FOUR-month-old Patrick Harris died on 6 December 1999. But his grief-stricken mother, Lorraine, lost more than her son. She also lost her freedom.
Shortly after Patrick鈥檚 death, a post-mortem revealed three distinct signs that he had been abused. His brain was swollen, its surface was covered in blood clots known as subdural haematomas, and there was bleeding behind the retinas in Patrick鈥檚 eyes. Any one of the symptoms might have been accidental, but all three together were taken as a sign that Patrick was a victim of shaken baby syndrome. An English court agreed and found Lorraine guilty of manslaughter.
More than five years later, the Court of Appeal disagreed. Last week, it declared that the original conviction was unsafe, and quashed it.
Advertisement
But the appeal court did more than that. In a 61-page ruling, it laid bare the uncertainties and confusion surrounding our understanding of shaken baby syndrome. And in a damning verdict, it condemned the whole notion that, by themselves, the three signs found in Patrick鈥檚 young body 鈥� the swollen brain, subdural haematomas and retinal haemorrhaging 鈥� proved that a baby had been shaken to death. Indeed, the court said that this trio of symptoms could only be considered a hypothesis for the cause of death, and questioned the usefulness of the term 鈥渟haken baby syndrome鈥� itself.
Last week鈥檚 ruling could lead to the reopening of up to 88 similar cases in the UK and up to 15,000 suspected child abuse cases in the US, says John Plunkett of the Regina Medical Center in Hastings, Minnesota.
Plunkett, an expert in child forensic pathology, testified during consolidated hearings on behalf of four British people, including Lorraine Harris, who were asking the appeal court to quash their convictions for physically abusing their children.
Of the four, two were acquitted: Harris on the single charge of the manslaughter of Patrick, and Michael Faulder on the charge of causing grievous bodily harm to a 7-week-old baby, known to the court only as N. Ray Rock鈥檚 conviction for murdering 1-year-old Heidi Smith was reduced to manslaughter, while the court dismissed Alan Cherry鈥檚 appeal against his conviction for the manslaughter of his partner鈥檚 daughter Sarah.
But the ruling also showed just how little research has been done into the causes of infant deaths, and just how little agreement there is among experts over what physical signs may indicate.
Some doctors have come to believe the trio of symptoms found in Patrick can be taken as evidence of physical abuse, even without other evidence. That鈥檚 because there is a strong correlation between the presence of such injuries in children and subsequent confessions by parents that they did indeed violently shake their child.
In a study published last year in the Archives of Pediatric and Adolescent Medicine (vol 158, p 454), Suzanne Starling of the Children鈥檚 Hospital of the King鈥檚 Daughters in Norfolk, Virginia, examined 81 cases where carers confessed to abusing children that had these three symptoms. Of these, 32 said they had done no more than shake the infant.
鈥淲e have confessions of hundreds of people who say they did it, and the patterns of injuries are unique to shaken baby syndrome,鈥� says Randy Alexander, a senior paediatrician at the University of Florida at Jacksonville. 鈥淵ou don鈥檛 see these injuries in the tens of thousands of other accident victims we see.鈥�
But other experts disagree, saying that brain swelling, subdural haematomas and retinal haemorrhaging could all have other causes. Paediatric pathologist Jennian Geddes, now retired from the Royal London Hospital, published research in 2001 that suggested even light and innocent shaking of a baby鈥檚 head might constrict the spinal cord, stopping the child from breathing and precipitating the swelling of the brain and subsequent blood-clotting and haemorrhaging. This research was favourably received by the appeal court, and it helped overturn Faulder鈥檚 conviction.
Faulder claimed he had dropped N while trying to sit him in a pushchair, and the appeal court judges accepted that this might have been true. They also said it was impossible to rule out the possibility that such a fall had injured the nerves in the base of N鈥檚 neck, producing the trio of pathological symptoms later found in the child. N survived, and went on to make a full recovery.
But highlighting the complexities of the situation, the judges also dismissed a wealth of other medical and forensic evidence relating to so-called battered babies. Between 2001 and 2004, Geddes had published further research challenging the notion that the trio of intracranial injuries is a signature of physical abuse. Geddes鈥檚 team had gone so far as to propose an alternative, dubbed the unified hypothesis, based on the idea that while all three injuries could have a single cause, it might not necessarily be trauma. If breathing difficulties restrict the flow of blood to the brain, for instance, then lack of oxygen could cause the brain to swell, which in turn would rupture blood vessels in the brain and eyes. Carers accused of shaking babies soon began to use the unified hypothesis as a defence in court.
