What joy from @RichardJMurphy

\”Richard Murphy ?@RichardJMurphy 2m
I am seriously wondering whether the so called #nhs crisis is simple statistical failure to understand variation from the mean is inevitable\”

I am seriously wondering whether the so called inequality crisis is simple statistical failure to understand variation from the mean is inevitable
I am seriously wondering whether the so called labour share of GDP crisis is simple statistical failure to understand variation from the mean is inevitable
I am seriously wondering whether the so called deficit crisis is simple statistical failure to understand variation from the mean is inevitable

One could go on for hours here……

 

12 comments on “What joy from @RichardJMurphy

  1. And I’m wondering if the so-called funerals of those patients who died unnecessarily is simply a statistical failure to understand that death is inevitable.

    Don’t Tim. If you went on he would call you a troll.

  2. Another entry on his rapidly-expanding CV: statistician.

    His response to criticism of his “fair tax mark” methodology suggests that numeracy may not be his strong point.

  3. I think that he has a valid point. A lot is done these days to deal with things which are above average, without stopping to consider that the whole point of an average is that you’ll always get some things above it and given the shape of a bell curve you’ll often get the occasional thing well above it.

    I’m thinking partly of the way speed cameras reduce accidents, because they get put in places which had one-off spikes in accident rates that then regress to the mean; and I’m also thinking of a cot death case (I think it was) where great play was made of the fact that there was a one in a million chance of something happening and so it can’t possibly have been an accident – apart from the fact that if there are a million babies born a year then a one in a million event is par for the course.

    In this NHS Trust thing I have no idea whether the difference from the average is at all significant in size or likelihood. But I’d be very slow to say that any trust which has more deaths than the average is ipso facto doing something wrong.

  4. I am seriously wondering whether so called tax dodging companies are simple statistical failure to understand variation from the mean is inevitable

  5. There’s a lot of this about.

    If a hospital was deliberately killing patients suffering from illnesses that weren’t life threatening, it would appear in the lower tail of a bell-shaped distribution and hospitals that were treating them and then sending them home would appear elsewhere in that distribution.

  6. Pellinor,

    The cot death case you are (probably) referring to was actually at least three cases of women banged up for murder because more than one of their children died a cot death. The “expert” “witness” who testified against them was Roy Meadows, now quite rightly disgraced since the obvious schoolboy error in his “reasoning” was revealed and the women were freed. Took years, though: the police, CPS, judges, and juries were completely convinced by his utter bollocks.

    The point with the cot death cases is that, aside from the total failure to understand basic mathematics, there was not a shred of evidence against the accused. The key difference with these NHS stats is that we have particular examples, not just aggregate stats. And the examples are appalling, and cannot be excused as merely below-average care. They constitute negligence, gross misconduct, a total disregard for the Hippocratic Oath, and arguably, in some cases, crimes.

  7. Oh, if there’s a lot of other evidence it’s a different matter. I’ve not been following the story in detail, though – I was just going on the bit about statistics.

    My understanding was that people had been extrapolating numbers from death rates to say that, in essence, if you pick the hospitals with the highest death rates and assume they should have had average death rates, then all the excess deaths were avoidable. I’ve seen a few pieces along those (logically dubious) lines, but as I say I’ve not really followed after that.

  8. Following Pellinor’s point and Squander Two’s points: let’s say you have a hospital with worse than average death rates – could be bad luck/ mere variation from the mean but clearly it’s worrying.

    So you investigate it and find concrete examples of bad treatment. But what if you haven’t investigated the good hospitals in the same detail? How do you know you wouldn’t have found examples of bad treatment there? Or do you just look at the numbers, and say – “we need to improve (or close) hospital X.”

    I spent 20 years dealing with professional negligence cases, some of which were really fraud. No medical ones, so I’m not claiming any expertise. But I am pre-disposed to thinking that some cock-ups/bad apples are inevitable. FWIW I think we should be more worried about any delay in discovering/reporting.

  9. Let’s face the fact that Murphy is uninterested in facts that do not suit his latest rant.
    The NHS review was instituted by someone who DID understand variation from the mean is inevitable and asked someone to review the 14 NHS Trusts with the worst statistics BEFORE passing judgement. Eleven had really serious management failings that caused high death rates; three were deemed unlucky in that it was not obvious that they caused patients to die unnecessarily.
    So reality is at 180 degrees to Murphy’s tweet (or perhaps Murphy’s tweet is at 180 degrees to reality)

  10. John 77, that’s helpful. I hadn’t spotted that three of the “worrying” cases were probably just bad luck.

Leave a Reply

Name and email are required. Your email address will not be published.

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>