Lucy Letby case expert witness was under fitness to practise investigation

[ITMA2000]

Wow! Pot, kettle, black? Need a new investigation?
Lucy Letby case expert witness was under fitness to practise investigation during trial | Lucy Letby | The Guardian

Also, getting a bit more basic here, what does it even matter if she was on shift when the poor babies died?

Did she kill babies with her special black magic powers just by her presence?

There were other people in the small room often actively trying to treat the babies. She quite possibly wasn't even thr designated nurse for the babies. Yet no one noticed anything?

And if she is supposed to have done things to them earlier, leading to them dying later then surely it could have been anyone who had been in the ward in the days preceding their deaths?

Why wouldn't he count deaths that happened at other hospitals? I thought the whole statistics argument was we need ALL the data? Not just when it suits the Letby camp. You also (think it was you) said that babies unfortunately can take a while to die. So if she was there at point of collapse it's very relevant.

Also we should be counting times she was there but not charged with harming the baby, because we have no idea if she still did but there wasn't enough evidence. It's very important data to leave out. So if there's not enough data for the Letby camp there's not enough for the guilty camp either. What we have makes her look pretty bad, the full picture would be even worse.

None of your other statistics probabilities mean she's innocent.

You need to prove murder, not just be unable to rule it out. If all you've got is, someone could have invisibly murdered this child despite indications they died of natural causes, and our proof is that nurse was there at some point, you're back to relying on a statistical case - but with insufficient data.

She went in when other nurses left for short breaks. As is the case with baby K, who the designated nurse had left LL "babysitting" whilst she went to talk to the parents. Don't you think it's odd all these suspicious collapses were happening just in the very short time one nurse (who wasn't even designated nurse sometimes) went in the room?

Well I think we now get back to what the others are saying. The prosecution defined collapses as suspicious based on her presence. Thats why the statisticians wanted details of all the collapses.

Yes, he can count the deaths that happened at other hospitals, but obviously Lucy Letby wouldn't have been at the other hospitals

So you have to assume he assigned a precipitating event, if he's saying she was present. And the problem then is that in the police investigation and at the trial, he was extremely flexible with his precipitating events. If Lucy Letby was on shift and event was significant. If not, it was dropped. Happened with Baby C (twice), baby B, baby I, baby J, baby O, baby P, baby Q - may have been others I'm forgetting. So the minute he's working outside the black and white, they died while she was actually on shift, he is clearly very willing to tweak the data against her. He gets lots of bites of the cherry - was she on shift? Was she on the shift before? How about the shift before that?

What looked like a striking figure - 15 in 17 - quickly becomes an illusion.

Let's remember too that Evans himself said that the deaths not included on the charge sheet had known natural causes. He even listed the causes. So why is he implying it's suspicious that Lucy Letby was on duty for (or before) some of them?

Nobody has argued that statistics can prove Lucy Letby innocent. But they can be used to argue that Chester's increase in deaths didn't mean murders, and that Lucy Letby's presence didn't mean murder. Does that mean there was no murder? No. It means you have to look beyond the statistics to prove murder.

Nurses would always be on their own with babies at some points during night shifts, because staffing levels were minimal. So if babies were going to collapse, nurses would often be alone with them. How often did babies collapse when other nurses were with them? We don't know.

It’s not about how much data there is though, Firefly, is it? We’ve been through this multiple times, as you know.

It’s making sure it’s the appropriate data, that it’s clean, relevant and then interpreted accurately to reach valid conclusions.

The point for statisticians is to determine which way data points, regardless of bias. I would go as far to say that they are not concerned with which outcome is reached, but rather that the accurate one is reached.

You are very concerned about ‘camps’ and ‘bias’, so are we, and in that case, I’m surprised you don’t think that the statisticians are the way to go here. You have often inadvertently argued for them, within this thread.

those deaths when she is around are phenomenal - even after all the possible reasons

To be able to say that, you need to first do an analysis that accounts for all possible reasons.

