Lucy Letby: have you changed your mind?

[Kittybythelighthouse]

I’ve been sensing a shift in opinions on the Lucy Letby case and I’m interested in hearing from people who have changed their mind either way.

Did you used to think she was guilty and now you don’t, or you aren’t sure? What changed your mind?

Also vice versa: did you used to think she was not guilty but then changed your mind to guilty? What convinced you?

The reason I’m using the term ‘not guilty’ rather than ‘innocent’ is because courts don’t prove innocence. Not guilty is a legal conclusion about whether or not the state met its burden of proof.

Do you have a source link for that new paper please Kitty? I’m having tech issues with them being presented as images, and it seems to start part way through. Thank you.

I completely agree about the one sidedness of the documentary. It pretends to give a balanced account but in reality, any points it makes that support the unfair trial theory are left hanging without counter.

The medical statistician being a good example but for different reasons than you mentioned above. She doesn’t like the use of the table that shows Letby at all the significant incidents they want to include but not others. This is used as an example of being unfair. However, there will be reasons why this data was produced like this, there will be raw data including all the signicant incidents that she was not present for and this will have been disclosed to the defence as undermining to the prosecution case. In fact, the documentary does not explore what other statistical charts were produced at all.
It doesn’t ask when the table was produced, what the criteria was for including those babies and not others. It was a segment early in the documentary and it immediately irritated me as it just made clear what agenda the programme had.

The whole premise of this thread is to drum up support and change people’s minds. People who don’t concur are quickly accused of not caring. It has been made clear that the goal is public pressure. Yes a panel of experts have raised concerns and the matter is being reviewed. As I’ve said upthread, there is no need at the moment for a counter panel of experts to argue and rebut as the trial has occurred.

The OP is so invested in this that I doubt they would accept the findings of the ccrc if they didn’t give the desired result.

However, there will be reasons why this data was produced like this, there will be raw data including all the signicant incidents that she was not present for and this will have been disclosed to the defence as undermining to the prosecution case.

Will there be?

How is it ironic? This poster is referring to an actual guilty verdict, so it’s not the same thing.

Yes.

The on duty list was accurate, she did do those shifts.

And she was there when all the babies died.

Just out of interest, particularly for those on the thread who feel the stats presented at court were sound, evidence based, and presented fairly and accurately in a way that a non statistician could see the full picture, understand them, and draw conclusions from them, here are the top ten hospitals that saw unusual jumps in deaths in neonates in 2014 - 2015 (as an aside, nationally there was an overall jump in that time period, just for an interesting fact)

Each bar represents a hospital.

Which bar on the graph represents the Countess of Chester?

Photo below

What do you think those reasons would be?

Because a coroner had found the deaths suspicious and further more in depth forensic testing was done as is standard when a pathologists findings point towards unnatural causes?

Because the other Drs outside of the countess of Chester who reviewed the deaths before Dewi Evans became involved found anomalies/evidence that raised alarms in some of the deaths?

For anyone interested in this case, I found this recent interview insightful.

It’s around 90 mins but flies by, as Michele is so engaging and easy to follow:

I totally agree with this.

The problem with the stats is that they have not been collected and produced using statistical method.

To be valid, the notorious chart should be independent of any subjectivity and must include all the data not simply cherry-picked (Texas sharp-shooter method). It also needs clear medical definition of “incident” and, preferably, as this is a murder enquiry, hard evidence of actual murder.

Leaked info from the prosecution found that the initial tally from Evans of “suspicious” events was 28, but when LL was found to be on duty for only 18, 10 were deleted until the chart indicated LL’s presence at “all” suspicious events.

Whether the police mangled the stats because they had no idea how to use them or because they wanted to manipulate them to support their case is not clear. The CPS is partly to blame as it requested the police cancel their instruction of a professional statistician to review the data.

I don't think she received a fair trial, and if the deaths were murders which Letby perpetrated, then the prosecution has badly let the families down because the conviction is not safe.

