Lucy Letby: how the charges were selected

The Jolly Contrarian on Crime and Punishment

0:00

-29:49

Lucy Letby: how the charges were selected

May 09, 2025

Transcript

Prelude: A daycare centre in New Zealand

Peter Ellis was a childcare worker in Christchurch, New Zealand. A creative, flamboyant and somewhat uninhibited character, Ellis loved his job, put a lot of energy into programme planning, and would entertain children with elaborate puppet shows and somewhat provocative productions, sometimes to the point of being “risqué and outrageous”. Ellis had worked at the daycare centre for several years and was popular with children and parents alike.

In 1991 a parent made a complaint. Her child had remarked that he did not like “Peter’s black penis”. An investigation ensued. Parents talked. Things escalated. The police were involved. They interviewed hundreds of children. The stories were horrific. Christchurch police eventually charged Peter Ellis with the systematic and horrendous abuse of twenty preschool children in his care.

Thanks for reading! This post is public so feel free to share it.

Before his trial, the police had disclosed to his legal team the statements of the 20 child complainants, as they were obliged to do. The children were all under 5. Their statements were broadly consistent. They told an appalling story.

The police did not disclose the statements from nearly 100 other children — on the pretext that, since no charges were laid regarding these children, their interviews were not relevant to the criminal proceedings against Ellis.

The nature of the allegations — that Ellis had abducted the children, removed them to a different location across town, ritually abused them and then returned to the daycare centre in time for pickup — made it very difficult for Ellis rebut the children’s evidence. There were no independent witnesses. He had no alibi. The consistency of the twenty complainants’ stories was striking. It was enough to convince the jury. Ellis was convicted and imprisoned.

It later turned out that while the witness statements the defence did see were consistent and theoretically plausible, many of the other children’s stories were not. Some were physically impossible. Some claimed Ellis had amputated limbs the children still possessed, or inflicted horrific injuries of which there was no trace. This is presumably why police did not press charges charges in these cases.

Furthermore, a significant number of children had flatly denied any abuse had happened. The defence were not told of these children either.

Plainly, these statements would have enormously helped Ellis’s defence: they would have provided context in which the “credible” accounts ought to have been set. They opened the argument that the police had cherry-picked accounts from interviews with six times that many children, thereby generating a wholly misleading impression of consistency and reliability.

This was a classic Texas sharpshooter: the police had ignored a greater volume of obviously made up stories, had shelved insistent denials and painted a target around the small number of outwardly plausible statements, and presented that to the defence and the jury as the whole story.

All the allegations against Peter Ellis were false. Ellis did not abuse anyone.

Even the “black penis” remark was explained: it arose when Ellis was showing the children how to identify the sex of a labrador puppy.

That this was a catastrophic miscarriage of justice was not finally acknowledged until Ellis had completed his sentence. Though he died of cancer in 2019, the New Zealand Supreme Court took the unusual step of posthumously exonerating Peter Ellis in 2022. The case was and is a cause célèbre in New Zealand. It led to significant changes in evidence disclosure rules and a thorough revision of the interview techniques used by police and social workers.

Relevance for Ms. Letby

Peter Ellis’ case shares two features with Ms. Letby’s: firstly, the danger of selective disclosure. By presenting only the collapses during Ms. Letby’s shifts and excluding the context of similar collapses when Ms. Letby was not present, the charge sheet, and particularly the staff attendance chart, creates an “artificial coherence” to one unlikely narrative — a healthcare serial murderer — whilst suppressing evidence suggesting a much more likely one — a poorly managed and under-resourced hospital suffering a cluster of collapses in desperately ill children it was not properly equipped to treat.

Just as Peter Ellis could not properly contextualise the credible complaints against him with evidence of a greater number of incredible complaints, nor could Ms. Letby demonstrate that an unusual cluster of collapses was happening whether she was there or not.

It is sometimes argued that other collapses were excluded because they were “explained” or “dissimilar”, but that is to beg a question the defence might wish to put before the court. Who is to say they were not similar? In what way? By whose criteria?

Selective disclosure risks transforming statistical noise into an apparent guilt signal by filtering out all the circumstantial evidence pointing away from the defendant. Remove the randomness, and what remains looks like a pattern.