Why wouldn't he count deaths that happened at other hospitals?

Because he's counting "deaths during Letby's shifts". If he wants to count something else, he has to define what.

Also we should be counting times she was there but not charged with harming the baby

For comparing rate of deaths per shift between Letby/non-Letby shifts, yes. But that's why you have to check for group differences between Letby/non-Letby shifts on factors that can affect the death rate. If you're going to count deaths against her regardless of whether they were natural, you have to show there's no reason to expect the rate of natural deaths to be higher on her shifts.

Because he's counting "deaths during Letby's shifts". If he wants to count something else, he has to define what.

@Dolphin37 Why? Letting us know that babies died at other hospitals yet Lucy had the opportunity to attack them (and was present when they deteriorated) is vital information. If we didn't have that info the stats would (wrongly) be more favourable to her.

For comparing rate of deaths per shift between Letby/non-Letby shifts, yes.

Yes that's what we want. Why the need to overcomplicate it? Afraid the stats make her look guilty?

If you're going to count deaths against her regardless of whether they were natural, you have to show there's no reason to expect the rate of natural deaths to be higher on her shifts.

You can't just pick and choose which deaths when we don't know if she harmed others that she wasn't convicted of due to not enough evidence. That's not accurate data. If she was innocent why would you have a problem with them including all of the deaths?

Yes, he can count when babies died at other hospitals, but pinning down the precipitating event is much more flexible than pinning down the time of death. So he can't present this as a simple counting exercise. He needs to show how he has chosen the significant event leading to their transfer and / or death.

@Dolphin37 isn't suggesting picking and choosing which deaths are counted. She is explaining that you need to put all deaths in context.

@Oftenaddled

Yes, he can count when babies died at other hospitals, but pinning down the precipitating event is much more flexible than pinning down the time of death. So he can't present this as a simple counting exercise. He needs to show how he has chosen the significant event leading to their transfer and / or death.

Agreed.

isn't suggesting picking and choosing which deaths are counted. She is explaining that you need to put all deaths in context.

Yeah but I'm saying we don't know the context. It's like doing stats for Harold Shipman and not counting the hundreds he wasn't charged with. Or claiming they were all natural so they go into a different category.

Because he's counting "deaths during Letby's shifts".
Why?

Ask him -- that's what he says he's counting in his emails with Elston.

If you want to instead define "death during Letby's shifts" as "death of baby who was at CoCH during any Letby shift", then you have to define "death during non-Letby's shifts" as "death of baby who was at CoCH during any non-Letby shift" for a proper comparison between the two sets of shifts. Several babies will likely end up in both sets. Evans doesn't give the numbers under this definition.

You can't just pick and choose which deaths

I'm not. I'm saying: take all deaths and all shifts, fix a definition for assigning a death to a shift, and apply it consistently. It can be "death happened on the shift", or "death happened to a baby who was cared for on the shift", or something else. It just has to be a definition that can be applied unambiguously to any death and any shift, to get a yes/no answer, without needing to know if it's Letby's shift or not.

You can then count deaths assigned to Letby shifts, and divide by the number of Letby's shifts, to get the death rate per shift for Letby shifts; and do the same for non-Letby shifts. If the rates differ, you can compute the chance of getting such a difference by chance on the assumption that rates of natural deaths are the same in the two sets. But the assumption is only valid if, on every factor that can affect the rate of natural deaths, the two sets of shifts (Letby and non-Letby) are similar. They have similar fractions of night shifts. They have similar fractions of shifts with 25-weekers. And so on -- not all factors may be known, but at least the known ones must be controlled for.

So, no, you don't need to deal with individual deaths' "context" (hard to do objectively), but you do need to show that no relevant aggregate metrics differ between the Letby/non-Letby sets of shifts.

Big mad 😆 Absolutely ridiculous.

But then she did suggest previously that maybe Shoo Lee believed she was guilty but was mad (presumably very big mad) that he didn't get enough credit for his work so he then set out to assemble an expert panel for revenge.

