Look here:
www.ted.com/tedtalks/index.cfm?flashEnabled=1

There’s lots of other interesting stuff there too.

Geoff

I had such a good time chatting last night that I thought I’d pay you the highest form of compliment I am capable of, namely by reaching down from my ivory tower to attack you for promoting the sort of myth you so rightly scorn when done by others.

The crime of which you are accused is in the very first line of the above article:

“There once was a time when your biggest worry, as a paediatrician, was being lynched by a herd of illiterate tabloid readers”

A good introductory gag, no doubt. Made me smile, in fact. But it is a gag based on a story which has done a good bit to fuel the idea the that once someone is accused of pedophilia the average British community will always hound them to their dooms, no matter how absurd or false the accusation. One gold star for a good choice of investigative topic for Brendan O’Neill, one trip to stand in the naughty rumour circulating corner for Ben

news.bbc.co.uk/2/hi/uk_news/magazine/4719364.stm

ta ta,

Tom

en.wikipedia.org/wiki/Roy_Meadow

There is also a press release issued after Sally Clark’s first (failed) appeal trial by the Royal Statistical Society, commenting on evidence consisting largely or entirely of statistics:

www.rss.org.uk/docs/Royal%20Statistical%20Society.doc

Finally, for a recent medical take (not too hard to read) on the whole issue of child protection, which sheds light on why many paediatricians have some sympathy with Roy Meadow, see the review in the Journal of the Royal Society of Medicine at:

www.jrsm.org/cgi/content/full/99/1/6

My twopenneth, which I’ve thought since first following the trial, is that Prof Meadows is guilty of serious misconduct and should be struck off..
Whether his “evidence” was crucial or irrelevant is not the point. Neither is the accussed’s guilt or otherwise. Nor the incompetence of the defence in not questioning the testimony.
He was in a position of authority, intended to be trusted by the jury as such. He had a duty to be sure he knew what he was talking about, yet he clearly did not. Forget even the prosecutor fallacy; what he asserted was clearly wrong to anyone with an O level in maths. If he did not understand basic statistics, as he clearly did not, then he had no right to quote them, indeed he has a duty NOT to. In short, he strayed outside his area of expertise, but allowed others to believe he had not done so, whilst giving potentially crucial testimony. THAT is what is unforgivable.
The blatant error which was obvious to me from the start is that, in the real world, outside theoretical coin tosses etc, the fact that someting has occurred once, can strongly influence the chance of it happening again, (whether because of genetic factors, environmental causes, a child murderer living next door, etc ). Prof Meadows has to assume SIDS is a totally random event in order to ignore these factors, in order to draw his conclusions. (Unless we now know definately, irrefutable and without exception, the cause of SIDS. And unless you doctors are keeping it to yourselves, I don’t think we do, do we?)
If the statistic was supplied by a statistician, then it is surely his role as an expert witness, to factor in the real world to the analysis, not to ignore it. This is a schoolboy error and is frankly, inexcusable.

(I should add that everything I know about the case comes from news reports, so if I myself am talking B*****s about the detail, I stand to be corrected)

BTW, fascinating original article. The “Prosecutors Fallacy” had never occurred to me before. (Unless I’ve just described it in a different way above. But I don’t think I have…)
Keep up the good work, Ben. I’m right behind you all the way, even if this – my first – posting is actually disagreeing with you about the scapegoating aspect. But then, I’m not a doctor

Interesting points about parents, Robert.

if you take Southall’s Pediatrics paper at face value, the majority of the parents who his video surveillance found physically harming children were diagnosed (subsequently) by psychiatrists as having “personality disorders”, although this is a pretty catch-all term (includes a wide range of things, and of varying degrees of severity). From the histories given it seems clear many of these parents themselves had grim childhood personal histories, perhaps not unexpectedly.

Of course, the kind of serious premeditated life-threatening harm his paper was discussing is a long way from (e.g.) a mum-at-end-of-tether slapping a screaming two yr old in a flash of anger after four sleepless nights. The abuse in the Southall paper’s cases was implied to be repeated, was rather deliberate, and was not associated with any triggering event (like screaming).

I do agree with you generally about the strain that having kids put on everyone, though. Given the pressures I am never that surprised that some parents do hit children. The tragedy is that most parents who don’t cope well could probably be helped effectively – when one watches the TV shows like “Little Angels” and “House of Tiny Terrors” that (child psychologist) Tania Byron does, the results do seem to be quite startling, and show what a bit of sensible advice and support can do for parents who often appear to be at their wits’ end. As a society I think we in the UK are not very good at supporting parents, and the ones that most need the support probably are the hardest to reach, for all sorts of reasons including social exclusion.

Dealing with the kinds of cases described in some of the published stuff on parents harming kids really must be a truly awful job. I have one friend who is a senior child protection social worker, and another who is a community pediatrician, and I wouldn’t do either of their jobs for any money there is.

Interesting points about parents, Robert.

if you take Southall’s Pediatrics paper at face value, the majority of the parents who his video surveillance found physically harming children were diagnosed (subsequently) by psychiatrists as having “personality disorders”, although this is a pretty catch-all term (includes a wide range of things, and of varying degrees of severity). From the histories given it seems clear many of these parents themselves had grim childhood personal histories, perhaps not unexpectedly.