鈥淟ast week鈥檚 ruling could lead to the reopening of 88 cases in the UK and up to 15,000 child abuse cases in the US鈥�
However, the appeal court dismissed the unified hypothesis as grounds for a defence, particularly after another expert showed brain images of children who had supposedly died from the condition. The images showed the usual blood clots and eye haemorrhages, but not the swelling that was supposed to have given rise to them. 鈥淚n our judgement, it follows that the unified hypothesis can no longer be regarded as a credible or alternative cause of the triad,鈥� concluded the judges.
Geddes herself conceded that the unified hypothesis was nothing more than that, and in her view should not have been used as evidence that a child had died accidentally. But equally, she says, she is happy that the trio of symptoms seen in Patrick Harris should not be taken as de facto evidence of abuse. 鈥淚鈥檓 pleased the judges concluded that you can鈥檛 assume the triad to be due to abuse,鈥� she says.
Geddes also says that medical professionals are at last coming round to the idea that babies can be injured by means other than violent shaking. But in a further indication of how difficult it is to ascertain how a young child has died, the appeal court judges refused to accept a whole range of other scientific evidence, saying it was too uncertain to be of value.
For instance, Plunkett presented evidence for three of the appeals that the three symptoms could be caused by a young child falling from a height of less than 2 metres. He analysed 18 cases in which an infant had died after such a fall, usually in a playground (The American Journal of Forensic Medicine and Pathology, vol 22, p 1). But the judges were not convinced.
Two experts also submitted biomechanical evidence as to whether violent shaking could produce forces that could cause the pathological symptoms seen in some young children. But because the experts had reached different conclusions, the judges discounted this evidence, saying: 鈥淭his is a complex, developing and as yet necessarily uncertain area of science.鈥�
One biomechanical study published last month, and not considered by the appeal court, was conducted by Faris Bandak of the University of the Health Sciences in Bethesda, Maryland. Bandak concluded that a baby鈥檚 neck would fracture before further pathological symptoms such as brain swelling appeared. 鈥淭he signature of shaken baby syndrome is cervical spine injury,鈥� he says. 鈥淚t would be a setback for our understanding of the mechanisms of shaken baby syndrome to say that as a result of these two [appeal court experts] disagreeing, biomechanical evidence can鈥檛 tell us anything.鈥�
So how can we avoid miscarriages of justice in the future? The English appeal court says that further trials should continue to consider all the evidence before them and judge the merits of each case on the individual facts. It also recommends that medical experts ensure they give honest, unbiased opinions in court and do not become advocates for any pet theory. It also suggests that such experts convene before each trial to thrash out what they agree upon and what they do not, a tactic also being promoted in the US. The court also pointed to an obvious but important restriction on how far science can go in resolving such issues 鈥� it is impossible to conduct experiments on live children.
If the trio of symptoms cannot be relied on, it presents significant problems to courts everywhere trying to decide how an infant may have suffered a head injury. In the UK, the ramifications go deeper, as it also adds to a crisis regarding the role of expert medical testimony in the courtroom.
The British General Medical Council鈥檚 suspension of paediatrician Roy Meadow earlier this month for serious professional misconduct (see 鈥淎 trial of evidence鈥�) will make it more difficult to convince experts to study infant deaths, and for juries to accept their evidence. 鈥淚t will have implications right across medicine and means that anyone who in good faith makes a mistake will be struck off,鈥� says Alan Craft, president of the Royal College of Paediatricians and Child Health (RCPCH), based in London.
鈥淭here is an obvious restriction on how far science can go. It is impossible to conduct experiments on live children鈥�
The dwindling number of paediatric experts available to conduct post-mortems makes the problem worse. 鈥淭wenty-five per cent of the consultant posts in paediatric pathology in the UK are vacant,鈥� says James Underwood, president of the Royal College of Pathologists (RCP). 鈥淲e desperately need to recruit people.鈥�
But there are signs of hope. A joint RCP and RCPCH report last year recommended introducing a standard national protocol for investigating all unexpected infant deaths. And all the experts 快猫短视频 contacted agreed this should include information about the local family environment, medical history and a post-mortem. That should help experts build a clearer picture of why some babies die naturally from causes such as sudden infant death syndrome and help identify others who are victims of abuse.
