This morning I got an e-mail from Frank Lockyer, the retired police superintendent and father of Sally Clark. His daughter spent three years in jail for murdering two of her sons, during which time she suffered untold torments. She was exonerated later and released, but she never recovered from her treatment at the hands of the British justice system - a system she had once served as a solicitor. She died in 2007.

Mr Lockyer was writing about a recent decision of the Press Complaints Commission to clear me of a complaint made under clause 1 of the Code of Practice (accuracy) by the paediatrician Professor Sir Roy Meadow. He was villified by the media for evidence he gave at Sally Clark's trial and that of Angela Cannings, another mother who was jailed and subsequently released. Professor Meadow had objected to a comment I had made suggesting that he had gone beyond his remit and given evidence that “led to the jailing of innocent people” - partly by presenting statistical evidence of the likelihood of cot death when he was not a statistician.

He felt that this was misleading, for three main reasons. First, he was only one of several witnesses in these cases; second, appeal judges had played down the importance of his statistical evidence in the Clark ruling, and third, we cannot know how much weight the juries actually gave his evidence.

The PCC decided not to uphold the complaint. It took into account that my views were set out in what it felt was an opinion piece and not presented as indisputable fact - and that Professor Meadow had rejected The Times's offer to print a letter from him, setting out his position. It is not an episode I am proud of - it is sobering for any journalist to be accused of inaccuracy. But that he was supported by other paediatricians in bringing the complaint reflects a continuing gulf between the medical professions and the press over what we should expect of expert witnesses.

Mr Lockyer writes that he would never have complained to the General Medical Council about Professor Meadow if he had apologised or admitted that he might have been mistaken. This is a different question from whether he was made a scapegoat when the courts themselves should have been more sceptical. Mr Lockyer wanted accountability, not revenge. He is concerned about what he considers to be the reluctance of so many members of the Royal College of Paediatrics to admit that evidence might sometimes be wrong.

He is not alone. This week I spoke to a Welsh woman whose son has been forced to go on seeing his father although both she and his school fear that dramatic changes in his behaviour are the result of abuse. An expert psychologist has dismissed their concerns by stating - astonishingly - that teaching in Welsh could cause retardation in some children. Despite there being no apparent research to back this up, the court served a penal notice that means that the mother will go to jail if she attempts to protect her son from unsupervised visits by her ex-partner, who she believes is an abuser. Welsh politicians have expressed outrage - Rhodri Morgan, the First Minister, said last month that he “would not dignify that person with the title of ‘expert'”. But the decision cannot be challenged, even though the expert has refused to indicate what his conclusions were based on.

The worst part of this story is not the expert's “evidence”, or the court's apparent acceptance of it. It is that the court has twice denied the mother the chance to call another expert in her defence. This is a common problem if the letters I get from parents are anything to go by.

After overturning Angela Cannings's conviction in 2003, Lord Justice Judge declared that no one should ever go to prison again solely on the basis of expert witness evidence. The criminal law was changed as a result. But in family courts, many decisions are still made on the basis of evidence from psychiatrists, psychologists or doctors, who often take the view that a mother is unstable, sometimes without cross-examination. Too many family courts are being run by experts, rather than judges.

If the courts are not prepared to challenge “expert” evidence, they should surely allow others to do so. In theory, experts are supposed to be independent professionals who have a duty to help the court to come to the right decision. In practice they are often hired guns, paid by local authorities that choose people they know will be a “safe pair of hands” - people they have used time and time again.

In December the Lord Chancellor and Justice Secretary Jack Straw took the heroic decision to open the family courts to the media and to end the gagging of parents who wish to speak out about their cases. In doing so he has struck a huge blow for justice. But blind faith in experts still poses a problem.

The opening up of the family courts will allow the media to scrutinise experts and their evidence - at least when journalists turn up. But although the Straw reforms will let parties disclose court documents to outside experts without a judge's permission, they will still not be able to call those experts in their defence without the court's say-so. Unless defence lawyers are savvy enough to exploit this, there may still be miscarriages of justice.

It is perhaps not surprising that many experts have an inherent tendency to believe that they are right. That is human nature - although you would expect members of the medical professions to be more conscious of scientific complexity and uncertainty. Courts should be more sceptical. If they will not challenge experts, they must let more parents do so.

Camilla Cavendish won the Paul Foot Award for campaigning journalism for her investigation into injustices in the family courts

Times Recommends

• In the real world, the public sector must pay
• Playing politics in a land of double Dutch
• Be like the plinth people: not set in stone