Help end medical misogyny. Sign our petition.

Help end medical misogyny.
Sign our petition.

Sign the petition

Please or to access all these features

AIBU?

Share your dilemmas and get honest opinions from other Mumsnetters.

To think that Lucy Letby will get a new trial?

552 replies

NameChangeMay2026 · 28/05/2026 17:40

The previous thread on Letby is almost full. Posting here for traffic.

If we have any lawyers here, what do you think the likelihood is of Letby getting a new trial? I'm a layperson, but I'm going to guess that she will get one. It seems that many, many rebuttals have appeared since her conviction.

YABU - she will not get a new trial. The case is settled.
YANBU - the new evidence/discussion is compelling and she will probably get a re-trial.

I've been mainly convinced of her guilt, but I have started reading the free Private Eye series on the case by Phil Hammond. Now I don't know what to think. Here's the series, if anyone wants to read it. https://www.private-eye.co.uk/special-reports/lucy-letby

Special Report: The Lessons of the Lucy Letby Case

After Lucy Letby was convicted in August 2023 of murdering seven babies, a number of experts contacted Eye columnist MD because they

https://www.private-eye.co.uk/special-reports/lucy-letby

OP posts:
Thread gallery
5
kkloo · Today 00:30

I just asked chatgpt the same question, then followed up by asking How strong and capable is this current model of chatgpt of assessing how rare that the insulin test results were and of accounting for things that may have affected the results?

Good question — and it gets right to the limits of what a model like this can and can’t do.
Short answer
This model is good at explaining medical logic and probability in general terms, but it is not capable of independently determining how rare a specific test result was in a rigorous scientific sense, nor can it fully account for all lab-specific variables the way a forensic endocrinologist or toxicologist could.

What it can do well

  1. Explain the biology accurately It can reliably describe things like:
  • how insulin and C-peptide normally behave
  • what patterns are expected in endogenous vs exogenous insulin
  • why certain combinations are considered suspicious in clinical toxicology
That part is grounded in well-established physiology. 2) Summarise how experts interpret evidence It can outline:
  • why courts or expert witnesses considered a result unusual
  • what arguments were made on both sides in a case like Lucy Letby’s
  • what kinds of methodological criticisms are commonly raised (e.g., assay limitations, sample handling, neonatal variability)
3) Compare plausible explanations It can reasonably evaluate:
  • “this pattern fits A vs B vs C explanations”
  • which explanations are more or less consistent with medical literature generally
  • where uncertainty exists in interpretation
What it cannot do reliably 1) It cannot calculate true “rarity” from real-world datasets To actually say “how rare this is,” you would need:
  • large neonatal ICU datasets
  • validated distributions of insulin/C-peptide under different conditions
  • lab-specific assay performance data
  • case-control comparisons
I don’t have access to that raw clinical database, and even in real life it often doesn’t exist in a clean, publishable form for exactly this scenario. So any “rarity” estimate from me would be inference, not measurement. 2) It cannot verify lab artefacts in a specific case In the Letby insulin issue, important technical factors include:
  • which immunoassay was used
  • cross-reactivity and antibody interference
  • sample timing and storage
  • neonatal blood volume constraints
  • lab reference ranges for preterm infants
A model like this cannot inspect those original lab conditions or validate whether they affected the result. 3) It cannot resolve contested expert disagreement In cases like this, experts may disagree because:
  • they weight lab uncertainty differently
  • they interpret “extreme rarity” vs “possible artefact” differently
  • they may have access to different subsets of data
I can summarise both positions, but I cannot adjudicate which expert is correct. So how strong is it, realistically? A fair way to think about it:
  • Strong at conceptual reasoning (high reliability)physiology
  • mechanism-based interpretation
  • explaining why evidence is considered significant
  • Moderate at evidential synthesissummarising competing arguments
  • identifying potential confounders
  • Weak at forensic quantificationexact rarity estimates
  • lab error probability in a specific real case
  • definitive conclusions about causation
Bottom line It’s best seen as: a very capable explainer and structured reasoning tool, not a forensic calculator or independent expert witness.
Oftenaddled · Today 00:37

Firefly1987 · Today 00:18

I asked chatGPT about how rare these insulin results were-

3. Reported literature

  • Case reports of insulin:C-peptide ratios like the Letby babies’ are extremely rare.
  • Most published cases of high insulin / low C-peptide in infants are either due to deliberate insulin injection or serious assay interference.