And such information is desperately needed. 鈥淲e need more research,鈥� agrees Geddes. The unpalatable truth is that without it, thousands of adults may serve prison sentences for child abuse crimes they didn鈥檛 commit. The flip side, of course, is that real abusers and child murders will also go free, by exploiting the inability of forensic scientists and paediatric pathologists to prove beyond reasonable doubt what, or who, killed a child.
A trial of evidence
Earlier this month, British paediatrician Roy Meadow was found guilty of serious professional misconduct for his part in the trial of Sally Clark. Clark had already lost one child when her second child also died unexpectedly. She was charged with murder, and Meadow testified that the chance of two cot deaths in the same family was 1 in 73 million, an assertion that the British General Medical Council called a grave error.
In fact, evidence is increasingly showing that sudden infant death syndrome (SIDS) tends to run in families, with common medical, genetic or environmental causes, and experts generally agree the chances of two such deaths in one family are closer to 1 in 200. A study published in The Lancet (vol 365, p 29) last January by Bob Carpenter of the London School of Hygiene and Tropical Medicine and colleagues tracked 6373 families whose first child had died from SIDS. They found a second similar death in the same family was in fact 9 times as likely to be natural as inflicted.
In 2003, Clark was acquitted on appeal, as was Angela Cannings, also convicted of a similar crime. In April this year, Donna Anthony was freed by the appeal court, having been convicted in 1998 of killing her two children. That led the British Attorney General to say a further 28 cases of infanticide might need to be reviewed in case convictions had relied on flawed medical evidence.
Meadow鈥檚 story highlights the immense difficulties facing medical professionals trying to ascertain why young infants die unexpectedly. In 1977, Meadow introduced the term Munchausen syndrome by proxy (MSbP) to describe parents who invent or cause illness in their children. In 2002, he refined his definition of the syndrome, specifying that the parent must fabricate illness in the child, repeatedly take the child to doctors, and that the illness must disappear when the child is removed from the parent (Child Abuse & Neglect, vol 26, p 501).
But in the same year, a joint study by the Royal College of Paediatrics and Child Health, in London, and the UK Department of Health concluded: 鈥淭he term 鈥楳unchausen by proxy鈥� is no longer appropriate or helpful and should be abandoned.鈥� Many forms of child abuse show similar symptoms, and MSbP is difficult to define and it is over-used, said the report (快猫短视频, 20 September 2003, p 40). 鈥淥ne of the reasons for changing the name is that no one knew what it meant,鈥� says Richard Wilson, a consultant paediatrician at Kingston Hospital in London, who led the report. Instead, doctors should talk about fabricated or induced illness.
Identifying sexual abuse
Parents and relatives have also been wrongly accused of sexually abusing older children because of questionable expert medical evidence.
Tests once considered reliable for proving whether or not a child had been sexually abused have now been discredited, while paediatricians are split over whether other physical signs can reliably indicate that a child has been vaginally or anally penetrated.
In 1987, 121 children in Cleveland, northern England, were deemed to have been sexually abused after experts relied heavily on a test known as the anal dilation reflex. That involved lying a child on their side, pulling apart the buttocks for 30 seconds and observing whether the anus dilates beyond 1 or 2 centimetres. It was subsequently proved that many children who hadn鈥檛 been abused showed similar physical symptoms. 鈥淏ut by the time it was recognised as unreliable, many children had been taken from their families,鈥� says Mary Pillai, a consultant gynaecologist and forensic physician who works with National Health Service hospitals in Gloucestershire.
This same error 鈥� identifying signs of sexual abuse without supporting knowledge of what happens without abuse 鈥� happened repeatedly. 鈥淣o one was checking what constituted normal or 鈥榖aseline鈥� anatomy in non-abused children,鈥� says Pillai, who in April published a review of evidence used in sexual abuse cases (Journal of Clinical Forensic Medicine, vol 12, p 57). 鈥淯sually, you establish what鈥檚 normal first, but in this case it鈥檚 happened in reverse.鈥�
More recently, she says, doctors have found that severe constipation can cause anal inflammation, once thought to be another sign of abuse, while others are debating whether photographs can reliably reveal genital marks symptomatic of abuse.
