Expert witnesses are on trial — again. This time their role has been put under the spotlight by two jurors who broke cover to speak out and question the conviction of a childminder for killing an 11-month-old baby in her care.

Keran Henderson, a mother of two and respected Scout leader, was convicted last month of shaking Maeve Sheppard so violently that she was left blind and and irreparably brain-damaged. The baby died days later.

Henderson, now serving three years in prison, has always protested her innocence. Yet medical experts at her trial lined up to say that the classic triad of symptoms that have been classified as shaken baby syndrome (SBS) were present. These signs — brain encephalopathy (swelling caused by accumulated fluid) and subdural and retinal haemorrhages (bleeding) — could have only been caused by violent shaking; the prosecution said that the baby’s neck ligaments were “over-extended”, indicating that her neck had snapped back and forth.

Henderson maintained that the baby had suffered a seizure as her nappy was being changed; defence experts attested that the injuries could have been days, weeks or months old.

Ten of the 12 jurors rejected the claim. But two felt so strongly that they have spoken out to say that they believe Henderson should not have been convicted. The foreman told The Times that the case, “relying on circumstantial evidence and forensic opinion based on evidential proof from other cases, should never have reached a court”. Up against the weight of medical and forensic experts, he added, the defence was on a “hiding to nothing”. But in his view: “Who caused the death, or whether anyone did, is not proved.”

A second juror, given the pseudonym Carol, also spoke out this week. She told BBC Radio 5 Live that she found the trial, with its conflicting experts, an “absolute shambles”. Jurors had the impossible task of trying to make sense of conflicting experts who “all seemed to contradict each other” even prosecution experts, she said. “It seemed as if they would bring an expert witness out for the prosecution, and if somebody said something that seemed to invalidate the allegation, they would bring another one out to emphasise the point.”

Like the foreman, she said that the evidence all seemed to be based on opinion, “possibilities and probabilities”, not “hard evidence”. How, with the experts themselves in dispute, could ordinary lay people make sense of it, she asked, and make “a right and fair decision”?

At one point the jury had to be given a lecture on how the brain worked. The pictures “were not quite right” but they decided that they would do for the purpose, she said. But key elements were “in the wrong place” — so “it was very hard to work out what they were trying to say”.

Then she describes a witness for the prosecution who said that she thought that the baby’s brain showed scars of old damage. “So we were confused as to what side she was on. Three days later they brought another witness who said that they were definitely not old injuries.” The first witness had urged “extreme caution” but the second had said “that’s not true. But it was all — ‘in my opinion’, all seemed to be based on possibilities, not actual evidence.” Nor could prosecution witnesses even agree on the degree of force used to cause the damage, she said, or the mechanism by which the injuries occurred.

Finally, Carol admitted that at times jurors had “struggled” to keep awake and “you wonder how much of it we missed, or how much we should have put emphasis on \ that we didn’t ... or put emphasis on things that we need not have done”.

At the end, she also felt “bothered” that the judge had — in her eyes — distanced himself after the verdict, making her “feel we had reached the wrong decision”; and she described the trauma of the verdict being delivered, when Keran broke down and sobbed and her husband, a former police officer, sarcastically clapped the jury.

The two graphic accounts give an unprecedented insight into how at least two of the jurors felt about their daunting task — raising a series of questions. Both felt inadequate to the responsibility given them. But was that their fault?

The consensus among lawyers is that a jury is perfectly capable of judging such cases; but that it is the task of lawyers, and experts themselves, to make their evidence comprehensible. John Cooper, a barrister who advised the family of Louise Woodward, the English au pair convicted of killing a baby in her care in the United States, also says that before trial, areas of dispute need to identified and honed, so that juries can focus on as “small an area of conflict as possible”. If there is no evidence but conflicting medical opinion, then — he argues — the prosecution should not be brought.

Sally O’Neill, chairman of the Criminal Bar Association, similarly believes that prosecutions should not be brought as readily as they are: that the test in such cases, where a parent or childminder stands accused of killing a child and there is no previous history of violence, should be applied with “much more rigour”.

Then there is the question of the scientific evidence itself. John Batt, a solicitor and part of the defence team for Sally Clark, the solicitor who died this year as a result of trauma and alcoholism after never getting over being convicted (although later cleared) of killing her two babies, urges a much tougher approach over SBS.

He cites the recent American case of Kathy Hyatt, charged with killing a baby in her care, but who, unlike Henderson, “walked”. Why? Because her lawyers challenged the whole basis of SBS and the judge accepted that it had no scientific validity, he says. In the UK, such procedural healings dealing with the admissibility of expert evidence are rare.

Two years ago, a similar challenge did come before the Court of Appeal in London in four cases. The established theory of SBS that had underpinned prosecutions for years was questioned. The leading exponent of the alternative theory, that there were different explanations (such as lack of oxygen) for the triad of injuries that did not involve shaking at all, was Dr Jennian Geddes. She gave evidence at the appeal; but admitted that her view was only that; a hypothesis, and added that she was sorry that it had been presented as fact in other cases. The appeal judges threw out the challenge and confirmed the approach of the Crown Prosecution Service in prosecuting SBS cases. The CPS firmly said: “The Geddes theory will no longer be used by the defence. The CPS will continue to prosecute appropriately, in order to protect infants — the most vulnerable victims in our society.”

But it is easy for jurors to be blinded with science. Sheer numbers or confidence will do it. Batt says: “Experts’ jargon baffles juries, but if the experts are confident and persuasive, a conviction will follow.” Abuse, he argues, falls broadly into two categories: those, as with Victoria Climbié, where inflicted injuries are so obvious; and those where the only evidence of abuse is medical opinion. When there is no direct evidence of killing, as with Keran, and the background of the accused makes it unlikely, why bring a prosecution?”

Meanwhile, there are calls for better qualified experts. Penny Cooper, a barrister and academic at City University, showed in a recent survey that most lawyers and judges do not bother to check the qualification of experts. A substantial number had also not bothered to undergo training as to how to present themselves clearly to jurors. There is no requirement for such training. Mark Solon, solicitor and founder of the training company, Bond Solon, says that often, experts think that their task is to impress their peers, rather than to enlighten the jury. Lawyers needed to be tougher in cross-examining experts, to elicit their qualifications and methodology as to how they reached their opinions; and judges needed to emphasise to juries that expert opinion “is just that — not fact”.

Tighter controls on experts’ standards and qualifications, fewer court disputes, more rigour over prosecutions brought, and tougher cross-examination of experts — all may help. But the two jurors this week have demonstrated the urgent need for a full-scale research into the “secrets” of the jury room. That will expose better than anything the failings in such cases. Without that knowledge, the innocent — as John Batt puts it — “will continue to pay the price”.

Times Recommends

• Are top lawyers worth their huge fees?
• Solicitor forced to retire claims 'ageism'
• Drug case failure may be last straw for SFO