Professor Jane L Hutton, Department of Statistics, The University of Warwick has sent a complaint to the BBC about the intubation dislodgement statistics misused in Panorama. Images of her complaint attached.
“Dear BBC complaints team
Panorama - Lucy Letby: Who(m) to believe?
I have been asked to comment on the numbers discussed by Judith Moritz and Jonathan Coffey on unplanned extubations (UE, accidental dislodgment of endotracheal tubes) at the Liverpool Women's Hospital (LWH). They essentially repeated information which Mr Richard Baker KC gave to the Thirlwall Inquiry on 12 September 2024:
...You will hear evidence that it generally occurs in less than 1% of shifts. As a side note, you will hear that an audit carried out by Liverpool Women's Hospital recorded that, while Letby was working there, dislodgment of tubes occurred in 40% of the shifts she worked...."
The evidence was not provided.
The judicial colleges endorse the principles that' 'When conclusions based on statistical science are drawn from data, it is crucial that the data and the reasoning supporting those conclusions are transparent" [The use of statistics in legal proceedings: a primer for courts 2020 Royal Society and Royal Society of Scotland]. No information from the LWH audit of UEs (the definition of dislodgment of andotrachael tubes, design, conduct or analysis of the audit) has been released to Letby or her defence team.
The journalists said that LWH used "ventilated shifts", a 12 hour period during which a baby is ventilated (rather than the usual ventilated days). They used the assumption of 10 ventilated babies on a unit for five days, 20 shifts, and stated that the UE rate was less than 1%. A systematic review published in 2012 [da Silva et al, Respir Care 2013;58(7):1237â1245] reported:
"UE is a common event in the NICU"; the median UEs/100 intubation days was 1.98.
This is 4% of ventilated shifts; the authors note that note as UEs are recorded.
The journalist claimed Letby worked about 50 shifts, without any reference to the number of babies ventilated, and said 'the (sic!) tube came out on around 20 of them. It's about 40%.
This is dreadfully bad arithmetic and seriously misleading journalist. There are 24 high dependency or intensive care cots at LWH. I have to make assumptions, in the absence of transparency. Suppose only ten of the 24 cots were occupied by infants who had breathing tubes - that gives 500 ventilator shifts, which would give the median rate of 4% of ventilated shifts.
A further obviously misleading element is the implications that the infants were under Letby's care, but no nurse would care for several intensive care patients.
Methods for identfication of variations from the norm were first proposed a century ago [A Brief History of Statistical Process Control, 2021, D Seland, Quality magazine].
The 2011 Department of Health and Healthcare Quality Improvement Partnership guidance on "Detection and management of outliers" states "The statistical analyses involved should be carried out by people with appropriate statistical expertise and experience." The journalists broadcast seriously erroneous information.
I request that you publish a formal apology and correction.
You could ask Tim Harford of BBC Radio 4's More or less to assist you in finding an experienced and qualified statistician. I know that the Crown Prosecution Service did not wish to pursue inquiries into the validity of the statistical evidence.”
(Via Dr Philip Hammond on X)