One of life’s difficulties, I have found, is that it keeps throwing up questions to which there is no indubitably correct answer. This means that the exercise of judgment is perennially necessary: and there is hardly a moment’s respite from this burdensome imperative. Alas, where there is judgment there is error, or the possibility of error. No one can be right all the time.
Of nothing is this truer than the vexed question of child abuse. Not to see it where it exists has terrible consequences for the child; to see it where it does not exist has terrible consequences for the parents or the others accused of it.
I have seen incontrovertible evidence of things done to children so terrible that, though no babe-in-the-wood when it comes to the human capacity for evil (having travelled through several countries in the throes of murderous civil wars), I should not previously have thought it possible for people to do them in conditions of peace and prosperity.
On the other hand, I am aware of how easy it is to make false allegations. Rare metabolic conditions may mimic factitious disease; there is even a very rare skin condition that gives lesions that are extremely difficult to distinguish from cigarette burns, which are a common form of abuse. Moreover, it has to be remembered that evidence of abuse does not by itself make any particular person guilty of it.
The subject of child abuse arouses deep passions, as two prominent British paediatricians, Professor David Southall and Professor Sir Roy Meadow, know to their cost. Both are the object of hatred of an exceptionally virulent and almost uncontrolled kind, far worse than if they had actually decimated the countryside or poisoned the wells.
Professor Southall first became the object of hatred when he videoed mothers suffocating babies who had had repeated apnoeic attacks (episodes when breathing stops momentarily) without any inherent physiological or pathological cause. He did this before the average Briton was videoed 300 times a day as he went about his business, that is to say in the days when filming people without their knowledge or approval seemed to be an intrusion of their privacy. However, most people believe that, in this case, the end did justify the means.
There is strong evidence that Professor Southall, a man with a worldwide reputation, has been the victim of a concerted campaign of vilification and even persecution, to which the General Medical Council may have been the unwitting accomplice. As a recent article in the British Medical Journal pointed out, the GMC has received a torrent of vexatious complaints against him, many from persons whose children he had never treated. It has preferred the evidence of a complainant about what was said on a distant occasion to the evidence of Professor Southall himself and that of a social worker also present at the time, and struck him off notwithstanding that the standard of proof necessary to do so was that of being beyond reasonable doubt.
According to the BMJ, the GMC, in two previous proceedings against him, has disregarded conflict of interest among both the prosecution witnesses and the prosecution lawyers; and finally it has taken 11 years to find that he (and two other paediatricians) had no case to answer in a complaint brought against them regarding a clinical trial that they had conducted.
Eleven years to find that there was no case to answer! Admittedly, the GMC’s task was not made easier by the almost kaleidoscopic nature of the allegations made against Professor Southall; but, again according to the BMJ, there is evidence that the GMC was far too friendly to, and almost in collusion with, the complainants. This is very damaging when a body acts both as judge and jury, the very opposite of natural justice and more like the proceedings of a Star Chamber.
The case of Sir Roy Meadow is even better known than that of Professor Southall. It was the infamous statistic of one in 73,000,000 that got him into trouble. He claimed at the trial of Sally Clark that the chances of two babies dying of genuine Sudden Infant Death Syndrome (Sids) in a household such as hers was one in 73,000,000. He arrived at this figure by simply squaring the chances of one baby dying (one in 8543), which was the figure in a recent, as then unpublished, survey.
Now there are several things wrong with this. First, it disregards the possibility (actually, the certainty) that variables other than those allowed for in the calculation affect the chances of repetition in a household.
Second, it is a useless statistic from the court’s point of view. To see why, imagine that the defence had entered the following argument: only one in 10,000 women of Sally Clark’s demographic profile kills her infant. Therefore the chances of her having killed two children are one in 100,000,000, smaller even than Professor Meadow’s figure.
It is obvious that no one would, or at any rate should, either convict or acquit on the basis of these statistics.
The real statistic, it seems to me, should have been the proportion of households in which two unexpected infant deaths occur in which those deaths were unnatural. In fact, this was mentioned in the court: Professor Emery, who had studied this question, had come to the conclusion that it was about a third, that is to say a proportion that would have comforted neither the prosecution nor the defence. From the evidential point of view it is worthless.
As it happens, the defence in O.J. Simpson’s trial made use of what is known as the prosecutor’s fallacy with great effect. It was entered that Simpson had been violent to his wife; the defence countered that only one in several thousand men who is violent to his wife actually kills her, and that therefore the fact that he was violent towards her before she died was of not evidential value. But again, the wrong statistic was used. Of murdered women whose consorts were violent to them, the overwhelming majority were murdered by their violent consort; therefore, the fact that he had been violent towards her before her death was of evidential value, though no one would convict on that basis alone.
