The officer asked again why she had written, “I killed them on purpose.”

“That’s how I was being made to feel,” she said. As her mental health deteriorated, her thoughts had spiralled. “If my practice hadn’t been good enough and I was linked with these deaths, then it was my fault,” she said.

“You’re being very hard on yourself there if you haven’t done anything wrong.”

“Well, I am very hard on myself,” she said.

After more than nine hours of interviews, Letby was released on bail, without being charged. She moved back to Hereford, to live with her parents. News of her arrest was published in papers throughout the U.K. “All I can say is my experience is that she was a great nurse,” a mother whose baby was treated at the Countess told the Times of London. Another mother told the Guardian that Letby had advocated for her and had told her “every step of the way what was happening.” She said, “I can’t say anything negative about her.” The Guardian also interviewed a mother who described the experience of giving birth at the Countess. “They had no staff and the care was just terrible,” she said. She’d developed “an infection which was due to negligence by a member of staff,” she explained. “We made a complaint at the time but it was brushed under the carpet.”

One of Letby’s childhood friends, who did not want me to use her name because her loyalty to Letby has already caused her social and professional problems, told me that she asked the Cheshire police if she could serve as a character reference for Letby. “They weren’t interested at all,” she said. Letby seemed to be in a state of “terror and complete confusion,” the friend said. “I could tell from how she was acting that she just didn’t know what to say about it, because it was such an alien concept to be accused of these things.”

Shortly after Letby’s arrest, the pediatric consultants arranged a meeting for the hospital’s medical staff, to broach the possibility of a vote of no confidence in Chambers, the hospital’s chief executive, because of the way he’d handled their concerns. Chambers resigned before the meeting. A doctor named Susan Gilby, who took the side of the consultants, assumed his role. Gilby told me that the first time she met with Jayaram it was clear that he was suffering from the experience of not being believed by the hospital’s management. “He was in tears, and bear in mind this is a mature, experienced clinician,” she said. “He described having issues with sleeping, and he felt he couldn’t trust anyone. It was really distressing.” She was surprised that Ian Harvey, the hospital’s medical director, still doubted the consultants’ theory of how the babies had died. Harvey seemed more troubled by their behavior, she said, than by anything Letby had done. “In his mind, the issue seemed to be that they weren’t as good as they thought they were,” Gilby told me. “It was ‘They think they’re marvellous, but they need to look at themselves.’ ” (Harvey would not comment, citing the court order.)

The week of Letby’s arrest, the police dug up her back garden and examined drains and vents, presumably to see if she had hidden anything incriminating. Four months later, while she remained out on bail without charges, the Chester Standard wrote, “The situation has caused many people to question both the ethics and legality of keeping someone linked to such serious allegations when seemingly there is not enough evidence to bring charges.” Letby was arrested a second time, in 2019, but, after being interviewed for another nine hours, she was released.

In November, 2020, more than two years after Letby’s first arrest, an officer called Gilby to inform her that Letby was being charged with eight counts of murder and ten counts of attempted murder. (Later, one of the murder counts was dropped, and five attempted-murder charges were added.) She was arrested again, and this time she was denied bail. She would await trial in prison. As a courtesy, Gilby called Chambers to let him know. She was taken aback when Chambers expressed concern for Letby. She said that he told her, “I’m just worried about a wrongful conviction.”

In September, 2022, a month before Letby’s trial began, the Royal Statistical Society published a report titled “Healthcare Serial Killer or Coincidence?” The report had been prompted in part by concerns about two recent cases, one in Italy and one in the Netherlands, in which nurses had been wrongly convicted of murder largely because of a striking association between their shift patterns and the deaths on their wards. The society sent the report to both the Letby prosecution and the defense team. It detailed the dangers of drawing causal conclusions from improbable clusters of events. In the trial of the Dutch nurse, Lucia de Berk, a criminologist had calculated that there was a one-in-three-hundred-and-forty-two-million chance that the deaths were coincidental. But his methodology was faulty; when statisticians looked at the data, they found that the chances were closer to one in fifty. According to Ton Derksen, a Dutch philosopher of science who wrote a book about the case, the belief that “such a coincidence cannot be a coincidence” became the driving force in the process of collecting evidence against de Berk. She was exonerated in 2010, and her case is now considered one of the worst miscarriages of justice in Dutch history. The Italian nurse, Daniela Poggiali, was exonerated in 2021, after statisticians reanalyzed her hospital’s mortality data and discovered several confounding factors that had been overlooked.

