Lucy Letby - have you changed your mind thread 4

[MistressoftheDarkSide]

With thanks to the original poster @kittybythelighthouse and @Tidalwave for continuing the discussion.

From private eye:
DR EVANS told MD: “There was a third insulin poisoning in November 2015. The insulin value was recorded as “>1000” and the C-peptide as “220” as handwritten entries in the notes. This is VERY abnormal. One expects the C-peptide to be 5-10 times the insulin value normally. Letby was most certainly on duty. She was the nurse who measured the six low glucose values, lowest of 1.0, when the baby was hypoglycaemic for nearly eight hours. The baby certainly survived and as far as I know is well. He was sent to Alder Hey and received a diagnosis of hyperinsulinism. But I think that is incorrect. If the baby had endogenous hyperinsulinism – ie producing his own insulin – his C-peptide would be high as well.”
https://www.private-eye.co.uk/pictures/special_reports/lucy-letby-4.pdf

And the Criminal Bar Association chair is right to raise this point. There are structural problems with achieving equality of arms in this country, which have become particularly obvious in the post LASPO climate. Three cheers for austerity!

I find the insulin cases troubling for a number of reasons.

1. There is no evidence of LL acquiring a high dose of insulin with which to spike the bags.
2. She had no way of spiking the bags without it being evident they had been tampered with. So if LL was a killer nurse then other staff on the unit who administered the tampered bags must have been in cahoots with her or were incredibly negligent in their administration.
3. The assay tests are not treated as a forensic test - if they throw up an anomalous test you are supposed to ask for a further testing of the sample - which did not happen.
4. A quality control test performed in the same lab shortly afterwards showed the same anomalies.
5. The tests are liable to distortion by antibodies which are often present in neonates due to placental transference. Sepsis can also create these distortions - and this was a filthy unit according the the Royal College report. Another reason why these samples should have gone for further testing.

Putting all of this together I would be uncomfortable convicting LL beyond reasonable doubt.

Anyone who actually believes that access to justice is fair and equal for all is walking around blindfolded (or havent been on the sharp end of it)

I don’t think he did row back at all actually. Norm has a definition, in statistics. Rare, unusual and uncommon don't.

For a standard set of data, over 90% of your results fall within the norm. So reasonably but not extremely rare, unusual and uncommon things can certainly fall within the norm. Of course that's not how we usually use language outside of science and statistics. But he is speaking in scientific terms. Chase's categorisation of the insulin results here could be clearer, but given how heavily edited the expert sections were in this episode of Panorama, it’s very likely that more was said but not included.

Sure. I'm pointing you to the article that has most information on the document. The authors are focusing on the legal implications of the existence of a similar insulin test result, with no charge, and no disclosure to the defence. That's the most important thing about this evidence.

In terms of timings, the seizures were certainly liable to arise from the hyperinsulinism, but that doesn't really matter. What matters is that the notes indicate that the test was carried out during the night shift, before Lucy Letby came on duty. And although the hypoglycemia continued all that day and much of the night, the child wasn't on TPN. So the mechanism the prosecution identified for babies F and L wasn't available either.

Now one could argue that Lucy Letby must have sneaked in early, must have had another way of poisoning children, must have luckily picked a child who carried on being hypoglycemic anyway - but I think anyone who reached that stage would have to acknowledge that they'd be able to frame anyone on the ward for anything once they reached these extremes.

As Hall said to the New Yorker, if this file had been submitted to the defence, that could have altered their arguments on child F and L. We saw from Private Eye this week that the insulin problem seems to have constrained the whole defence strategy. The existence of this case is highly significant.

Absolutely. I was about to make exactly the same point.

That's presumably why Chase described the words as not best chosen. They weren't, for a general audience. But there's no fundamental contradiction. And of course, even if there results are rare, but happen, that overturns the information given to the court: that they are impossible except through exogenous insulin.

Yes. Evans is (or was) determined to bring that baby's case into the light, because he thinks it's another mark against Letby. But it's the opposite.

So the information from Panorama is definitely incorrect then, as they say Baby Y’s blood sugar was normal at 06:56?

Medical records seen by Panorama show how quickly the boy became poorly after Letby came on duty. A blood test taken at 06:56 showed the infant had a normal blood sugar level of three millimoles per litre (mmol/L).
^^
Letby started her shift at 08:00, and by 13:54 his blood sugar level had plummeted to one mmol/L – a dangerously low level, and a strong indication the baby had too much insulin.

Are the new experts saying that the insulin/c peptide ratios in babies F, L and Y would definitely fall within the norm then? Because there’s obviously some disagreement about that from other paediatric endocrinologists who weren’t involved in the case.

The number is probably correct - I don't see why not.

3 is borderline, so I've been told a medic seeing a child with associated and urgent symptoms like seizures would consider hypoglycemia protocols based on that result. In any case, she did take the sample for testing, whether Moritz and Coffey think she needed to or not.

There was an excellent report on statistics posted a few days back that is really worth a read.

The first step is that you work out the probability that something occured randomly and we don't have enough data even for that.

Honestly (I read a book called Insulin Murder) I think Insulin test results have been badly misused by courts in general. They can be wrong for other reasons aside from lab errors and prematurity such as sepis and there have almost certainly been other miscarriages of justice like Deborah Winzar. Whose case demonstrates there are no guarantees a miscarriage of justice will ever be righted.

