Until relatively recently a prosecution for murder arising from the death of an infant from suspected “shaken baby syndrome” rested entirely upon scientific evidence. Without any sign of abuse or rough treatment or any background suggesting that the child was at risk of ill treatment, a person could still be convicted solely upon expert evidence confirming the presence of the ‘classic triad’ of symptoms. The majority view was that the existence of cerebral edema, subdural haematoma and retina haemorrhage was enough to prove, not only that an infant had been shaken, but that it was shaken with a sufficient degree of force for a jury to conclude that there had been intent to cause death or grievous bodily harm. Experts claimed to be able to identify the perpetrator as the last person to have care of the lucid infant as the loss of consciousness was thought to be prompt. Scientific research and understanding in this area has progressed and concerns that the reliance upon the triad of symptoms as the sole evidence of the condition can lead to miscarriages of justice. Although the Courts no longer treat the presence of the triad alone as determinative evidence of shaking, there remains an active debate as to whether “shaken baby syndrome” or non accidental head injuries attributable to shaking can be conclusively diagnosed.
It is universally accepted that shaking an infant is a dangerous action that can cause death or serious injury. Concern that this seems to have been accepted practice in the 1970s was well placed and there are few who would seek to criticize the drive to educate parents and care givers on the inherent danger. The phenomenon is unique to infants because of their anatomy; the weak and undeveloped neck muscles and disproportionately large head mean that, upon shaking or vigorous impact, the brain bounces back and forth against the skull causing bruising, swelling and bleeding. The only debate concerns whether the triad of symptoms alone is enough to prove that an infant has been shaken with the requisite force or whether there may be an innocent explanation.
The connection between the shaking of an infant and brain injury was first noted by the British pediatric neurosurgeon, Norman Guthkelch in 1971. In a review of thirteen cases of infant subdural haematoma, five showed no signs of direct violence to the head. Dr. Guthkelch concluded that the reason for that absence was that the babies had been shaken not battered. A few years later, this theory was confirmed by Dr John Caffey, a pediatric radiologist, who published two seminal articles: ‘Theory and Practice of Shaking Infants’ 124 Amer J Dis Child 161 (1972) and ‘The Whiplash Shaken Infant Syndrome’ 54 Pediatrics 396 (1974). Notably, neither Dr. Guthkelch or Dr. Caffey contended that subdural bleeding and retinal haemorrhages could only be caused by shaking. It was accepted that there could be innocent causes. Lucid intervals after the shaking were also acknowledged.
The issue of lucidity is an important one, as experts giving evidence in relation to the presence of the triad persisted in the view that the onsent of neurological symptoms and the consequent loss of consciousness would follow the shaking event. As such, a window of opportunity for the crime could be identified; the person with care of the child during the loss of consciousness would be the perpetrator. This conclusion was thoroughly undermined, and pediatricians, including Dr Caffey, accepted that periods of lucidity could follow the initial assault. Although a number of practitioners continue to maintain that shaken baby syndrome can be diagnosed, all now accept that the injuries cannot be timed, and that post assault lucidity does occur.
There was a prolonged campaign following the discovery of the effects of shaking to raise awareness of the dangers of shaking an infant. Within fifteen years, accepted medical opinion was that the presence of the triad of injuries established that a baby had been shaken. Both the medical community and the criminal justice system had arrived at a situation where the expert testimony established each element of the crime. It was claimed that the science could show that shaking was the only explanation for the presence of the triad, that the shaking must have been violent enough to prove the intent required to establish guilt and that the person responsible was the defendant. This led to a number of convictions based solely on the triad in the UK and the US, a situation which persisted until the case of Louise Woodward.
The 1998 prosecution of the British au pair, Louise Woodward, for the killing of a Massachusetts infant, Matthew, revived the issue. There had never been concerns that Louise Woodward was abusive towards Matthew or that she had handled him roughly. No other bruising was found upon Matthew’s body but the post mortem revealed a fatal brain injury and a skull fracture that could not be dated. Prosecution experts testified that Matthew’s brain injury proved that he had been shaken to unconsciousness; the defence argued that he had suffered from an older brain injury. Louise Woodward was convicted but later the charges were reduced to manslaughter and the sentence to one of time served. The Woodward trial had succeeded in bringing the issue to the attention of medical and legal professional on both sides of the Atlantic.
Challenges to the theories underpinning shaken baby syndrome gathered momentum as a result. In 2001 Dr Jennian Geddes published two papers which suggested that a relatively minor shaking might cause a fatal failure of oxygen to the brain. The relevance of this was that it cast doubt over the level of force required to cause a fatality, which in turn raised the question of whether it was possible to use the fact of shaking itself to prove that it was the intention of the defendant to cause death or grievous bodily harm; it had previously been thought that severe force comparable to high speed whiplash incidents were necessary. As a result of this hypothesis, the possibility was raised that a mere jostle or later revival attempts could cause that fatal failure of oxygen to the brain. These theories came to be known as Geddes I and II.