Absolutely zero understanding of normal human motivations, just ridiculous, nonsensical theories.

The data collected would simply just be 'Cases where Dewi thinks LL may have harmed a baby' and we already know that Dewi suspends his medical knowledge in order to deem things suspicious.

You say why not count it but what's the purpose of that data? What do you think it can be used for?

Let’s be very, very clear here.

Neonates receiving breathing assistance often get air in their stomach and intestines. Google “CPAP belly” for any number of scholarly articles. This is problematic, but not life threatening.

Injection of air via a NG tube is not a method of murder because it isn’t feasible that it would cause death, not because air in neonates’ stomachs has never been observed.

I still think you don’t really understand the basic anatomy and physiology relating to these supposed methods of causing harm, which is resulting in you going off at tangents, repeatedly.

I thought that was the total deaths for the unit so all included

The number corresponds with the 18 deaths between June 2015 and July 2016 reported at Thirlwall.

(The 18th baby died in the labour ward so didn't spend time in the NNU)

During this period, 13 babies died on the ward and four after transport to other hospitals.

The cases are still selective, in two ways:

1. Why only deaths between June 2015 and July 2016?

1. How can you claim Lucy Letby was present for the deaths of three babies at other hospitals? She didn't follow them there. So Evans can only mean, he picked out significant incidents before they were transferred. But his track records shows (for babies B, C, I, J, N, O, P and Q at least) that he changed his mind about which incidents were suspicious or led to death so that they all matched up with Lucy Letby's shifts.

She was on duty for ten, not fifteen of those seventeen deaths, as Thirlwall documents show. We know seven deaths were prosecuted. If she is innocent, it is reasonable to assume that she would also be there for deaths by natural causes not on the indictment sheet. This could well be because, as Elston implies with his questions, her work pattern put her on duty when the most vulnerable babies were on the ward, at the time when they were most vulnerable and the unit was worst staffed.

Evans can only mean, he picked out significant incidents before they were transferred

If he did that to assign deaths to Letby shifts, he has to do the same to assign deaths to non-Letby shifts. Per his numbers, none of the 15 babies he counts for Letby shifts had any significant incidents while she was off. From timeslines I've seen, that's not so. Such an analysis would also only be valid if "significant incidents" were picked without knowledge of Letby's name and suspicions against her, which wasn't the case here. He could skip reliance on "significant incidents" and just count a death to be on Letby shifts if the baby who died (at CoCH or elsewhere) was at CoCH during any of Letby shifts. But then he'd need to do the same for non-Letby shifts.

@Firefly1987 @Rhubarbandcustardd If you want a sanity-check on the points being raised here, I suggest pasting a link to these threads into ChatGPT or Claude and asking what it thinks. It is by no means an arbiter of truth, but it can identify considerations people have missed -- you can then decide for yourself if it has a point. I've found it useful that way.

ChatGPT points out key flaws when I’ve posted stuff said on here before so its 🤷🏻‍♀️

and just supports my pov and why people claiming MOJ are being over zealous

last time it concluded CCRC wouldn’t grant appeal on the reasons put forward in these type of posts

ChatGPT points out key flaws when I’ve posted stuff said on here before

Share an example?

I really would be careful with AI and legal stuff.

https://thebarristergroup.co.uk/blog/responding-to-ai-hallucinations-in-uk-jurisprudence

https://www.legalfutures.co.uk/latest-news/court-issues-stark-warning-to-lawyers-over-ai-generated-fake-cases

https://freemovement.org.uk/tribunal-slams-two-immigration-lawyers-for-suspected-citation-of-ai-invented-fake-case-law/

Not saying there's no use or benefit for it, but it also makes things up! Best taken with a large pinch of salt and enough idea what you're talking about to be able to identify obvious errors.

Yes I’d be interested in this.

I can’t get ChatGPT to admit that air via a NG tube was considered a cause of death, only that it might have contributed to a collapse. It’s also very keen for me not to mix up AE and air via NG tube.