allied with the prosecution, the child abuse experts are in the habit of rejecting the explanations offered by doctors outside of their own specialty, and maintaining faith in what they’d been taught by their teachers. After seeing their honest opinions confirmed by legal outcome, they believe that shaken baby syndrome has been proven, that fractures in an infant too young to walk prove abuse, and that their critics are either incompetent or cynical opportunists hired by the defense to get guilty parents and caretakers off the hook.

This lack of communication will not help children or families. We need an objective organization to take a careful look at the lack of science behind the opinions that are still sending innocent people to prison. A number of activists in the field are calling for a review of the scientific basis for diagnoses of child abuse by the National Academy of Sciences, which in 2009 produced an influential report ( about forensic science in the courtroom.