Yet another woman was sent to prison last week, following expert evidence that she had shaken to death a baby in her care. Keran Henderson, a 42-year-old childminder, was said to have killed 11-month-old Maeve Sheppard, by shaking her so violently she was left blind and brain-damaged. The infant died in hospital a few days later.

The case has grim echoes of those of Sally Clark, Angela Cannings and Trupti Patel, all of whom were accused of killing their children only to be found innocent later. Clark, a solicitor, who was released from prison after serving three years, died last March as a result of psychological trauma and alcoholism caused by her ordeal.

At the Court of Appeal, two days after the judgment on Henderson, a retrial was ordered in the case of Barry George, the loner convicted of killing the television personality Jill Dando in 1999 with a single shot to the head. Expert testimony as to the significance of a particle of gunshot matter in his pocket is being challenged.

There has also been the recent conviction of the true killer of schoolgirl Lesley Molseed, 32 years after the event – and after Stefan Kiszko had served 16 years for the sexually motivated killing, even though medical evidence could have pointed out his infertility proved his innocence. Once again the review of the evidence threw a spotlight on the role of expert witnesses, whose testimony is often crucial in criminal cases but can be unreliable.

Our blind faith in scientific opinion makes us reluctant to question pronouncements by “experts”, but while the law requires everyone from plumbers to nurses to be trained, registered and checked, there is no such requirement for witnesses who may be pronouncing on matters of life and death in court.

A study by senior barrister Penny Cooper of City University in London, has shown that the majority of lawyers and judges do not bother to check the qualifications of experts they approach to bolster an aspect of their case. She also found a substantial number of the expert witnesses had undergone no training to understand their legal duty.

The disquiet this arouses has led to a clamour for legislation to require expert witnesses to be regulated. But how to do that without calling into question thousands of court decisions will not be an easy task.

There is already acute unease over the proliferation of parents convicted of causing cot deaths, shaking babies to death, or harming them by creating symptoms of fictitious illness.

Henderson, for instance, a mother of two herself, a long-term childminder and stalwart volunteer of her local Beaver Scout group, was sentenced to three years in prison for shaking baby Maeve so violently that she was left with fatal brain damage, despite the fact there was no evidence of any “grip marks” on the child, which would normally be expected to accompany such an action.

Her husband, a former police officer, has said she will appeal and hopes to create a campaign similar to that run by Sally Clark’s family, to try to prove his wife’s innocence.

Many character witnesses spoke up for Henderson in court and the family has dozens of supporters in their home village of Iver Heath, Buckinghamshire.

Some even believe her prosecution was only pursued because of the successful appeal by Roy Meadow, the expert paediatrician whose evidence led to the conviction of Sally Clark.

Following the Clark case, in which Meadow quoted a completely erroneous statistic suggesting the chances of Clark’s babies having died naturally were one in 73m, he was struck off by the General Medical Council (GMC) for misconduct. The Court of Appeal agreed he had acted in good faith.

In the meantime, Alan Williams, the Home Office pathologist who conducted post-mortems on Clark’s two infant sons, was less lucky. His appeal against a GMC finding of serious professional misconduct was rejected. Williams was accused of tailoring his diagnoses of the nature of the babies’ deaths to fit the police case against Clark.

The GMC is currently hearing a claim of gross professional misconduct against paediatrician Dr David Southall. The council has received evidence alleging that Southall falsified his curriculum vitae.

Southall’s evidence has figured highly in at least 50 criminal cases and possibly hundreds of family court cases held in secret, which have led to children being removed from their parents.

Questions of how frequently babies really are shaken to death, and indeed if it is possible to do so, have divided medical opinion for some years. There have, however, been up to 200 convictions annually for related forms of violence against babies and young children.

After Clark, Cannings and Patel, another bizarre case was overturned. Ian and Angela Gay, who had been convicted of poisoning their three-year-old adopted son with salt, were cleared when it was revealed the boy was suffering from a rare, and fatal, congenital abnormality.

Recently, the attorney-general ordered a review of almost 300 criminal convictions and 30,000 family court proceedings where children were taken into care. Only four were referred to the Court of Appeal. This, according to critics, was a function of the way the review was done, with authorities being asked to review their own decisions.

Social workers say the crusade to root out dangerous adults is to some extent a reaction to a previous era of regular criticism of their profession when children were left to die at the hands of their parents. Although some acknowledge the pendulum may now have swung too far, others are furious: “Do people think we spend all our time trying to break up families for no good reason?” said John Coughlan, a joint-president of the Association of Directors of Children’s Services. “In comparison with the volume of cases, the number of errors is tiny. We never rely on expert witnesses alone.”

Others argue that the opinion of expert witnesses is often the decisive factor. And as we have seen most recently with Barry George, it is not just child murder cases that have turned on such evidence.

Last year the Home Office took the unprecedented step of holding a disciplinary tribunal against Michael Heath, one of its most senior forensic pathologists: 20 charges against him were upheld. One man was subsequently cleared of murder, and numerous other convictions have been called into question.

Police forensic scientist Peter Ablett, who is now chief executive of the Council for the Registration of Forensic Practitioners, points out there are only three ways to prove a crime: a reliable eyewitness, a confession, or forensics. The advent of DNA technology and other advances in recent years has brought increasing reliance on forensics, yet only about 3,000 of the estimated 8,000 expert witnesses operating are members of the council and signed up to its code of practice.

He said many of those who are not are unaware that their duty is to give impartial evidence to the court, not to bolster the case of their paymaster.

City University’s Cooper, who is also a governor of the Expert Witness Institute, was concerned to discover during her research that not only have one in five experts undergone no training to understand this duty, but one in 10 was so arrogant they said they saw no need for it. “There should be a requirement for them to be trained, and there should be rules requiring judges and lawyers to consider their credentials before accepting them as expert witnesses,” she said.

Such a provision cannot come soon enough. A review is still going on of 700 cases in which bogus forensic scientist Gene Morrison gave evidence. Morrison, 48, from Manchester who was sentenced to five years for fraud in February, admitted he pretended to be an expert witness and bought his qualifications on the internet because it “seemed easier” than getting real ones.

For many of the genuinely qualified experts, legal work isa lucrative sideline, and if they are perceived to be able to “tailor” their evidence convincingly, the commissions keep flowing in. John Hemming, a Liberal Democrat MP campaigning about the misuse of medical evidence, says fees for a basic written opinion, based on reading through existing files, start at £4,000. If the expert concludes there is a case to answer, they attract court attendance fees as well.

“I have known experts get as much as £28,000 for one report,” said Hemming, who is lobbying for experts to be required to produce the scientific publications on which their opinion is based: “Unless we start using evidence-based evidence in court, we will get nowhere.”

Explore Law

• Columnists

• Law Panel

• Law Reports

• Student Law

Times Recommends

• UK legal industry faces loss of 10,000 lawyers
• Curious case of chocolate bunny trademark
• Divorce settlement granted after 22 years