Summary

  • In most medical literature, the ratio seen in the Letby trial is considered strongly indicative of exogenous insulin.
  • Rare exceptions exist, mostly in complex neonatal metabolic disorders or technical lab issues, but these are far less common than insulin injection explanations.
  • That’s why the original prosecution and most endocrinologists considered it compelling evidence — while critics argue that in neonates and with assay limitations, the result might not be absolutely definitive.

Bottom line

  • The insulin/C-peptide pattern seen in the trial cases is very rare naturally.
  • There are known exceptions, but they typically don’t match the full pattern as neatly.
  • That’s why it has been treated as one of the strongest medical signals in the case — while still being debated in terms of evidential certainty.

ChatGPT also said it's unknown where Letby was when the third insulin test was conducted as the timeline hasn't been established in court. So any claim that she was only on shift AFTER the test was done (convenient) is pure conjecture.

Of course the timeline for the third insulin case wasn't established in court. The case wasn't put before the court. Documents referring to it are among the non disclosures Lucy Letby's defence has listed as grounds for referral in the CCRC documentation.

These documents show that the tests were ordered before Lucy Letby came on shift.

The question of whether I had breakfast this morning hadn't been established in court either. But I did.

Oftenaddled · Today 00:42

Note that the information on children with results comparable to the LL cases doesn't come from case reports, but from a dataset.

Note too that the court wasn't told the results were strongly indicative of poisoning. It was told that they left no other possibility.

ChatGPT isn't doing much to help your cause there. It's well known that it will do its best to give you the answer you ask it for, so it is not much use as a tool to convince.

Firefly1987 · Today 00:44

Lol how does chatGPT know I'm not a truther?

amateursleuth · Today 00:44

Oftenaddled · Today 00:37

Of course the timeline for the third insulin case wasn't established in court. The case wasn't put before the court. Documents referring to it are among the non disclosures Lucy Letby's defence has listed as grounds for referral in the CCRC documentation.

These documents show that the tests were ordered before Lucy Letby came on shift.

The question of whether I had breakfast this morning hadn't been established in court either. But I did.

But if the door entry data wasn't reliable and lots of staff were using the back stairs, do we know Letby wasn't in the building earlier than the start of her shift or at another time?

This is what's bugging me about the 'back stairs' revelations and entry cards. If what's now being said is that we can't actually know for sure who was there at what time, AND it's been said that staff were coming and going all the time even when not on duty (and who stated this? It's in part 7 of the Private Eye reports but not attributed) then you can't rule out anyone being there. Including Letby.

Oftenaddled · Today 00:48

Barbie222 · Yesterday 20:50

The door data would have actually harmed her case more if it had been presented correctly in the first trial IIRC.

Insulin - even LL admitted at trial that someone had administered insulin, it ‘just wasn’t her’. Any new proof that c-peptide could have been produced in an infant, without exogenous insulin, on two separate occasions, would obviously be of great interest to the CCRC so let’s see what they’ve got!

Statistics - The deaths in the list were the unexplained deaths, where infants were not expected to crash. That’s why not all deaths were included. You could try and argue that these deaths were actually to be expected, which is what Lee did after the fact, but that didn’t work at retrial because the jury were invited to consider this at trial in the first place and didn’t find it convincing.

Air present in x rays post mortem could have been from decomposition, but again the jury were presented with this and didn’t find it convincing - it was clearly more likely to them that it had been administered with purpose.

Whether or not Letby called for a doctor on one occasion was really here nor there when viewed against the great stack of evidence against her. It did not seem to take away from Jayaran’s credibility as a witness at second trial.

So the issue is - these key arguments were presented at trial, and the jury didn’t find it believable. Unlike social media, in the jury system you can’t just shout what you think again and again, and hope people believe it a little bit more. She had her chance, maybe she’ll get another chance, but my view is it’s doubtful given the weight of argument she has to overturn, so I’m left with the conclusion that the correct decision was made.

There's a significant period for which the unit had lost its swipe card data. The prosecution implied that this would have implicated Lucy Letby in a couple of cases, if it had existed. Obviously that's pure speculation.