William C. Thompson, one of the authors of the Royal Statistical Society report and an emeritus professor of criminology, law, and psychology at the University of California, Irvine, told me that medical-murder cases are particularly prone to errors in statistical reasoning, because they “involve a choice between alternative theories, both of which are rather extraordinary.” He said, “One theory is that there was an unlikely coincidence. And the other theory is that someone like Lucy Letby, who was previously a fine and upstanding member of the community, suddenly decides she’s going to start killing people.”

Flawed statistical reasoning was at the heart of one of the most notorious wrongful convictions in the U.K.: a lawyer named Sally Clark was found guilty of murder, in 1999, after her two sons, both babies, died suddenly and without clear explanation. One of the prosecution’s main experts, a pediatrician, argued that the chances of two sudden infant deaths in one family were one in seventy-three million. But his calculations were misleading: he’d treated the two deaths as independent events, ignoring the possibility that the same genetic or environmental factors had affected both boys.

In his book “Thinking, Fast and Slow” (2011), Daniel Kahneman, a winner of the Nobel Prize in Economics, argues that people do not have good intuitions when it comes to basic principles of statistics: “We easily think associatively, we think metaphorically, we think causally, but statistics requires thinking about many things at once,” a task that is not spontaneous or innate. We tend to assume that irregular things happen because someone intentionally caused them. “Our predilection for causal thinking exposes us to serious mistakes in evaluating the randomness of truly random events,” he writes.

Burkhard Schafer, a law professor at the University of Edinburgh who studies the intersection of law and science, said that it appeared as if the Letby prosecution had “learned the wrong lessons from previous miscarriages of justice.” Instead of making sure that its statistical figures were accurate, the prosecution seems to have ignored statistics. “Looking for a responsible human—this is what the police are good at,” Schafer told me. “What is not in the police’s remit is finding a systemic problem in an organization like the National Health Service, after decades of underfunding, where you have overworked people cutting little corners with very vulnerable babies who are already in a risk category. It is much more satisfying to say there was a bad person, there was a criminal, than to deal with the outcome of government policy.”

Schafer said that he became concerned about the case when he saw the diagram of suspicious events with the line of X’s under Letby’s name. He thought that it should have spanned a longer period of time and included all the deaths on the unit, not just the ones in the indictment. The diagram appeared to be a product of the “Texas sharpshooter fallacy,” a common mistake in statistical reasoning which occurs when researchers have access to a large amount of data but focus on a smaller subset that fits a hypothesis. The term comes from the fable of a marksman who fires a gun multiple times at the side of a barn. Then he draws a bull’s-eye around the cluster where the most bullets landed.

For one baby, the diagram showed Letby working a night shift, but this was an error: she was working day shifts at the time, so there should not have been an X by her name. At trial, the prosecution argued that, though the baby had deteriorated overnight, the suspicious episode actually began three minutes after Letby arrived for her day shift. Nonetheless, the inaccurate diagram continued to be published, even by the Cheshire police.

Dewi Evans, the retired pediatrician, told me that he had picked which medical episodes rose to the level of “suspicious events.” When I asked what his criteria were, he said, “Unexpected, precipitous, anything that is out of the usual—something with which you are not familiar.” For one baby, the distinction between suspicious and not suspicious largely came down to how to define projectile vomiting.

Letby’s defense team said that it had found at least two other incidents that seemed to meet the same criteria of suspiciousness as the twenty-four on the diagram. But they happened when Letby wasn’t on duty. Evans identified events that may have been left out, too. He told me that, after Letby’s first arrest, he was given another batch of medical records to review, and that he had notified the police of twenty-five more cases that he thought the police should investigate. He didn’t know if Letby was present for them, and they didn’t end up being on the diagram, either. If some of these twenty-seven cases had been represented, the row of X’s under Letby’s name might have been much less compelling. (The Cheshire police and the prosecution did not respond to a request for comment, citing the court order.)