There is another parallel: in both cases, the alleged victims were vulnerable children entrusted to the defendant’s fiduciary care.

We have an instinctive, emotional reaction to protect the weak from predators. While we can be sanguine when grown men hurt one another, we are not inclined to give those who might have harmed children the same benefit of doubt. Abuse of that sacred trust is unforgivable. If, in vouchsafing children from predations of carers the odd innocente is caught in the gears, that is a price some seem prepared to pay.

We see this attitude in appeals to refrain even from discussing Ms. Letby’s case, prioritising the consideration for victims’ families, however valid her grievance may be. It neither follows — exonerating those falsely convicted of offences that did not happen should hardly upset the bereaved — nor is it a suitable moral calculus even if there was an offence, but the defendant did not commit it. Two wrongs do not make a right. The victim’s damage is done: we should be doubly sure we are not compounding it with further injustice on another.

Implied evidence

Few people in history have been charged with more than twenty murders and attempted murders. Fewer still where no direct evidence links them to any of the charges. But Ms. Letby has.

There were no eyewitnesses. No fingerprints. No CCTV footage. No missing vials of insulin. There wasn’t even any tittle-tattle amongst the other ward nurses: inquisitors at the Thirlwall Inquiry asked them, and they all roundly denied it, in writing.1

No one saw Ms. Letby do anything. No one suspected her of anything. The strongest circumstantial evidence incriminating her was that she was there. That’s it. That is the clincher. Ms. Letby was at work.

As causal links go, this is not strong.

What made it stronger, the Crown argued, was repetition.

Ms. Letby wasn’t just there for one collapse. She, and no one else, was at work for all of them. The only common thread across the twenty-two collapses was Nurse Letby.

Do not underestimate the significance of this assertion. When it comes to evidence identifying the perpetrator, it is all there is.

That attendance chart

To illustrate this, the Crown ginned up a “staff attendance chart” for the twenty-five shifts over the suspect period on which unexplained collapses happened. The chart set out the attendance of every nurse in the neonatal unit. As it happens, the chart wasn’t awfully accurate, but let’s disregard that for now.2

The chart painted a stark picture. Whereas each other nurse’s attendance was sparse, Ms. Letby shows as a uniform column of black crosses. To labour the point, it is highlighted in sort of purplish, bruised, blue. The chart invited a clear inference:

Ms. Letby, and only Ms. Letby, was present at all twenty-two events. So, you know —

The Crown did not present much evidence that these incidents were related, much less that they were all caused by the same person.3 The jury was just left to put two and two together.

But that calculation is trickier than it looks.

It was not a “staff rota”

For one thing, the attendance chart was not, as it is often described, a “staff rota”. No one pulled this off a pinboard in the staff room and presented it in court.

This was painstakingly — if not necessarily carefully — compiled and arranged by the police for the trial. It is designed not to show, but to persuade. It invites an inference.

It should have been a staff rota. A staff rota would have included all the shifts in a given period, showing who was on and who was off at every moment in the hospital’s operation.

Now if the “suspect period” occupied fifteen months — taking a month or so on either side of the actual charges for good luck — a staff rota would show nine hundred twelve-hour shifts. It would note every collapse, whether or not Ms. Letby was on duty, and whether or not they were explained.

A staff rota would have told a very different story.

It missed some events

Nor did the attendance chart, as presented in the trial, feature all “suspicious” or “unexplained” collapses. Dr. Evans claims he reviewed at least sixty unexplained collapses, so it can’t have. We do not know who selected those sixty cases, or on what basis, and the Crown did not disclose it. It may be that even these 60 were all taken from Ms. Letby’s shifts.

You will hear it said that “they weeded out the explained collapses” but this rather begs the question: how were they explained? How were these cases different from the ones for which Ms. Letby was charged? Who got to decide what was relevant?

Compare this to Peter Ellis’s trial: there they excluded plainly fabricated statements and kept in the handful of stories that weren’t on their face preposterous. But all statements came from an essentially homogeneous group of witnesses. They were all important context.

In Ms. Letby’s case, some “innocent” collapses might have had salient features in common with the “suspicious” collapses. We don’t know: the defence wasn’t told about them. It cannot be right that the Crown is permitted to determine this behind a cloak of privilege: it must surely be tested in open court. And it is easy enough to test, if all the information is made available.