One of the grounds for the first appeal in the case was Professor Meadow’s allegedly misleading statistic. The appeal judges dismissed this round of appeal for a number of reasons. First, it was challenged by an expert for the defence, Professor Berry, professor of paediatric pathology. He told the court that it was wrong simply to square the risk of one child dying of Sids to reach the chance of two children dying of Sids in the same household, and drew attention to the caveat against doing so in the study upon which Professor Meadow relied.
The judge, in his summing up, properly drew attention both to Professor Meadow’s and Professor Berry’s opinion, but left the jury to decide which opinion it preferred. Moreover, he specifically warned upon reliance on statistics:
I should, I think, members of the jury, just sound a note of caution about the statistics. However compelling you may find those statistics to be, we do not convict people in these courts on statistics. It would be a terrible day if that were so... it is of course necessary for you to have regard to the individual circumstances relating to each of these two deaths before you reach your conclusion...
The judge devoted about 4 per cent of his summing up (if I have my statistics right) to the question of Professor Meadow’s statistics, so unimportant did they seem to him.
Furthermore, it was, or ought to have been, abundantly clear that the figure was irrelevant to the whole case. It was not part of the defence argument that the children had died of Sids: no medical expert, whether for the defence or the prosecution, said that they had. The defence case was that they die d of natural causes and that the prosecution had not proved that they had died unnatural deaths. The infamous statistic was therefore a red herring, and perhaps the judge in his summing up should have said so more firmly, though he did say several times that the babies had not died of Sids, a conclusion quite independent of the statistics. If a jury is not capable of seeing the irrelevance of a statistic related to a cause of death claimed neither by the prosecution nor the defence, then doubt is cast on the viability of the whole jury system.
The subsequent case against Professor Meadow at the General Medical Council, that led to him being struck off before he was reinstated by the Court of Appeal (a decision itself appealed by the GMC, which it lost by 2:1), was that by wandering into the field of statistics, he was straying beyond his expertise. He failed to warn the court that he was doing so, and thereby misled the court; finally, in misleading the court, and even though he was acting in good faith, he was guilty of serious misconduct worthy of erasure from the medical register.
Now Professor Meadow never posed as a statistician. The appeal court said ‘Professor Meadow (who does not claim to be a statistician)...’. The fact is that all doctors use statistical and probabilistic thinking — of course not always correctly — in their daily work. This is proved by the old clinical saw that common diseases occur commonly. A patient comes with a pattern of symptoms that is 500 times more common in disease A than in disease B; therefore, the doctor diagnoses disease A. Often, of course, the reasoning is much more finely balanced than this.
If expert witnesses were prohibited from uttering their interpretation of statistics in court because they were not statisticians by training, the courts would act even more slowly than they do. Moreover, if having done so was a justified grounds for striking off Professor Meadow, it would also be justified grounds for striking off Professor Berry, a self-evident absurdity. After all, he was a pathologist, not a statistician; the mere fact that he was right and Professor Meadow was wrong was no defence against his failure to warn the court that he was not a statistician.
Incidentally, when the GMC struck Professor Meadow off, it made a serious statistical error in its own calculations of the statistical chances of two babies dying of Sids in the same family. Quis custodiet ipsos custodies?
It is the essence of the adversarial system that conflicting views should be tested by the trial process. It is inevitable that experts are not infallible, and furthermore that they should disagree. Mistakes are unavoidable. It is rarely that an expert leaves the witness box thinking he has put his points as well as he might. He has to think on his feet and consider arguments that he has never heard before. Recently, I failed to explain as clearly as I might the difference between relative and absolute risk (though perhaps I should not even have tried, not being a statistician). I was also involved in a case in which I thought an opposing expert was mistaken as to a point of pharmacology, and was obdurate in refusing to accept it, but it did not occur to me that he was guilty of anything other than error, no doubt with a bit of vanity thrown in. But to demand of experts that they have no human characteristics, and leave all qualities behind them except rational calculation, is to demand the impossible. It is something to be aimed at but it will never be achieved.
What will people make of the Southall and Meadow affairs in 50 or 100 years, that have resulted in an acute shortage of people to undertake the difficult and harrowing work of child protection, in its wider historical and cultural context? I think they will see them as a small part of a pattern of the destruction of the independence not only of the medical profession, but of all professions, that is part of the great bureaucratic Gleichschaltung of British society. The populist media outcry is an important part of the process by which the population is induced to run to authority to protect them from — well, authority. And, like all attacks on authority, this one results in a transfer of authority rather than its destruction. Personally, I doubt that it is for the better.