I do agree with you generally about the strain that having kids put on everyone, though. Given the pressures I am never that surprised that some parents do hit children. The tragedy is that most parents who don’t cope well could probably be helped effectively – when one watches the TV shows like “Little Angels” and “House of Tiny Terrors” that (child psychologist) Tania Byron does, the results do seem to be quite startling, and show what a bit of good advice and support can do for parents who often appear to be at their wits’ end. As a society I think we in the UK are not very good at supporting parents, and the ones that most need the support probably are the hardest to reach, for all sorts of reasons including social exclusion.

Dealing with the kinds of cases described in some of the published papers on parents harming kids really must be a truly awful job. I have one friend who is a senior child protection social worker, and another who is a community pediatrician, and I wouldn’t do either of their jobs for any money there is.

Must just share this www.blacktriangle.org/blog/?p=1485 re my own professional regulatory body.

And the disgrace of Professor Sir Roy Meadows does not mean that Sally Clark was not charged with a crime that she had committed. For that matter, I gather that Dr WIlliams’s final opinion was that one of the Clark children was shaken to death and one was smothered, the former being the case with the mishandled evidence of infection. That view – for whatever it’s worth – doesn’t support a similar genetic or accidental cause for both deaths,

As for future inquiries and which set of other cases should be used for any comparison… I think the set of parents with multiple infant deaths will include multiple child murderers as well as multiple natural death victims, so you couldn’t tell the difference, whereas the set of families with one death will proportionately include fewer murderers – I assume that overall there are fewer murderers – and there wouldn’t be much difference between parents who have had one child, and parents who, due to a death, have had one child at a time. But I expect families of more than one child at one time to have measurably different outcomes. Of course that will partly because, purely statistically, these children were lucky enough to be born without either a disease or a murderous parent that would do them in before the second child arrived.

I also think that parents are not from their own birth either evil people who will inevitably murder their child or normal people who will certainly not do so. Caring for an infant is incredibly stressful, women typically also have a hormone roller-coaster to deal with, and your sanity is under terrible pressure. When and whether this should or should not be a valid defence against a charge of harming a child, I consider to be lower on the agenda than arranging as much support as possible for the family so that these tragedies don’t happen.

it has changed my understanding of the role of the GMC and such professional bodies in general which as a member of the medical register is quite important!

In Southall’s case he was/is famous for his work using covert video recording in a group of children who had a history of “unexplained life-threatening events” and where the doctors suspected the parents might be responsible. In a celebrated study Southall showed using the video surveillance that in a number of these cases a parent was indeed harming the child.

pediatrics.aappublications.org/cgi/content/abstract/100/5/735

The upshot of this work was that no matter how unthinkable it was that parents might intentionally harm their own small children, to the point of suffocating them, it clearly did occur.

Of course, that it CAN happen, and sometimes does, doesn’t tell you HOW LIKELY it is to be occurring. Southall and his co-workers made this point very explicitly at the time, although you might conclude later events showed he lost sight of this.

So we are back to the begining with the questions “What are the real relevant statistics?” and “Who should be capable of working out how they are derived?”

PS It’s amazing what you can find on the net – the full High Court Judgement on Professor Southall vis-a-vis the Clark case is at:

www.bailii.org/ew/cases/EWHC/Admin/2005/579.html

Although you want to avoid one of the majore flaws in Bayesian reasoning, the use of arbitrary subjective probabilities, which are particularly open to abuse, as in:
en.wikipedia.org/wiki/Regina_versus_Denis_John_Adams#Use_of_Bayesian_analysis_in_the_court

“My recollection – no doubt twisted by crap media reporting was that Meadow saw the news story on Tv and approach the police of his own volition saying he was convinced it was murder, with no evidence beyond the TV story. i thought that was why he got struck off.”

You’re confusing Roy Meadow with Professor David Southall, who is the paediatrician this story applies to, steve – see:

news.bbc.co.uk/2/hi/health/3538320.stm

bmj.bmjjournals.com/cgi/content/extract/329/7462/366-a

Southall was not struck off, but was reprimanded and had restrictions placed on his work.

But as RS said back at 5, we need to be a lot more Bayesian. At best, that is what a criminal trial should be – lots of small pieces of evidence which, on their own, may not have huge positive or negative likelihood ratios, but put together give us a pretty good negative or positive predictive value. Omitting evidence like the possibility of infection – which would have had a reasonably strong negative likelihood ratio wrt foul play – was IMVHO at least as bad as attempting to quantify a risk far better expressed as “very rare”.

———–

Though I do see that there is potential for expert witnesses to give biased opinions, one may also say that both sides can get their own expert witness. Moreover, the court could also be hiring its own witness for the case to make sure the information is valid. Being an expert witness also mean that your expertise is your credibility and qualifies you to give your testimony in the court. If one fails to do that or give biased opinions they will be punished legally and loses their credibility as an expert witness – an example would be Sir Roy Meadow.