How to lie with data

This spreadsheet was shown on TV both yesterday (Friday August 18, the day of the verdicts) and at the start of the trial of Lucy Letby. Apparently, Cheshire Constabulary find this absolutely damning evidence against Lucy. And indeed, many journalists seem to agree.

The 25 events are almost all of the events at which LL was present during the periods investigated. They are suspicious because she was under suspicion when the police started their investigations. Not surprisingly, most nurses are not present at many of these events. And of course, many nurses probably work far fewer hours than LL. Many are often on administrative duties.

The doctors on the ward are of course missing. Doctors were never investigated as suspects but from the start of police investigations apparently always believed to speak gospel truth. During cross-examination, during the trial, some of them have changed various parts of their stories. Of course, unlike Lucy, they do not lie, since they could never (under oath in court, or earlier, when being interviewed as witnesses by police) be saying untruths in order to deceive.

Back to the spreadsheet. When drawing conclusions from any data it is important to know how it was gathered. It is important to know what data is missing, but would be needed draw even the most preliminary and tentative inferences.

There was an NHS investigation into the raised rates of deaths and collapses at Countess of Chester Hospital (CoCH) in summer 2015 and summer 2016. It was published in 2017 by the Royal College of Paediatrics and Child Health (RCPCH). The investigation blamed the consultants for the appalling low standard of care, and the terrible situation regarding hygiene. The RCPCH investigators actually wrote that nurse Lucy Letby could not be associated with the events, but that passage was redacted out of the published report for privacy reasons. We know that already, consultants had presented their fears to hospital management. One of them (successful TV doctor and FaceBook influencer dr Ravi Jayaram) was on TV yesterday proudly telling the world that he had been vindicated. Management was inclined not to believe them, and did not act on them, but they certainly came to the ears of the RCPCH. On publication of the report, four consultants had had enough, and went to the police with their suspicions that LL was a murderer.

Thanks to FOI requests and statistical analysis by independent scientists, we now know that the rate of events (deaths and collapses) is just as much raised when Lucy is not on the ward as it is when she is on the ward. A lot of medical information (as well as the state of the drains at CoCH) points to a seasonal virus epidemic.

The elevated rate went back to normal after the hospital was down-graded (no longer accepting high risk patients), and when the drains were rebuilt, and when the senior consultant retired, all of which happened soon after the police investigation started. Incidentally, the rate of still-births and miscarriages show exactly the same pattern.

Lucy must certainly have been a witch in order to kill babies in the womb and even when she is far from the hospital.

Those familiar with miscarriages of justice involving serial killer nurses will be familiar with this police and prosecution tactic. Is it evil or is it just stupid? (cf. Hanlon’s razor). I think it is quite simply “learnt”. Police and prosecution learn what convinces jurors over the years, and that is why the same “mistakes” are made again and again. They work!

31 thoughts on “How to lie with data”

  1. County Chesshire is very biased due to the general practitioner Shipman, who was convicted for 15 murders. Probably there were up to 250 victims. Is a clean legal process is possible?

    1. Peter Elston is publishing various statistical analyses on his (“Chimpinvestor”) blog. I would like to do further analyses with data which presently is not available. The defence team could help me get that data, but they do not respond to requests for information or offers to help. I did not yet manage to get contact with Lucy or her parents. But I keep on trying. It’s going to be a long haul, anyway.

      1. If you eventually manage to gain contact with Lucy or her parents.
        Then one would hope if she really is innocent you may provide a glimmer of hope.

  2. Dear Richard, thank you for all the work you do to highlight the facts in the Lucy Letby case. I find it shocking that a case was brought against Lucy Letby on such “evidence”, and her conviction is enough to shake all belief in so-call “justice of the courts”. I do hope that help can be given to Lucy and that she will have support from those believing in her (especially since she has already been severely traumatized by the whole proceedings against her) and that she willl have a better legal team to fight for her. Those who were involved with her defence have done a shockingly bad job. I can´t quite get my head around why no better defence was put up for her and why no experts in support of her were brought in. It seems that from the beginning her guilt was assumed, with no presumption of innocence – both in the courts and in the press.

  3. Thank you for this article Dr Gill.

    “Thanks to FOI requests and statistical analysis by independent scientists, we now know that the rate of events (deaths and collapses) is just as much raised when Lucy is not on the ward as it is when she is on the ward.”

    Though I trust that you would not claim this unless it was so, I would really like to see where it has been established. I’ve looked at the FOI requests referenced in another article and the article on Chimp Investor, but can’t find anything that makes this claim. If true, it would be absolutely critical.

    1. Chimpinvestor has three posts now on the case. Also, we know that the RCPCH investigated this very question, and found that answer, and concluded that the events weren’t linked to Lucy Letby. This result was redacted out of the published report in order to protect her anonymity. Immediately after the report was published, the four consultants – whose work was severely criticised in the report – went to the police. Almost immediately, dr Dewi Evans drove up to Chester and offered his services to the police.

