The Guardian today on the safety of the Lucy Letby convictions

[Kittybythelighthouse]

This article was apparently months in the making but it was delayed by the reporting restrictions https://www.theguardian.com/uk-news/article/2024/jul/09/lucy-letby-evidence-experts-question

“A Guardian investigation has interviewed dozens of these experts and seen further evidence from emails and documents. Those raising concerns include several leading consultant neonatologists, some with current or recent leadership roles, and several senior neonatal nurses. Others are public health professionals, GPs, biochemists, a leading government microbiologist, and lawyers. Several of those still working in the NHS have asked to remain anonymous, fearing the impact if they are named.

These experts said they were acutely aware of the suffering of the families involved and did not want to reopen their trauma, but were so troubled they felt compelled to become involved”

I think in a case like this, with complicated, largely circumstantial evidence and requiring the jury to understand statistics- which is a subject that is often not intuitive and can go against how you expect things to work- that evidence from expert statisticians should be a compulsory part of the trial. If that isn't the case, then the legal system needs to be changed.

When you are potentially sending someone to prison for the rest of their life, every piece of evidence needs to be thoroughly looked at, dispassionately, and understood. I have no idea why anyone would not see how vital this is for justice to be done. Just because historically, this hasn't been a requirement doesn't mean it shouldn't be.

Thank you for answering my question so clearly. What is meant by "more heat than light"?

I would appreciate if anyone answered the question I asked earlier about what exactly is wrong with the prosecution case. But Yes I'm happy with this case.

I have been involved in a lot of criminal trials.

This jury will have been selected knowing this will be a very long and very difficult case. Effectively it became their day job for nine months. It would appear they took it very seriously. If you have attended court or sat on a jury you know the evidence gets repeated endlessly. There is very little risk the jury did not appreciate the key points the prosecution and defence wanted them to understand.

Juries can be problematic where underlying cognitive biases about who offends or is a victim and why distort their decision-making. This happens in rape cases when pre-existing assumptions about how victims behave may distort judgment, which is why prosecutions can admit special evidence to rebut rape myths. It can also affect defendants from minority groups. But Letby was a white, educated woman. It is unlikely the jury had a strong cognitive bias in favour of finding her guilty.

Juries can also be problematic when the underlying legal points or evidence are simply too abstract for them to make sense of. In these circumstances, e.g. in following the evidence in complex fraud cases, we do appoint specialist juries.

But this evidence is not too abstract to make sense of. It is straightforward deductive logic. If a is true and b is true it must follow that c is true...

The job of the defence is to disprove or at least demonstrate a and b are not proven. On some of the charges, where Letby was acquitted, they clearly did this.

Despite arguments to the contrary no one on this thread has pointed to anything in the prosecution case that relied on inductive or abductive inference which could be flawed. Because this is precisely the kind of case the prosecution avoided making.

It either confused the issues or implied that LL did it rather than the opposite.

The prosecution had very deliberately not adduced statistical evidence. Had the defence chosen to invite Gill to give evidence that it was not significant that Letby was on shift on all those occasions the obvious thing to do for the prosecution to do in response is say 'Yes, we're not claiming we're inferring from her presence that she is suspicion. We're claiming that because deliberate killings occurred on days 1,3 and 5 and L was the only person on shift on all of days 1, 3 and 5 we have deduced she is the only common factor who could have caused the killings.' I.e. the defence looked at the offer and thought this will only draw attention to a weakness in our case.

The Gill claim - based on his work on the de Berk case - is that the SAME mistake happened. I.e. they only investigated after noticing a high death rate when de Berk was on duty and then only examined the deaths that occurred when de Berk was working in the hopspital and then having retrospectively established that the deaths when de Berk was on duty were the suspicious ones, calculated the odds against her being present at 342 million to one against.

That did not happen here. The police investigated all the deaths in the unit, isolated those which could not be explained, and then isolated those which looked as if they had been caused deliberately. The final stage was looking for who had the opportunity to commit the killings or attempted killings. Its the opposite approach, they have not inferred de Berk is suspicious from her association with high death rates. They deduced Letby is suspicious from the fact deliberate killings occurred and she is the only person who had the opportunity to commit the murders.