This is all the more so since the defence’s case was that Ms. Letby was only in court was thanks to misapprehensions, cognitive biases and misdiagnoses by those involved in selecting cases to bring to trial.

It must be germane to know which cases were not selected for trial, therefore, and why. Without this information this is the Peter Ellis case all over again.

Why the other nurses are important

Now an attendance chart asserting only that Ms. Letby was present for all charges would not be controversial. It would not even be very interesting: of course she was present: she could hardly murder by remote control.

But the record of the other nurses’ presence is controversial. None were charged, so it hardly matters whether they were on duty or not. But they should not be on the chart. We should ask what they are doing there.

It is, of course, clear: they are there to make that key identification point, which otherwise is strangely lacking from this case.

There is one common thread here: Lucy Letby.

How did the chart get into evidence?

The staff attendance chart is not evidence but advocacy. It is all the more damning because it looks like evidence. No one seems to have noticed it wasn’t evidence: the defence appears not to have taken the point. So the question arises: just how did it get into evidence?

A summary document compiled from other sources is, on its face, “hearsay” evidence. These days, hearsay may be admitted in criminal proceedings,4 where its “probative value outweighs potential prejudicial effect”. But the methodology used to create the document must be reliable, no better evidence must be available, and the prosecution must call a witness to verify the document, its accuracy and how it was compiled.

None of these seems to have happened. Defence barrister Mr. Myers appears to have regarded the prosecution as “entitled” to present the chart by way of summary of their evidence, though he was critical of it.5 I am no criminal lawyer, and my lecture notes from the Laws of Evidence hail from the far side of the planet and date from 1992, but it is not obvious how this document was even allowed before the jury given its obviously prejudicial effect, and its lack of probative value.

Do get in touch if you know the reason.

How were the charges selected?

It may be that the Crown’s incident review was thorough and methodically sound. It might have scoured every one of the 900-odd shifts, fastidiously assembling its final list of “suspicious” collapses by reference to defensible criteria without regard to who was on duty for any of them. It may be that, having done that, the Crown was left with 22 events from the fifteen months, and every one involved Ms. Letby.

If so, that would likely be the end of it: if the sample is fair, the odds against Ms. Letby’s presence at all 22 events in those circumstances by chance are truly astronomical. That identification evidence would be compelling.

But that is a big “if”.

All the same, it is an “if” the Crown could easily have answered. Mr Myers asked it to, in January 2023. But it chose not to, citing “irrelevance”. Rather peremptorily, the court agreed.

I have just a fragment of a scroll from the application. It is articulated in the impenetrable bluster that passes between silks, so I have translated it back into passable English as follows:

Mr. Myers: We are concerned that the Crown only investigated cases where Ms. Letby was on duty. We are seeking disclosure so we can see to what extent Ms. Letby’s presence was a “search term” for the events the Crown investigated.
Mr. Johnson: Well, it would be odd if we didn’t think there was a solid case against Ms. Letby, wouldn’t it?6 And so, wouldn’t it also be odd if we weren’t looking at cases where Ms. Letby was present?
So, confirming or denying whether we looked at cases where she wasn’t there has nothing to do with our disclosure obligations under the Criminal Procedure Investigations Act.

Mr Johnson’s logic is plain. How are collapses Ms. Letby isn’t being charged with possibly relevant to collapses for which she is? But he made a subsidiary concession that gave the game away: In his full barristerial English he said:

...it would be fairly odd if in investigating what else she may or may not have been up to, they weren’t looking at cases where she was there.

Of course we were only looking at Ms. Letby’s shifts, in other words. And the Court accepted it and upheld the Crown’s refusal to disclose it as irrelevant:

4. [...] Whilst the information sought, or the basis for further police inquiries into the defendant may be of great interest to the defendant herself, they are irrelevant to the issues in this case. So there is no material to disclose, there being no material that “might reasonably be considered capable of undermining the case for the prosecution against the accused or of assisting the case for the accused ... that has not been disclosed to the accused” (CPIA s. 7A). Inevitably, the investigators were looking at cases where the defendant was there, but that has nothing to do with disclosure; [...]
5. In these circumstances, there is no reason to doubt the prosecution’s response to this application that the material requested is not disclosable under the CPIA.