  4. You comment that most nurses are not present at these events. But is it not significant that at all of them there were about six other nurses present as well as LL, and that as I recall, none of the nurses ever suspected LL of doing anything wrong? One would have thought that with 6 others present, that someone would have noticed suspicious behaviour, yet none spoke up. Or am I missing something?

    1. No, you are very perceptive. Cheshire police, BBC Panorama, the jury, and the British public, thinks the table proves guilt. I think it strongly suggests innocence, and it strongly suggests that the police are incompetent and the defence is incompetent too. Well: the police are competent at getting convictions. That’s what we pay them for. Who cares whether the convicted are innocent or guilty.

      1. I took an interest in this because I’m a data scientist, working with medical data. I guess what we really need to know is how many of the available shifts did LL work on (compared to her colleagues), and how many shifts where no incident occurred. Would it then be possible to work out a p-number for her attending those particular shifts, as I believe you did in the case of Lucia de Berk?

  5. Dear Richard, the media are reporting that Lucy Letby may be charged with 6 additional deaths, and it is said (see a report by Sky news dated 19 August 2023) that “There were 13 deaths on the neonatal unit where she worked over a one-year period … and the nurse was on duty for all of them”. You say above that “Thanks to FOI requests and statistical analysis by independent scientists, we now know that the rate of events (deaths and collapses) is just as much raised when Letby is not on the ward as it is when she is on the ward”. I think it is becoming important now to explain what the evidence is for that claim. It is indeed disturbing that the chart is widely understood to show that Letby was on duty every single time a baby died or collapsed on the ward when it plainly shows no such thing. But the suggestion now is that Letby was present at least for all of the deaths even outside of those that are recorded in the chart…

    1. Prof. Gill replied to my comment above asking the same question, but I can’t seem to find evidence for this point in any of the Chimp Investor articles. I think the entire question of whether Letby has grounds for appeal hinges on this, because the ‘medical’ evidence is so abysmally weak. If she was present for every single one of the deaths over a one year period then it would at least rouse suspicion (but would of course on many other factors).

      1. Yes. I agree that the Chimp Investor articles do not give hard evidence on this point. Indeed, the Chimp’s 15 July article states, of the other deaths on the unit, “it would seem that the reason that Letby was not charged with any of them was that she wasn’t on duty”. This does not seems like an unreasonable starting assumption in the absence of more information. However, media reports have it that there were “13 deaths on the neonatal unit where she worked over a one-year period… and the nurse was on duty for all of them” (see It is unclear what the source for that claim is. My impression is that anything to do with cases for which Letby was not charged was excluded from the trial so it is unlikely to come from evidence given at the trial itself – but, again, I’m not sure.

        I agree that this is the absolutely pivotal question. If the jury were mislead into thinking that Letby was present for all of the deaths on the unit whereas in fact she was there for only around half of them, as the roster chart implies, then the conviction is plainly unsafe – both because this suggests there was no extraordinary coincident in her involvement (as you may note everyone who thinks she is guilty brings up the “unlikeliness” point) and because it may suggest (per the Chimp’s analysis) that there was some other factor (eg a virus) which was killing children on the ward. Conversely, though, if she was on duty for all of the deaths then I am afraid it is much harder to believe there is a miscarriage of justice here, notwithstanding the criticisms of some of the other medical evidence in the case.

    2. Thank you (and the Repliers) for focusing on this. It does seem to be a very important matter to clear up. I notice there have been various FOI requests in this area. I’ll try to find them, and if any speak to this point, I’ll post the information here—with the site owner’s permission.

  6. I am afraid all I can find is an FOI request which I expect is already well known to you all, giving monthly deaths by various categories.

    Assuming (i.e. I do not know) that the police define a “suspicious event” as a medical emergency that occurred while LL was present, then the infamous Chart 1 showing LL present at all suspicious incidents on the ward also lists only the deaths LL was on the ward for (and is therefore an instance of circular reasoning, as has been pointed out). The FOI chart says there were 14 deaths in the period June 2015–June 2016, implying (if the assumption is correct) 7 deaths when she was off rota. Still it would be nice to see some explicit rota information covering the other deaths, so that this does not have to rest on assumption.

    (I feel I should add that it is horribly distasteful to be speaking about the deaths of babies in such instrumental terms. I am well aware of the almost sacred atmosphere that rightly prevails around a vulnerable baby ward. But there is nothing to be done about that for now.)

    1. And I now see that the Law Health and Technology substack post knows a lot more than I do! Sorry — I’m learning about this case as I go.

    2. Is that file accurate? It lists no neonatal deaths in August 2015 or October 2015. Two of the deaths that Lucy was convicted of occurred in those months. I also believe that it is wrong to only be suspicious of the deaths when Lucy was present. Her being present shouldn’t be the reason for suspicion, because that makes it impossible for her to be absent.

