And where the “welfare of the child” is used to gag them from discussing the case ever after. I live in that country.

In that country, a mother who has had her three children taken away broke the gagging order on BBC One on Monday night. She was brave to do so. For she is in grave danger of also losing the baby she is due to give birth to in two weeks’ time. Nicky Hardingham’s three children were removed because one had unexplained fractures in his bones. Two paediatricians categorically stated that the boy could not have the brittle bone disease that can cause such fractures, even though the disease is in four generations of Mrs Hardingham’s family. Medical geneticists have since said that the boy could indeed have inherited the disease, although Mrs Hardingham herself has not suffered fractures. The police dropped all charges against the Hardinghams. But the court’s decision to take away her children was final.

I won’t name those two doctors. It is partly the fault of my own profession that doctors now have so much more to fear from failing to spot signs of abuse — of missing another Victoria Climbié — than they do from diagnosing abuse where it does not exist. Many social workers and judges are so terrified of false negatives that they hasten to false positives for which they will never be accountable. Yet this alone does not explain why professionals make categorical statements in situations where they should merely be posing questions.

In classic child abuse cases, there is often a family history of abuse and vile evidence of assault, such as bruising and cigarette burns. Yet there seems to be a growing number of cases in which these factors are totally absent. The metaphyseal fractures suffered by Mrs Hardingham’s son are said to be caused by an adult violently twisting little limbs. But it would be virtually impossible to do that without bruising the skin. And in case after case, there is no bruising.

Other parents are accused of violently shaking their children, because doctors have found haemorrhages in the brain or the eye. Again and again parents are told that they have shaken their babies at a velocity of “30 or 40 miles an hour”. Yet the evidence is flimsy. Lorraine Harris was cleared after serving a jail sentence for shaking her four-month-old to death, when she proved that he had a blood disorder. She has little hope of ever seeing her other child again. Joe Wainwright was jailed for shaking his baby son, even though the scan used in evidence was taken after the hospital had smashed the boy’s trolley into a concrete pillar. He has not been given leave to appeal. The list goes on.

At a conference in London two days ago, I watched American scientists rip the medical canon on child abuse to shreds. Kirk Thibault, a biomechanical engineer from Pennsylvania, has demonstrated that even a child falling over its own feet can sustain much more serious head injuries than one being deliberately shaken. So perhaps the courts should have given more credence to all those parents who pleaded that their daughters had slipped on the floor. Marvin Miller has found that low bone density is responsible for many infant fractures, and that this can be caused by lack of movement in the womb as well as by a brittle-bone gene. Low bone density does not show up on the conventional X-rays used by the radiologists who pronounce on abuse. Patrick Lantz has studied 890 bodies to find that a whopping 15 per cent of adult corpses, and 19 per cent of those under the age of 1, have the haemorrhages that the textbooks cite as proof of abuse. John Plunkett, the forensic pathologist from Minnesota who has helped to free more than 100 wrongly accused parents in America, neatly summed up the problem. “If all you have is a hammer,” he said, “everything looks like a nail.”

Most of these men started out on the other side of the argument, as doctors testifying against parents. At some point each seems to have become alarmed by the number of allegations being made in families with no history of abuse, and stunned by the flimsiness of the research. The belief that metaphyseal fractures are incontrovertible evidence of abuse, for example, is based chiefly on one 1986 study by the American radiologist Dr P. K. Kleinman. He found such fractures in four infants who were known to have been abused, and none in one who had not. How could any court convict Nicky Hardingham, and many others, on the basis of such an unscientific sample? Yet they have.

At Tuesday’s conference I met a young couple looking completely shell-shocked. Social services took away their child last month. It is hard to escape the feeling that any one of us could be in their shoes.

I do not know quite how we have created a situation where abuse has become the default diagnosis in the face of the unexplained. Or why social workers and judges have come to rely so heavily upon medical theories that are presented as fact. The only way to avoid miscarriages of justice must be to throw open these decisions to thorough scrutiny. Children in Scotland, Canada and New Zealand have not been damaged by family courts being open to the press. The longer these processes remain secret, the more danger there is that children will suffer from misguided decisions.

It is too late for Nicky Hardingham to get her three children back. But if her next baby is taken away, we must suspect that the system has learnt nothing. In telling her story she has at least begun to expose how the system is skewed. There are so many like her who dare not speak. We should not stand for that in this country.

Times Recommends

• Exit Clinton. But what a show next!
• Angel of the South: the wrong horse
• Cannabis: who cares if it's B or C?