Importantly assessing whether deductive logic is plausible is straightforward, young children can do it.

You can parse the prosecution case as a simple logic problem a 5yr old can do:

a) Last week I went into the kitchen every morning at 10am. On Mon, Tues and Friday mornings I saw that someone had taken cookies from the jar.
b) The only people who went into the kitchen before me were Annie and Sam.
c) Sam went into the kitchen before me on Mon, Tues and Wed.
d) Annie went into the kitchen before me on Mon, Tues, Thurs and Friday mornings.

Provided you can solve that successfully you can assess the validity of the evidence of L's presence of the ward. Gill are in error in claiming the prosecution are saying the equivalent of Annie must have taken the cookies because she was there. They are saying the only person who had the opportunity to take the cookies was Annie.

It is OK to dispute whether or not the cookies were stolen/the killings were deliberate. The defence had every opportunity to do that and definitely commissioned their own expert witnesses. They decided not to call them, my deduction from that is that they could not rebut the overwhelming and tragic evidence of deliberate killing.

There are some interesting articles but from what I can gather none of these deaths were considered suspicious until someone produced the clustering theory.

It then appears that someone retrospectively assessed these for foul play by LL.

There must have been some evidence though that allowed a distinction between murder and natural causes, and she wasn't charged over every death, noting all nurses in this environment are going to oversee natural deaths

That mischaracterises the clustering theory, as John Quigley has. In actual clinical practice doctors were not initially suspicious of Letby at all, in fact the opposite. But they were highly concerned when four babies suffered very unexpected collapses and three then died within a two week period in June 2015. When they investigated they found she was the only nurse on shift, but they did not infer foul play, they thought perhaps medication errors were to blame. Its not until a year later and after a further 11 babies who she has now been found to have murdered or attempted to murder that she was removed from the ward.

In healthcare we have to investigate when untoward outcomes affect our patients. If four of my stable and recovering patients collapsed in a two week period unexpectedly when I usually expected that to be a never event of course I would investigate. But foul play would be the last thing on my mind. As it was for the doctors on this ward.

There was evidence for all the charges made - evidence of air embolus, suffocation, overfeeding with milk, and poisoning with synthetic insulin. 61 cases were reviewed externally, she was only charged with 21 cases of murder and attempted murder and found guilty of 14 (now 15 after the retrial re Baby K).

So most of the deaths and collapses on the ward were treated as natural and explained. She was then acquitted in 6 cases and only retried on one outstanding charge where the jury could not reach a verdict.

Thank you again for being so clear. Now that I can see your position in black and white, I have to say that I disagree with several of your assumptions. Whilst, as noted, I know little about law, I do have expertise on cognitive biases, and I'm deeply skeptical that the jury would have overcome theirs simply by repetition, trying hard and taking the job seriously.

The claim of "very little risk" is an empirical one. It goes against most of the research on the topic, and also my own experience of learning and teaching statistics. I've cited research up thread which shows that medics themselves are terrible at statistical inference, despite it being something they're trained in and part of their daily job. I would need to see some compelling empirical evidence to convince me that your assumptipn of low risk was correct, given that it contradicts my priors so strongly.

I contend that if I had not been trained in statistical reasoning, I would have regarded that shift data as damning, even if it had been presented merely as "evidence of presence" and, probably, even if I'd been told not to. This is because it typically takes a long time and a clear exposition in order to overcome such biases, even if you have a natural affinity for mathematics (which most people don't). As an example, most people argue against the Monty Hall problem for a long time after it's been explained to them.

Once I'd adopted this view, it would likely have influenced my concern with, and interpretation of, the strength of the medical evidence. Again, there's lots of evidence that this is how humans think (Bayesian conditioning on priors). That report by the RSS covered this in the section on suspicious deaths. I would therefore have adopted a lower threshold for "beyond doubt" for the evidence of murder than I should or otherwise would have..

I reiterate my point about there being too much emphasis on what the prosecution did, and not enough on how the jury responded. It's not sufficient to say that they were "directed" to respond a certain way by the prosecution. That's simply not how human cognition works (see, for example, "don't think of a pink elephant "). And it's not sufficient to say that the prosecution relied exclusively on a simple deductive argument - because it's unlikely, a priori, that the jury did similarly.