For me, there's always been a bit too much Amanda Knox flavour about this case, particularly how Letby was gossiped about by other staff and the prevailing belief that people raising doubts are doing it "just because she's pretty" (while failing to interrogate their own biases on that matter, which was also a factor with Knox - who along with being pretty was seen to act 'oddly' and liked sex - there are parallels to the Letby case though in different ways), to really think she's guilty. I've always felt it carried the scent of a witchhunt - Letby was pretty, present, acted oddly, liked a particular doctor, and multiple other female staff were bitchy about her. To me it looks like they went looking for murders and a murderer, rather than an investigation leading to that conclusion. And if they're incorrect in their conclusion, and families have been led to believe their babies were murdered when they weren't - that would be absolutely unforgivable.

I am also biased as my nephew was born very early, spent time in NICU and had two collapses out of nowhere. He nearly died twice. It happens with very ill or premature babies. Standard of care should be looked into when collapses and deaths are happening a lot on a unit, but the fact of them happening isn't suspicious in and of itself. I've always wondered why the jump was to murder, and the methods described seem bizarre and the evidence circumstantial at best. If I were a NICU nurse this case would make me want to wear a body cam at all times on shift just in case.

To add to that, it appears that Lucy wasn't present for all of the events shown on the prosecution’s chart either.

Im no statistician, but this man cited below is

https://journals.sagepub.com/doi/10.1177/00258024241242549

I’d imagine the average person off the street is going to struggle to understand his explanations.

What chance does a jury have of being able to discern if a set of ‘stats’ presented to them are logically sound?

The only other link I can find is to Dropbox - https://www.dropbox.com/scl/fi/bxrjyn796a7pau8dxq6kl/Letby-Briefing-Paper.pdf?rlkey=dkvzwvb5mvgpyejv6c7x1jots&e=2&dl=0

Have you watched the ITV one yet? That goes into much more detail.

Fab thanks - much better!

Absolutely. “Present” in LL’s case is rather elastic in its application.

Wow. There’s a lot to unpick here! First, the point of the documentary was to show that there are deep problems with the prosecution’s case. That was their editorial line. This is entirely valid, which is why the documentary has been praised and highly reviewed across the board. If this was out of line other journalists, particularly pro prosecution journos, would have been very vocal about it. Besides, it would be literally impossible to find a statistician to defend the rota because it is that bad. It is literally indefensible. Have you ever seen a statistician defend it? Even in court? No. That’s why and I’ll explain further below. They did approach prosecution witnesses for comment btw.

Did you feel anything was “one-sided” during the previous 5ish years where ALL of the public narrative was focused on “Ooh she’s a bad b*tch that one” with a constant stream of misapplied, or straight up incorrect, data and information almost every day? Did you take issue with the Moritz/Coffey Panoramas which have a public duty to provide balance (BBC “due impartiality”) but were riddled with errors and no real attempt to provide balance? Already we know that tonight’s documentary speaks about Dr Dimitrova’s report and baby O without handling the report itself or speaking to her directly.

In the first Panorama, to give an example, graphics highlight Dr. Brearey’s note on the 22nd of June 2015 that the only personnel all of the three first deaths had in common was “one nurse”. Except this was not true, as it later transpired, that Elizabeth Marshall was also on shift during the first three deaths. Was this ever addressed or corrected? No. Did we ever find out what doctors, orderlies, respiratory therapists etc were on shift and when? No. Funny that. Are only nurses capable of murder?

If you are as concerned about that as you are about this documentary it might seem less like you merely don’t enjoy your “side” quite rightly being challenged after years of pro prosecution narrative control and misinformation.

Regarding the table. As I say, no statistician would or could defend it, which is why none ever has. You make a lot of assumptions about what would have been done in order to produce it. The problem is that the prosecution did not use an actual statistician to analyse the data and when the only one they spoke to said “I’ll need ALL the data to make a proper analysis here” they fired her and never engaged another statistician.