Both Mr. Johnson KC and Mr. Justice Goss appear to regard Mr. Myers’ application as a slightly cheeky fishing expedition: designed see what else the Crown might have up it sleeve that could lead to more charges — as if that was what Mr. Myers was after.

In fact, Mr. Myers wanted to know quite the opposite: how far had they considered collapses that didn’t incriminate Ms. Letby?

Mr. Johnson gave his answer: he hadn’t. The Crown had their woman: it was Ms. Letby.

Let’s return to Peter Ellis for a moment and consider what Ellis was confronted with, compared with what he might have seen had the crown disclosed everything (this is hypothetical data by way of illustration):

Charged with:
Credible allegation of harm by Complainant A
Credible allegation of harm by Complainant B
Credible allegation of harm by Complainant C
Actual Picture:
Credible allegation of harm by Complainant A
Denial of any harm to anyone by Witness D
Non-credible allegation by Witness E
Incoherent babbling by Witness F
Non-credible allegation by Witness G
No knowledge of anything by Witness H
Non-credible allegation by Witness I
Denial of any harm to anyone by Witness J
Credible allegation of harm by Complainant B
Denial of all harm to anyone by Witness K
Non-credible allegation by Witness M
Credible allegation of harm by Complainant C
Denial by Witness N
Non-credible allegation of harm by Witness O

And apply the same filter to Ms. Letby’s situation (this is hypothetical data by way of illustration):

Charged with:
Unexplained collapse Baby A: Ms Letby on duty
Unexplained collapse Baby B: Ms Letby on duty
Unexplained collapse Baby C: Ms Letby on duty
Unexplained collapse Baby D: Ms Letby on duty
Actual Picture:
Unexplained collapse Baby E: Not present
Unexplained collapse Baby A: Ms Letby on duty
Unexplained collapse Baby F: Not present
Unexplained collapse Baby G: Not present
Unexplained collapse Baby B: Ms Letby on duty
Unexplained collapse Baby H: Not present
Unexplained collapse Baby I: Not present
Unexplained collapse Baby C: Ms Letby on duty
Unexplained collapse Baby J: Not present
Unexplained collapse Baby K: Not present
Unexplained collapse Baby D: Ms Letby on duty
Unexplained collapse Baby L: Not present

It tells a very different story.

Interlude: the unique nature of healthcare serial murder cases

There is something else to bear in mind about this case. During the suspect period, it was by no means clear that anyone was attacked, let alone murdered at the Countess of Chester Hospital. At the time they happened, many of the collapses prompted further tests, examinations and post mortems. None resulted in suspicions of foul play.

In murder cases, not being sure that anyone has, in fact, been murdered, is quite an unusual state of affairs.

In serial murder cases — except in one type of case — it is almost unprecedented.

Usually, no-one doubts there’s been a murder. Body parts are found clogging up drains,7 turned into lampshades,8 in shallow graves,9 under floorboards10 or riddled with gunshots.11

Usually, there is strong evidence implicating the defendant: eyewitness evidence, fingerprints on the gun, or it being the defendant’s drain, bath, garden or crawlspace where the bodies were found.

That one exception is healthcare serial murder. This type of case is strikingly different. Usually, there is doubt there has been a murder. Usually, there is no direct evidence implicating the accused. Often, the evidence is entirely circumstantial.

Ms. Letby’s case is a perfect example. The inference that there has been foul play comes, predominantly, from technical evidence: proposed air emboli. Immunoassays. Insulin and C-peptide readings. These are matters requiring deep expertise and subtle inference well beyond my meagre faculties. I am not going to express any view on medical evidence I have no hope of comprehending other than to say that people who are field-trained endocrinologists and neonatologists seem vigorously to disagree with each other about the technical evidence in this case.

Remember, the standard the Crown must attain is beyond reasonable doubt. If experts in the field are disagreeing about it, there is reasonable doubt, Q.E.D.

But this is as may be. We are here talking about evidence specifically implicating Ms. Letby, and no-one else, as the author of categorical crimes, presuming they can be proven to have occurred.

The major evidence in this regard — almost the only evidence in this regard — is her opportunity: Ms. Letby seems to have been the only one on deck when every incident happened.

But if that evidence falls apart, as it seems to have, what remains that identifies Lucy Letby as the perpetrator of any crime at all?