Among the new suspicious episodes that Evans said he flagged was another insulin case. Evans said that it had similar features as the first two: high insulin, low C-peptide. He concluded that it was a clear case of poisoning. When I asked Michael Hall, a retired neonatologist at University Hospital Southampton who worked as an expert for Letby’s defense, about Evans’s third insulin case, he was surprised and disturbed to learn of it. He could imagine a few reasons that it might not have been part of the trial. One is that Letby wasn’t working at the time. Another is that there was an alternative explanation for the test results—but then, presumably, such an explanation could be relevant for the other two insulin cases, too. “Whichever way you look at this, that third case is of interest,” Hall told me.

Ton Derksen, in his book about Lucia de Berk, used the analogy of a train. The “locomotives” were two cases in which there had been allegations of poisoning. Another eight cases, involving children who suddenly became ill on de Berk’s shifts, were the “wagons,” trailing along because of a belief that all the deaths couldn’t have occurred by chance.

The locomotives in the Letby prosecution were the insulin cases, which were charged as attempted murders. “The fact that there were two deliberate poisonings with insulin,” Nick Johnson, the prosecutor, said, “will help you when you are assessing whether the collapses and deaths of other children on the neonatal unit were because somebody was sabotaging them or whether these were just tragic coincidences.”

But not only were the circumstances of the poisonings speculative, the results were, too. If the aim was to kill, neither child came close to the intended consequences. The first baby recovered after a day. The second showed no symptoms and was discharged in good health.

On the first day of the trial, Letby’s barrister, Benjamin Myers, told the judge that Letby was “incoherent, she can’t speak properly.” She had been diagnosed as having post-traumatic stress disorder following her arrests. After two years in prison, she had recently been moved to a new facility, but she hadn’t brought her medication with her. Any psychological stability she’d achieved, Myers said, had been “blown away.”

Letby, who now startled easily, was assessed by psychiatrists, and it was decided that she did not have to walk from the dock to the witness box and instead could be seated there before people came into the room. The Guardian said that in court Letby “cut an almost pitiable figure,” her eyes darting “nervously towards any unexpected noise—a cough, a dropped pen, or when the female prison guard beside her shuffled in her seat.” Her parents attended the entire trial, sometimes accompanied by a close friend of Letby’s, a nurse from the unit who had recently retired.

Press coverage of the case repeatedly emphasized Letby’s note in which she’d written that she was “evil” and “killed them on purpose.” Media outlets magnified the images of those words without including her explanations to the police. Much was also made of a text that she’d sent about returning to work after her trip to Spain—“probably be back in with a bang lol”—and the fact that she’d searched on Facebook thirty-one times for parents whose children she was later accused of harming. During the year of the deaths, she had also searched for other people 2,287 times—colleagues, dancers in her salsa classes, people she had randomly encountered. “I was always on my phone,” she later testified, explaining that she did the searches rapidly, out of “general curiosity and they’ve been on my mind.” (Myers noted that her search history did not involve any references to “air embolism.”)

The parents of the babies had been living in limbo for almost a decade. In court, they recalled how their grief had intensified when they were told that their children’s deaths may have been deliberately caused by someone they’d trusted. “That’s what confuses me the most,” one mother said. “Lucy presented herself as kind, caring, and soft-spoken.” They had stopped believing their own instincts. They described being consumed by guilt for not protecting their children.

Several months into the trial, Myers asked Judge Goss to strike evidence given by Evans and to stop him from returning to the witness box, but the request was denied. Myers had learned that a month before, in a different case, a judge on the Court of Appeal had described a medical report written by Evans as “worthless.” “No court would have accepted a report of this quality,” the judge had concluded. “The report has the hallmarks of an exercise in working out an explanation” and “ends with tendentious and partisan expressions of opinion that are outside Dr. Evans’ professional competence.” The judge also wrote that Evans “either knows what his professional colleagues have concluded and disregards it or he has not taken steps to inform himself of their views. Either approach amounts to a breach of proper professional conduct.” (Evans said that he disagreed with the judgment.)

Evans had laid the medical foundation for the prosecution’s case against Letby, submitting some eighty reports. There was a second pediatric expert, who provided what was called “peer review” for Evans, as well as experts in hematology, endocrinology, radiology, and pathology, and they had all been sent Evans’s statements when they were invited to participate in the case. The six main prosecution experts, along with at least two defense experts who were also consulted, had all worked for the N.H.S. Evans wasn’t aware if Letby’s lawyers had sought opinions from outside the U.K., but he told me that, if he were them, he would have looked to North America or Australia. When I asked why, he said, “Because I would want them to look at it from a totally nonpartisan point of view.”