In 2005 the Court of Appeal heard four separate appeals which involved doubts about the use of the triad as the sole mechanism for a finding of shaken baby syndrome and subsequently, a conviction. The case of Harris  EWCA Crim 1980 considered the findings of Dr. Geddes. Of course, the Court of Appeal was not asked to, and could not, rule upon whether the triad was a reliable indicator of head injury; that is a medical diagnostics issue. The sole concern of the Court was whether the convictions were safe. To determine that the Court had to consider whether the triad alone was sufficient to diagnose shaken baby syndrome conclusively. That question was answered negatively;
“… the mere presence of the ‘triad’ does not automatically or necessarily lead to a diagnosis of non accidental head injury and/or a conclusion of unlawful killing. All the facts of the individual case must be taken into account.”
This was a landmark ruling in some respects, but not in others. From one perspective, it confirmed the view held by a number of members of the medical community, namely that there could not be an automatic or presumptive diagnosis. There was a subtle shift in the language used; the phenomenon formerly known as ‘shaken baby syndrome’ was now known as ‘non-accidental head injury’. As a result the use of emotive words such as ‘shaken’ would trouble the courts no further. Further and more importantly, the observations contained a message to both medical professionals and the prosecuting authorities that the triad alone would no longer conclusively prove an unlawful killing. This would have an impact on charging decisions in the future.
In the case of Lorraine Harris, the conviction rested solely on the triad without any other medical evidence or suggestion of abuse. The Court concluded that the finding of retinal haemorrhage was powerful evidence of shaking, but alone was not enough. The conviction was quashed. In the case of Raymond Rock, considered at the same time, the triad did not stand alone. There was no dispute that Mr. Rock had shaken his partner’s daughter and that she suffered an impact to the back of the head. The conviction was reduced to manslaughter as the required level of intent could not be proven. However, in the case of Alan Cherry, the defendant had been convicted of manslaughter rather than murder. Two of the elements of the triad were present and there were two separate scalp bruises so the conviction was upheld.
After the ruling in Harris the then Attorney General, Lord Goldsmith, conducted a seven-month review of eighty eight cases involving shaken baby syndrome including guilty verdicts and guilty pleas. The review found three that caused concern as they were based on the triad only. The actual figure may have been higher if Lord Goldsmith had not considered guilty pleas and admissions by defendants to be corroborative of the diagnosis. A defendant may plead guilty for other reasons than a full acceptance of the facts alleged against him. A defendant jointly charged with his partner and told that a guilty plea from one of them will be acceptable to the Crown may be under pressure to plead guilty in order to protect others. Also, a defendant may prefer the certainty of the sentence he faces for a plea to manslaughter to the risk of a conviction for murder. Similarly, terms like ‘admissions’ are unhelpful with non-accidental head injury. It is rare that a caregiver could avoid self incrimination; some explanation for the injuries of the infant wll usually be offered. A defendant who accepts shaking in the course of revival attempts or play fighting may be taken to have made ‘admissions’ but they would not be adequate for a safe conviction.
Nevertheless, as a result of the review by Lord Goldsmith, the Crown Prosecution Service issued new guidelines in March 2011. Although the CPS consider that non-accidental head injury can still be diagnosed by the classic triad, as part of their charging considerations and the application of the Full Code Test, they now require the triad plus some form of additional supporting evidence.
There remain further challenges to be made to findings of non accidental head injury, or shaken baby syndrome. We must remember that new generations of experts have been discrediting what, for decades, was accepted as a certainty. Dr. Geddes continued to pursue the matter, and published research in 2003 that has come to be known as Geddes III or the “unified hypothesis”. This challenged the proposition that the triad was diagnostic of shaking at all. It suggested that the triad could be caused by lack of oxygen to the brain leading to swelling. Lack of oxygen can be caused by many events that have nothing to do with abuse. This was considered in the case of Harris but it did not stand scrutiny at the time; Dr. Geddes herself accepted under cross examination that the hypothesis was to stimulate debate rather than assert a fact and that further research was needed. Inded, the CPS’s current policy is to resist any challenges to a diagnosis on the basis of the unified hypothesis.
However, this was considered in the case of A Local Authority v S  EWHC 2115 (Fam). This is a decision of the Family Court but can still be used as persuasive authority in criminal cases. This case involved an application by the local authority for a care order in relation to child S. The proceedings were born out of the death of the second child of the family, Z. The triad of injuries were found in Z. A number of experts gave evidence in that case but the court heard from two defence experts, Dr. Cohen and Dr. Squier, who both subscribe to the unified hypothesis in one form or another. Put simply, their view was that hypoxia in a child can lead to the triad of injuries in the absence of shaking. Mrs. Justice King did reject the evidence of Dr. Cohen and Dr. Squier, preferring that of other experts, but was at pains to make clear that four years had passed since Dr. Geddes accepted that her theory was in need of research and that new experts claimed to have built upon that hypothesis, and by implication, undertaken their own research. The area remains ripe for future research and development.
Although there are still queries and challenges being made to the unified hypothesis that the triad can be explained by lack of oxygen to the brain, some progress has been made in this area as the presence of the triad will not longer be conclusive and determinative of the case.. Geddes I and II can be relied upon to challenge the degree of shaking. Play fighting will not be sufficient to prove intention to cause serious injury; although the reality is that the more severe the injury the more likely it is that it was caused by rough handling, we are no longer facing unchallengeable medical evidence that the triad evidences whiplash like force that must have been deliberate. The scope for challenging an allegation of murder by non-accidental head injury has widened considerably, and the pace of continuing medical development and research will mean that there remains the potential for further challenges before the Court of Appeal.
First Published by the South Eastern Circuit in The Circuiteer