The jury are entitled to deliberate in secret so we cannot know how they used the evidence. Your objection applies to any trial by jury. It is not a perfect process and I do not think anyone would defend it as that. The point is that it allows the evidence to be tried by Letby's peers - i.e. people we trust are motivated to treat her fairly because they could equally be in her position, and not by someone exercising power due to their senior position like a judge.

The other objection is simply allowing the jury to see the shift data would cause an irreversible cognitive bias that makes the subsequent presentation too biased to be fair. Or that if the shift data were presented with more deaths included (all 61 reviewed externally) and thus showing she was not on shift for other deaths which were not determined to be suspicious that would alter the juries perception of her.

I think that is far-fetched. If the shift data were excluded or not presented in a tabulated form but the charges were merely presented as a series of independent cases at which Letby happened to be present the first Q I would ask as a juror is have we established if anyone else had the opportunity to commit all these deliberate killings i.e. why haven't you done this simple exercise in deductive reasoning? The prosecution have to exclude other suspects for their case to meet the threshold for charging let alone trial and any juror of average intelligence would spot that.

The other question - do they need to exclude non-suspicious deaths and collapses? I assume a table of that nature could be generated and she would likely be present for some but not all the non-suspicious events. This would be meaningful ONLY if the jury were being asked to draw inductive inference from the evidence and would need a statistician to help them understand how to do and probably a calculation of how likely it is x would be present for so many events. So it isn't generated because the prosecution made no claims of that nature.

Of course you could argue I have hindsight bias because I am assuming the events on the table ARE suspicious. I think, having read the appeal court judgment, they are highly suspicious and difficult to explain except as deliberate killings. Letby herself admitted this.

The only way to make the table less damning is to also assume the events tabulated are not themselves deliberate killings. That's a huge assumption given the volume of evidence.

My points do not apply equally to any trial by jury. If the trial concerned a single incident they would not be pertinent. My argument is that the type of evidence presented in this trial is especially likely to trigger a pro-guilty bias when handled by non experts.

It is true that it would have been difficult for the prosecution to make their argument without presenting the shift data in some form or another. But this does not support the safety of the conviction. It supports the argument that such data should have been presented only to a panel of expert jurors, who are specifically trained to avoid such biases.

Have you attended many trials? The prosecution job in this country is to trigger a pro guilty bias. It's not an inquisitorial system. So whether it's a single count of theft or multiple murders the opening case for the prosecution will try to present the evidence in a way which is as damning as possible for the defendants

Miscarriages of justice are, I would hazard, happening daily in our crown courts involving ordinary young men who look the type and who the judge and defence know have a history of priors as long as your arm.

But your point still doesn't address my point. How is the conviction unsafe? What about the shift table and the point I have raised about how it should be read makes you think Letby did not commit the offences taken in conjunction with the overwhelming and unrefuted evidence the children were deliberately killed?

Anyway, I thought I would add a link to Tuesday's article in the Guardian. The article in the New York Times (not the New Yorker as stated in the article) is much longer and goes into a huge amount of detail. I think if people read the articles from a stance of enquiry rather than "she's been found guilty by a jury so she is", they will appreciate that there are, at least, a lot of questions that need to be answered. Let's not forget that there's a woman looking at spending the rest of her life in prison. Guardian article. Hope the link works, I've not tried before....

Many thanks for your patient explanations.

I strongly doubt that any of the ‘deniers’ have been anywhere near a criminal courtroom or know anything about criminal procedure. You - and Tinylittleunicorn - are saintly in your willingness to spend time correcting the nonsense spouted on this thread.

This subject is, in my view, a fantastic example of bonkers conspiracy theory and, as is always bound up with it, self-importance through ‘secret’ or ‘inside’ knowledge or ‘special’ insight.

Thank you. I think for my sanity I need to drop out here..I agree there is a wilful aversion to looking dispassionately at the overwhelming evidence that a large number of babies were deliberately killed.