      1. Which file? All data sources disagree with one another. Nobody knows the truth. Definitions are not the same. A death in days 0 to 6 can be filed as a stillbirth. Both doctors and parents may well prefer that. No death, no post mortem. Better statistics. Grieving parents will often prefer it, for obvious emotional and social reasons

      2. Hi Richard, I couldn’t reply to your reply. I was referring to the file showing deaths per month (B.B. posted a link to it). A death in days 0 to 6 should be recorded as early neonatal, but I could understand why they might record as stillbirth to improve the hospital reputation. I can’t remember precisely, but I believe that one of the unrecorded deaths in August 2015 or October 2015 was of a newborn baby, so that could be recorded as stillbirth (2 in each month). But I believe the other unrecorded ‘murder’ was of a post-neonatal baby (28-364 days). I think he or she was about 2 months old, so couldn’t be recorded as stillbirth.

  7. Hello Richard.

    I’m a journalist looking into some of the issues raised about the prosecution case. You are raising some interesting and concerning questions.

    I do think trust in the institutions/process is important though, so we should be really clear about the facts and evidence we’re using to raise these issues.

    There’s a couple of things above where I’m not certain where the info has come from – could you help?

    Firstly, this section: “It was published in 2017 by the Royal College of Paediatrics and Child Health (RCPCH). The investigation blamed the consultants for the appalling low standard of care, and the terrible situation regarding hygiene. The RCPCH investigators actually wrote that nurse Lucy Letby could not be associated with the events, but that passage was redacted out of the published report for privacy reasons.”

    This doesn;t match my understanding and reading of the report and the unpublished addendum (which I have a copy of) – could you point me to the specific bits you’re referring to?

    Secondly, this section: “Thanks to FOI requests and statistical analysis by independent scientists, we now know that the rate of events (deaths and collapses) is just as much raised when Lucy is not on the ward as it is when she is on the ward.”

    Again – could you point to precisely which FOI data and analysis?

    Many thanks

    1. The *published* report is a censored version of the full report, which has remained confidential.

      My main blog post on the Lucy Letby case, pinned to the top, gives links and references to sources, including data.

      1. Yes, I believe I have a copy of the full report inlcuding the redacted section. Could you share or publish exactly what you are referring to, so we can cross check?

        Thanks, I’ve had a look at the references and data, but the data doesn’t tell us whether or not LL was on the ward. It just details the number of deaths per month.

  8. Dear Lawrence, we know exactly when Lucy was on the ward. January 2015 to July 2016. I can find the date in July when she was transferred to administrative tasks.

    The ward was downgraded when the report came out.

    But, I believe that a full and confidential version also named Lucy by name.

  9. Hi -I’m not sure there is data anywhere showing all the days that LL was on shift? She certainly wouldn’t have worked every day between Jan 2015 and July 2016. What we’d really need to see, in order to properly question this, is whether she was working on the days when incidents happened.

    The NICU was actually downgraded in June/July 2016 – the RCPCH report wasn’t finished until Nov 2016, and not published until early 2017.

    The version I have does not name LL, but refers to an individual nurse. It doesn’t sound like you’ve got a copy of it?

    1. Please, please email me your copy!

      And yes: what we need is the complete roster data of all nurses. The police must have had it. But instead they prepared their silly spreadsheet. They are hiding critical evidence, which was in their possession, from the jury and from the public, because they are too dumb to interpret it; too dumb to know what kind of expert was needed; and too blinded by ambition to pay any attention to the obvious.

  10. Will do –
    Please bear in mind it was commissioned by the trust management, who had believed and sided with LL.
    It was not set up to actually look in detail at the concerning cases, it was simply an overall review of the dept. It said another review was needed to look at the specific cases (the Jane Hawdon review, which suggested another ‘forensic’ review was needed).

    With the spreadsheet – I’m trying to check this, but I believe this may have been the eventual output of the process, as opposed to the starting point (which is how the media have presented it). There’s a danger here that people are commenting on the media summaries of the trial (which are understandably simplfied for their audiences), without actually knowing the full evidence that juries heard and saw.

    In the i/v below, which is worth listening to, Dewi Evans says he was initially sent the case notes of 32 infants where they had died or collapsed from early 2015 and July 2016. He says he identified 15 of these where he believed there had been foul play – and asked Police to check the staffing rotas for those 15 cases. And I think that’s what the chart may represent.
    I’m not saying that’s an appropriate/inappropriate way to do it – but it certainly seems more appropriate than what being presented and interpreted.

    1. Yes, the spreadsheet came out of Dewi Evans’ work. He was given medical records of 32 events. This led to 22 charges by the CPS. The rows of the table correspond to 22 events, not 22 babies. Now we should ask: who chose the 32, which were then whittled down to 22 by Evans? There is a suggestion that these are all deaths and collapses, but that suggestion is false. They are all the deaths in the period when Lucy was on the ward, together with arrests during Lucy’s shifts in that period. Maybe there were five times as many arrests in total during that time interval? I believe that the 32 cases were compiled by the medical consultants. What is the definition of an arrest, by the way? Nobody has given us a definition.

      Dewi has correctly identified 10 events which are obviously completely natural. What’s left are six or seven deaths which occurred on Lucy’s shifts together with almost all arrests during her shifts. He has spotted some tiny odd thing in each of those cases, fantasized a murder or attempted murder, and it is no coincidence at all that Lucy was there every time. She was framed by the initial selection of 32.

Leave a ReplyCancel reply

Exit mobile version