That's different from the issue of the reversed swipe card data. I've never seen a claim that makes this unfavourable to Lucy Letby. It did, however, undermine at least two of the cases on which she was convicted. These were the cases of baby K and baby N.

Firefly1987 · Today 00:56

Oftenaddled · Today 00:37

Of course the timeline for the third insulin case wasn't established in court. The case wasn't put before the court. Documents referring to it are among the non disclosures Lucy Letby's defence has listed as grounds for referral in the CCRC documentation.

These documents show that the tests were ordered before Lucy Letby came on shift.

The question of whether I had breakfast this morning hadn't been established in court either. But I did.

Well yeah but I thought they might have a timeline for it but that was the answer-it wasn't established in court so no. I guess even chatGPT doesn't have the info you have. I'd put money on her being there.

As a PP said, she was known to come in early or stay late and that's exactly the type of thing she'd want to do that day to distance herself from the poisoning.

Oftenaddled · Today 01:00

amateursleuth · Today 00:44

But if the door entry data wasn't reliable and lots of staff were using the back stairs, do we know Letby wasn't in the building earlier than the start of her shift or at another time?

This is what's bugging me about the 'back stairs' revelations and entry cards. If what's now being said is that we can't actually know for sure who was there at what time, AND it's been said that staff were coming and going all the time even when not on duty (and who stated this? It's in part 7 of the Private Eye reports but not attributed) then you can't rule out anyone being there. Including Letby.

That is very different from saying that she was on shift when the sample was taken, and if we do reach the stage of assuming anyone was coming in by the back stairs covertly while off shift, the evidence against Lucy Letby is obviously undermined since there is no reason for that person to be Lucy Letby.

There was also (if we want to go further into the case of Baby Y) no way for Lucy Letby or anyone else to produce prolonged hypoglycemia in that child, since the child was not on TPN - unlike baby F - not dextrose - unlike baby L - at that point. Their hypoglycemia persisted through new prescribed bags, again while Lucy Letby was off shift. Injected insulin would have affected the child only briefly and couldn't account for the later longer bouts of hypoglycemia. So we are having to postulate a scenario where Lucy Letby (or someone else) sneaks into the unit in the early hours of the morning, enters the nursery where a child is receiving one-to-one care, and injects him with a syringe, unnoticed. Then later he develops hypoglycemia independently for a longer period ...

That really is not plausible

Oftenaddled · Today 01:05

Firefly1987 · Today 00:56

Well yeah but I thought they might have a timeline for it but that was the answer-it wasn't established in court so no. I guess even chatGPT doesn't have the info you have. I'd put money on her being there.

As a PP said, she was known to come in early or stay late and that's exactly the type of thing she'd want to do that day to distance herself from the poisoning.

That's getting a bit desperate now. I've explained why it wouldn't work in my previous post.

That failure with ChatGPT is a perfect example of why we can't say, we can only believe what has been tested in court. If the jury wasn't given relevant information, any conclusions involving that information haven't been tested in court.

Oftenaddled · Today 01:15

Here, on the other hand, is an instance where the swipe data exposes serious problems with the prosecution account:

Original Trial (from Chester Standard)

The court is shown swipe data for Joanne Williams, who left the neonatal unit at 3.47am.

Mr Myers says it is very precise in coinciding with Dr Jayaram's recollection of waiting two-three minutes before the desaturation is timed at 3.50am, and asks if Dr Jayaram always has such a precise memory.

Dr Jayaram says "In this event, I did."
He adds: "I kept telling myself, don't be ridiculous [about my suspicions]. I looked at my watch - I didn't have a stopwatch."

Dr Jayaram says he has never seen the swipe data, nor had cause to look at any data.

The problem? This was the reversed data. That means the child's nurse didn't leave the unit at 3.47 (leaving the child with Lucy Letby). It means she returned at that time. In the retrial, accordingly, Dr Jayaram made no mention of the time of this event ...

There's an excellent account of his ever shifting story on this incident at
https://www.reddit.com/r/LucyLetbyTrials/comments/1irru42/who_what_when_why_how_the_evolution_of_dr_ravi/

Brokentoes85 · Today 02:11

Step father, brother, brother in law all barristers, grandfather judge. All say 100% she'll get a retrial.