As a minimum I think that labs should be required (some kind of legal duty) to send these sort of results for further testing.

It could all be as simple as a transcription error with a careless person at the lab prone to putting the results for insulin and cpetide the wrong way round, which they did in a sample set for quality control.

I feel like the jury was misled being told that only the administration of insulin could lead to these results. The reaction of the doctors at the time to the results (to ignore them) suggests to me that in practice odd results come back not infrequently and it's usually just an error. Though really they should have followed up.

Also was there an expert who said that if the insulin levels were actually as high as the test showed the babies would certainly have died. If so can anyone remember who it was?

Utterly bizarre! Evans decided the baby didn’t have hyperinsulinism despite being diagnosed with it at the Women’s and requiring treatment for it for weeks??

No, and I'm glad you you asked because that's a common misunderstanding.

What they said is that the C-peptide values on their own are typical for the cohort, but that the insulin-C-peptide ratios are meaningless where the insulin values are artificially high. They posit that artificially high insulin values (from insulin antibodies) are "relatively common' in this cohort. And all of that needs to be contextualised with the introductory statement:

"The following summary analysing the evidence against Lucy Letby for Babies 6 and 12 (“the insulin cases”) focuses on a detailed analysis of the evidence presented, as well as the full medical records made available. Thus, this release presents only the briefest high-level overview of a full ~100 page report with 250+ supporting references from peer-reviewed science publications to support the points below, many of which were published since the trial."

In other words, they're at the cutting edge of scholarship and one wouldn't expect all endocrinologists (or experts in testing) to be on the same page as them - and most especially not in statements from 2023 and 2024 or before, when they wouldn't have had a chance to assess new research in the field.

Shoo Lee's first report (Annex 1) as well as the insulin-specific summary from his panel is useful here. Quotation is from annex 1.

lucyletbyinnocence.com/#shoolee

That tech has no similarity to the one we’re discussing, but it does show that kit breaks down and gives duff results!

That was the only reason I posted it 🙃

To be fair, he's following the prosecution's logic. Tests with very high insulin - C-peptide ratios can only mean exogenous insulin - they say.

So either that's not true, and there's no reason to rule out transient hypoglycemia for the other two. Or, it's true, and congenital hyperinsulinism can't explain baby Y's test.

He saw that child's file before the first trial. He wants a charge and conviction. But the police for some reason don't. Could it be that they know they won't get a conviction, because Lucy Letby wasn't on shift?

Really? How likely is it that the paediatric endocrinologists and other relevant experts who have recently been commenting on the insulin/c peptide ratios (and who will no doubt have read the expert panel’s summary reports) are completely unaware of all these peer-reviewed publications and supporting references that this cutting edge scholarship is based on?

If all this research exists, they will have seen it. However, they are obviously still confident enough about the prosecution position on the insulin/c peptide ratio to comment publicly on it, which suggests to me that they have not been persuaded by this new research.

Just to add, I didn’t mean to suggest that the new research doesn’t exist - I realise my wording made it seem like that’s what I was saying!

And presumably there would have been pushback from whoever diagnosed at LWH and continued to treat the child for I think seven months. Thereby exposing the assertion that the ONLY possible reason for the results was deliberate poisoning as false.

Depends how recent. "Since the trial" is since August 2023. I know it wasn't you who was quoting Frayn upthread, but he and Wolfsdorf are both good examples of people who commented within a year or so of the trial, and haven't since. There's also a huge difference in the extent to which people keep up on research in a niche area in their fields. And this is very much a niche area, because if you get an unexpected result like this normally, you send your sample off for full analysis. You don't need to mess around a decade later trying to prove whether or not the insulin is exogenous

I've no problem with people citing experts disagreeing with Shannon, Chase and colleagues, but there is no point in assuming disagreement from people who haven't expressed an opinion on their propositions - i.e. since February 2025. They're consciously synthesising and presenting the new findings of recent researches - they're not claiming to be presenting established consensus views. So the importance of people disagreeing with them, can be overstated. And I'm not particularly worried if people, very reasonably, wait to see more before assessing their work. That is normal research practice.

Prof John Gregory commented on the most recent Panorama, so within the last few months?

Re: high insulin: cpeptide ratios.

All the experts are right when they say that the only thing that can mean is exogenous insulin - it is impossible for the body to produce insulin without producing cpeptide in equal amounts - in fact cpeptide is frequently used to measure how much insulin someone is making - because of the very nature of the precursor molecule needed for release of insulin into the bloodstream.

So they are right that something else has gone on in those babies. Or with the tests.

I guess it’s going to remain to be seen what Geoff Chase’s research says re antibodies and prem neonates

If baby Y’s ratios were also high (cant remember) then that really does point to something else going on (though he wasn’t Orem - he was overdue wasn't mum diabetic?)

I’ve not had a chance to read the panel’s report yet.

Does Chase go into more detail beyond what was said in Panorama?

@Kittybythelighthouse- I read somewhere that Gibbs got his dates wrong and baby F received insulin to treat the hyperglycaemia as late as the 4th of August (instead of just July 30th) I can’t find anything to back up this claim.

Do you know anything about it?

Oh yes, I agree. I just meant that the test results were more likely to be erroneous than indicate transient hyperinsulinism.