“In fact, the documentary does not explore what other statistical charts were produced at all. It doesn’t ask when the table was produced, what the criteria was for including those babies and not others.”

Let me help you, that’s because there was no statistical work done at all. If you don’t have all the data e.g all the other staff (again, why just nurses btw?) or how cases were picked, you can’t treat the chart as reliable no matter which “side” you’re on. Calling that out isn’t pushing an agenda, it’s the bare minimum for trusting a piece of evidence. This is quite simply not how stats work. It’s not how evidence works either. You’re asking the documentary, the general public, and every statistician in the world just to “trust me bro” on this chart? Are you taking the p*ss?!

“The whole premise of this thread is to drum up support and change people’s minds. People who don’t concur are quickly accused of not caring.”

This is such a misrepresentation. When people (including myself) asked “don’t you care about a miscarriage of justice” it’s in response to the constant appeal to emotion “Don’t you care about the parents?” being used as an attempt to shut down conversation. No one has countered a factual argument with no factual rebuttal and just said “don’t you care about a MoJ?” If I’ve missed it please link to it if you can.

If you or anyone else cannot counter the factual rebuttals that are made then your response to that should be to seek better arguments or change your mind in line with the logic that has been presented, not to moan and misrepresent how you were shut down with an appeal to emotion when this has simply never happened.

”The OP is so invested in this that I doubt they would accept the findings of the ccrc if they didn’t give the desired result.”

I have been clear always that my interest in this case is purely about the fact that I find it genuinely quite scary that our justice system could be in such poor shape that something like this can even happen! That worries me because if it could happen to one person it could happen to any of us. In fact it already has happened to 900 postmasters and multiple other recent MoJ.

The truth is that I’ll be delighted if someone credible presents something solid, and based on the best science, that proves that she did it and it’s all okay actually. Having researched this case for over a year I just don’t think that’s going to happen. The evidence is just that bad. That worries me. It should worry all of us.

@Spaghettimoth oo no I haven't, I didn't realise there was another one on ITV because we don't ever watch that channel but I do watch iPlayer. I'll check it out.

Which is the new bbc documentary you referred to? I don’t think there has been a new one. Do you mean the second panorama?

No. There won’t be. And we know this for a fact. The prosecution never engaged a statistician after they let Professor Jane Hutton go when she said “I’ll need all the data”. A full statistical analysis was never done.

Well said. I’m just re-listening to the The Teacher’s Pet podcast (where Chris Dawson was convicted on circumstantial evidence of killing his wife) as I missed the last few episodes the first time round and the host makes the point that circumstantial evidence is often the best kind of evidence because it’s direct evidence.

Ffs. This is why dodgy stats are an issue. People with no grounding in statistics think it’s “common sense” but it isn’t. It’s easy to convince lots of people using dodgy stats that appear convincing, including doctors and judges btw. Both professions have form for being very bad at statistics in seriously damaging ways. This is why the Royal Statistical Society produced a report highlighting the dangers of courts mishandling stats, which is a constant issue.

As a piece of statistical evidence, the rota commits selection bias, post-hoc inference, omission of key covariates, and failure to adjust for exposure and multiple testing. Restricting the view to nurses while excluding doctors and other staff compounds confounding and suppressed-evidence problems. Even before you debate causation, the design does not meet basic standards for inference.

You can tell the chart is unreliable just by its design. It wasn’t a neutral plot of all incidents and staff shifts; it was a filtered presentation built on the assumption of her guilt. This is why statisticians say the underlying probability calculations are meaningless. The dataset itself is already contaminated by confirmation bias.

Journalists and defence witnesses have noted there were other deaths/collapses in the same time frame with Letby not on shift, which were absent from the chart.

Dewi Evans himself admitted some deaths initially flagged as suspicious were reclassified as natural when Letby was not present, but these were excluded from the prosecution’s rota graphic.

There's a new article by Coffey today

https://www.bbc.co.uk/news/articles/cj0y9673rjno