I'm going to bow out also, but for one last time, I am going to ask you to please read the articles. You're pronouncing on things with great certainty, but clearly many experts feel there is at least, a degree of uncertainty.

And finally, the Court of Appeals rules on whether for correct legal procedures have been followed. As an example, they ruled against an appeal around the defence not calling expert witnesses who could have supported their case. This was not a comment on the quality of the evidence of the potential witness, but because the defence could have called him but didn't.

@Bigcoatlady
Forgive me if I'm misunderstanding your point, but I'm confused about how the inadequacy of the legal system in general is supposed to lend weight to the claim that this was a safe conviction? Surely it undermines it, if anything.

As for your last paragraph, it's not the case that I'm confident she's innocent - as I've said upthread, I don't know whether she's guilty. The point of my last few posts has been to discuss whether trial by lay jury was appropriate in this particular case. I don't think it was.

Regarding the "overwhelming and unrefuted" evidence of deliberate harm. I assume you mean that it wasn't refuted in court, as it has obviously been challenged by medics in the public arena since. I would have to be an expert on those very specific areas of medicine to have an opinion on that, and to be honest, I'm a bit suspicious of anyone outside of this domain who does. I predict that more experts will come forward with opinions in due course and I'll be watching that space with interest

I have read the articles. And rebutted the points consistently about the issues regarding statistics which I understand well on this thread.

I understand entirely the role of the CA but as you will know from the judgment the appeal on ground 6 regarding fresh evidence submitted to the court after the trial regarding the significance of skin discoloration in air embolism was exhaustively considered abd discounted in.paras 160-192. It was not the case that the CA refused to order a retrial was not for the technical reasons that the defence should have adduced it at trial but also the substantive reason that the court did not assess the evidence that the specific pattern of discoloration Dr Lee has observed in one infant as specific to air embolism was likely to disprove air embolism in these other cases. Para 182 is particularly instructive.

I'm not at all convinced by the posts of @Bigcoatlady as the arguments put forward rely on trust of them knowing everything but with little back up of actual evidence and links.

This is an aggressively stupid comment.

Up until now I've been painstakingly polite and, yes, patient. Nothing I have proposed could remotely be construed as a conspiracy. I've been consistently open about my lack of legal knowledge, but I do have expertise in other relevant areas and have used this to highlight some completely reasonable concerns regarding how cognitive biases might influence a jury. These biases are widely documented and in the public domain - just because you personally don't know about something doesn't make it "secret knowledge" (after all almost everything would qualify then....)

I don't feel like @bigcoatlady has really been able to answer all my concerns, but at least she's been broadly civilised. Why don't you give it a go.

Nah, mine was a spot on and well measured comment.

Do you know what a troll is? Clue: it’s not somebody who posts about how a criminal conviction of a baby killer is legally safe.

Legally safe by what definition given the medical complexities at play? Simplistically the prosecution constructed a case that was successful in getting the defendant convicted, yes. However, the victims were already fragile and at greater risk of collapse than full term infants. Of course this makes it much easier to paint Lucy Letby as the most vile of murderers. But does it reflect good and robust scientific evidence?

I've just spent a couple of hours reading a blog by a retired doctor who has gone through the medical evidence and good lord it's enlightening. There are differential diagnoses for each collapse and death far more complex than a lay person could imagine.

I've tried really hard to maintain impartiality in this case but it's becoming harder the more I read to be honest.

At the very least this case should be re-examined in far greater detail from every aspect.

Would you mind linking to that blog? Sounds interesting

@Reallybadidea

I'm a bit crap at links but it's someone writing as James Egan.

Be warned though, other Internet sources claim he and Richard Gill are "cranks" spouting pseudoscience etc.

So I'm somewhatback to throwing my hands up in the air regarding science, medical complexity, expert witnesses, motive, psychology, our justice system and human fallibility in general. And that's mainly because I recognise I lack the expertise to analyse it all. Yet still I am uneasy and I can't imagine how the jurors felt.

The only certainty I am left with is that for so many professionals to express concerns over the safety of the conviction, without further analysis untold harm will be done regarding the safety of our justice and medical institutions, all other emotive ramifications aside.