It is important to remember that Baby P was not killed by social workers. He died at the hands of sadists of unimaginable depravity, who were skilled at deception. That is where the greatest fault lies, and there must be real questions about why they face only a maximum of 14 years in prison. Had they treated an adult that way, the verdict would surely have been murder and the sentence would surely have been higher.

Yet evil is abetted by incompetence and aided by secrecy. All three elements were present in this appalling case. There is no doubt that some of the most evil people are the most persuasive, and the most adept at hiding their sadism. Yet it is not easy to understand how so many different care workers and agencies could be involved with a child who was on the at-risk register for nine months, and fail to see what we, from the outside, feel that we would have spotted immediately.

Bureaucracy and lack of communication are often given as reasons for child protection failures. But the tragedy of Baby P is not explained by bureaucracy. There is clearly a problem with social workers being overburdened with form-filling and not having enough time to see families, or to think. But Baby P was visited 60 times by health and social work professionals. They were not failing to talk to each other, according to the Haringey safeguarding report. But they were repeatedly failing to touch or examine him.

To have failed to protect him in the face of so much evidence seems inexcusable. But there is an explanation - given by Beverley Hughes, the Children's Minister. She believes that social workers jumped to the conclusion early on that Baby P's mother was inadequate but not a physical abuser. They then stuck to that in spite of the staggering and mounting evidence to the contrary.

Her explanation strikes a chord with me, because I have seen that pattern over and over in child protection cases with which families have asked me to get involved. Social workers deal almost constantly in uncertainty and ambiguity, with many families whose cast of characters and income is in constant flux and where children cannot be sure who will be in the house tomorrow. So it is particularly important for them to anchor themselves in a working hypothesis about what kind of risk might be posed to the child. But too often, that hypothesis becomes the official line without sufficient challenge. This is despite the changes brought in after the Laming review into Victoria Climbié, which encouraged agencies to work more closely together. Or perhaps it is partly because of it. When so many people are involved in case conferences, groupthink can be very powerful. And it can lead to bad decisions.

Eileen Munro, reader in social policy at the London School of Economics and the respected author of the Effective Child Protection handbook, makes the point that it is human nature to form a view based on first impressions, and stick to it. She says that this “has a devastating impact in child protection work, in that professionals hold on to their beliefs about a family despite new evidence that challenges them. It can be equally harmful whether they are over or underestimating the degree of risk to the child”. To counter this, she believes that social workers need to make their first impression explicit, and recognise that it will require a deliberate effort to think of evidence to the contrary.

This kind of psychological approach seems enormously powerful to me. Most people seem to focus on the legislation. But no amount of child protection legislation can ever substitute for properly trained professionals knowing how to think straight amid chaos, with strong leadership from their managers.

Camilla Cavendish, the Times columnist, won this year's Paul Foot award for journalism for her series of articles calling for greater accountability in family courts to prevent miscarriages of justice

Times Recommends

• Fundamentalist brew: faith and power
• Big guns wheeled out for the marginals
• Wanted: new merchants of light