Letby shift data was scientifically worthless, statisticians warn

Shift pattern data used in the Lucy Letby trial was “scientifically worthless” a meeting of the Royal Statistical Society has heard.

One of the most damning pieces of evidence presented to the jury at Letby’s trial was a chart indicating that the nurse was always on duty at the Countess of Chester Hospital when babies collapsed or died.

The table, which covered a 13-month period between June 2015 and June 2016, showed that while the 38 other nurses were in attendance on just a handful of occasions when suspicious incidents occurred, Letby was at the scene of all of them.

However, statisticians were so concerned about the chart that they convened a meeting at the RSS headquarters in London on Thursday to discuss how statistics are used in court cases.

Dealing specifically with the Letby data, the meeting heard how the rota pattern was a “scientific fake” which could not be reproduced independently.

Statisticians said there were “issues with the selection and collection of data” used in the chart and warned there may be “many other possible causal factors not being considered that might explain these deaths”.

‘Bigger impact than it deserves’

The duty roster evidence was described as “feeble” and “had a way bigger impact than it deserves” with experts arguing that not enough attention had been paid to alternative explanations, such as the different shift patterns of the nurses and whether some were more likely to be on duty than others.

The meeting, which was held in private under Chatham House rules which allow attendees to remain anonymous, heard that statistical evidence was “worthless” and there had been no concrete evidence of a spike in deaths at the hospital.

In August 2023, Letby was convicted of the murders of seven newborns and the attempted murders of six other infants. A retrial in July 2024 also found her guilty of the attempted murder of another child.

Since the conviction, numerous scientists, statisticians and doctors have expressed their concern about the evidence presented to the jury regarding shift patterns, medical conclusions and the standard of care at the Countess of Chester.

The police and prosecution have defended the chart, saying it was just one brick in a wall of evidence which showed that Letby was undoubtedly guilty.

Speaking about the evidence after the meeting, Dr Jane Hutton, a professor of statistics at the University of Warwick, said: “It’s a large pile of crockery, much of which is broken. Such a pile does not hold water however big it is.

“We’re not saying the conviction is unsafe, but we consider that if the concerns we are raising are essential to the decisions of the court, then the convictions are unsafe.”

Nurses were later exonerated

The meeting heard how similar shift pattern data was used to convict a Dutch paediatric nurse and Italian nurse Daniela Poggiali, both of whom were later exonerated.

Statisticians warned that statistical evidence was being “abused by the courts in the way that it’s used in trials”.

Members of the RSS said they were asking for parity in court cases, with statisticians invited to be present if statistical evidence was shown to the jury.

“If toxicology was represented you would expect a toxicologist to be present,” they said.

The meeting heard how shift pattern charts had the potential to spark confirmation bias which could taint further evidence.

Statisticians also said that there was no evidence of a spike in deaths at the Countess of Chester once the hospital was seen against other similar units and warned that clusters do happen in medical settings.

“There is evidence of coming last,” said one expert. “But coming last is not the scene of crime.”

In 2022 the RSS produced a report entitled “Health care serial killer or coincidence?” to help legal teams present data correctly.

The report warned that investigators must always bear in mind that there may be innocent explanations for apparent and even striking correlations between a medical professional’s presence and deaths or other incidents.

It recommended that investigators and prosecutors consult professional independent statisticians who could give instructions to the jury on how to interpret the data.

It also advised that confounding factors that could distort the figures should be made clear. But none of the recommendations were followed in the Letby trial.

The meeting concluded by calling for greater collaboration between statisticians and the legal profession in the future.