NameChangeMay2026 · Today 03:24

Brokentoes85 · Today 02:11

Step father, brother, brother in law all barristers, grandfather judge. All say 100% she'll get a retrial.

So interesting!

Do they have any thoughts about whether ensuring a fair trial would be an issue? Is having her conviction quashed and being given a new identity likely to be an alternative outcome? Can you plse ask them for us??

OP posts:
kkloo · Today 04:48

NameChangeMay2026 · Today 03:24

So interesting!

Do they have any thoughts about whether ensuring a fair trial would be an issue? Is having her conviction quashed and being given a new identity likely to be an alternative outcome? Can you plse ask them for us??

Edited

This is what I'm wondering, I would imagine it would be very difficult for the CPS to bring the case again, Dewi Evans is retired, Sandie Bohin said she wouldn't take the stand again. There's also problems with Jayarams evidence etc.

Also it's the special division of the CPS who would be bringing the case, not the Chesire CPS and it's the special division that decided they weren't proceeding with the second set of cases that Chesire police put forward because there wasn't enough evidence.

SnakesAndArrows · Today 06:23

Edited for quote fail. See next post.

SnakesAndArrows · Today 06:24

Firefly1987 · Yesterday 22:56

I mean, you're taking his opinion over actual endocrinologists, I just wouldn't personally that's all.

I don’t think you have very good scientific understanding at all.
Insulin testing and modelling is not within the skillset of endocrinologists, generally. They are physicians - often brilliant ones. They understand and treat disease, and make massive differences to people’s lives. They do not design laboratory equipment; they do not design drug assays, conduct them or research them.

Engineers on the other hand, biomedical mechanical engineers, do.

It’s quite bizarre that you persist with your ignorance despite having had this explained to you many times.

Moreover, you seem to think that all endocrinologists agree with Hindmarsh. This is demonstrably untrue.

SnakesAndArrows · Today 06:39

Firefly1987 · Today 00:44

Lol how does chatGPT know I'm not a truther?

Wild that you’d actually post something straight out of AI as though it was the truth.

If you read what it regurgitated for you, you can see that it’s lacking in detail, equivocal and flabby, and doesn’t really support your view in any meaningful way.

Hint: “rare” is not synonymous with “impossible”. “Strongly indicate” does not mean “prove”.

To obtain good information you have to ask it much more structured questions, and then check the answers independently, in case of hallucination.

IonianNerveGrip · Today 07:19

kkloo · Today 04:48

This is what I'm wondering, I would imagine it would be very difficult for the CPS to bring the case again, Dewi Evans is retired, Sandie Bohin said she wouldn't take the stand again. There's also problems with Jayarams evidence etc.

Also it's the special division of the CPS who would be bringing the case, not the Chesire CPS and it's the special division that decided they weren't proceeding with the second set of cases that Chesire police put forward because there wasn't enough evidence.

Same. I could foresee problems getting expert witnesses for the prosecution given what this has turned into- pretty much the dictionary definition of going to cause you more trouble than it's worth. Multiple consultants from Chester have very good reason not to want to take the stand, so if it did get to court, wouldn't surprise me if we had some conveniently timed long term sick leave episodes.

And the case has sparked so much online discussion, the risk must be much higher than usual that it emerges during the trial one of the jurors was in an argument about it on social media at some point or whatever.

Piglet89 · Today 07:43

Thanks for sharing that Private Eye, OP. I am a lawyer (tho not practising in crime) and I’m very interested in this case. I don’t know the thresholds that must be met before a retrial is granted. Having watched some documentaries about her case though, I’d be very shocked if the case plus subsequent debunking of evidence in her trial plus new evidence didn’t result in an acquittal.

I think the fact her defence team seems not to have adequately challenged that chart the prosecution relied on (showing her being on shift for every one of the baby death cases in respect of which she was being tried) is a facet of lawyers being shit at maths - so not immediately looking for the holes in the stats. The challenge for that was to ask for disclosure of the other baby deaths she wasn’t on shift for and wasn’t being blamed for. But nobody seems to have bothered doing that.

Either way, the standard of proof is high in criminal cases and I don’t think there’s any way a jury can be sure to the beyond reasonable doubt standard that she committed these baby murders.

happybug1234 · Today 07:45

Secretseverywhere · 28/05/2026 17:57

I have a law degree but am not practicing. I do think she will get a new trial eventually, these things could take years/ decades to roll around. I think future statisticians and law students will be taught how the erroneous use of statistics can mislead juries in the same way SIDS parents were convicted.

If there is a chance she is innocent, Surely the system shouldn’t take years to come around.

IonianNerveGrip · Today 08:01

happybug1234 · Today 07:45

If there is a chance she is innocent, Surely the system shouldn’t take years to come around.

That would be nice, but unfortunately we have recent examples of our system failing to identify clear MOJs for a long time. The Post Office and Andrew Malkinson are the highest profile recent ones. So if there has been an MOJ in this case, and I'm not saying there was, no guarantee it will be addressed quickly.

CheeseNPickle3 · Today 10:46

Piglet89 · Today 07:43

Thanks for sharing that Private Eye, OP. I am a lawyer (tho not practising in crime) and I’m very interested in this case. I don’t know the thresholds that must be met before a retrial is granted. Having watched some documentaries about her case though, I’d be very shocked if the case plus subsequent debunking of evidence in her trial plus new evidence didn’t result in an acquittal.

I think the fact her defence team seems not to have adequately challenged that chart the prosecution relied on (showing her being on shift for every one of the baby death cases in respect of which she was being tried) is a facet of lawyers being shit at maths - so not immediately looking for the holes in the stats. The challenge for that was to ask for disclosure of the other baby deaths she wasn’t on shift for and wasn’t being blamed for. But nobody seems to have bothered doing that.

Either way, the standard of proof is high in criminal cases and I don’t think there’s any way a jury can be sure to the beyond reasonable doubt standard that she committed these baby murders.

Edited

The defence did ask for details of how the collapses were selected but they were refused permission for this.

They were also not allowed to present information on any babies not in the indictment.

TroppoCaldoPerLondra · Today 11:18

CheeseNPickle3 · Today 10:46

The defence did ask for details of how the collapses were selected but they were refused permission for this.

They were also not allowed to present information on any babies not in the indictment.

The judge also bears a lot of responsibilty for what happened in the case as quite a lot of the issues are down to decisions he made. He probably decided she was guilty and didn't want such a terrible serial killer to get off.

I think a snowball effect starts with everyone thinking she must be guilty because a previous group of people they respect did.

Consultants -> Police -> CPS -> Judge -> Jury

This is why some of my friends think she is guilty. It isn't based on any actual evidence. One (a doctor) thinks the consultants wouldn't have got it wrong.
Another thinks she is guilty because the CPS and police only bring cases in the first place if they have lots of good evidence.

Brokentoes85 · Today 16:23

NameChangeMay2026 · Today 03:24

So interesting!

Do they have any thoughts about whether ensuring a fair trial would be an issue? Is having her conviction quashed and being given a new identity likely to be an alternative outcome? Can you plse ask them for us??

Edited

They all think with the documentaries a fair retrial is v v hard to achieve. They also think quashing it is also v difficult aswel because of the families wanting "justice". All we know is the conviction is not safe.

kkloo · Today 16:37

Agree with the snowball effect, however the police and CPS should have done better, it shouldn't have mattered to them what the consultants thought. Of course they gave them a theory and pointed them in that direction, but I'm sure the police encounter that all the time but on a smaller and less complicated crime, they still have to do their job properly.

Another thinks she is guilty because the CPS and police only bring cases in the first place if they have lots of good evidence.

Yes, if I'd been a juror for a case held in 2022/2023 I would have believed that this evidence must be robust and tested, I would never in a million years have thought that they could have or would have put forward a case that was so amateur and misleading. I obviously realise now that that is very naive but it just wouldn't have occurred to me that they would have ever done otherwise. I

Secretseverywhere · Today 17:02

happybug1234 · Today 07:45

If there is a chance she is innocent, Surely the system shouldn’t take years to come around.

Sadly criminal law in the UK is littered with wrongful cases that take decades to come back around. If you’ve got any interest in the subject I’d highly recommend memoirs of a radical lawyer written by Michael Mansfield. Will leave you shocked and appalled.