In Hyde, near Manchester, Harold Shipman practised alone at 21 Market Street. Directly opposite, at number 20, was the Brooke Practice. Before a body could be cremated in England, a second doctor had to examine it and sign a confirming form, Form C. Shipman almost always asked the Brooke doctors to sign his.
One partner, who had joined in 1996, noticed that she was asked to sign rather more often than she was used to. Reading his forms, she also noticed that Shipman had often been present at the death. In 19 years of practice, that had happened to her only two or three times. Her partners reassured her. Shipman had a lot of elderly patients, they said, and was known for visiting them unannounced. Then a local funeral director raised her own worry: elderly women, found dead at home, sitting up in a chair and fully dressed, who had not seemed to be ill. When the partners met to talk it over, some still thought a large list with many elderly patients could be expected to produce more deaths.
On Monday, 23 March 1998, Shipman asked one of the partners to sign another Form C. That day, the inquiry concluded, the practice probably pulled its figures. In the previous three months, its doctors had signed 16 cremation forms for Shipman’s patients, and that count left out anyone who died in hospital, was buried, or went to the coroner. The Brooke Practice itself, five doctors and about 9,500 patients, had recorded 14 or 15 deaths in the same period, counting those in hospital. They guessed Shipman had about 3,000 patients. The inquiry later found that his death rate appeared, at the very least, about three times theirs. Some partners thought it might be eight or nine times.
The next morning the partner who had first noticed telephoned the coroner with the numbers, who asked the police to look into it. The detective, the inquiry found, did not ask how the rates had been worked out and “attached no importance to the numbers or death rates.” On 17 April he closed the investigation, finding no substance in the concerns. In a report that August he described Shipman’s apparent excess of deaths as common gossip among doctors, nurses and undertakers, “although there was no evidence to support this rumour.”
Shipman killed three more patients, on 11 May, 12 June and 24 June. The last left a forged will naming him as her sole beneficiary, and he was arrested in September. In 2002 the inquiry reported that he had killed at least 215 patients. Its chair concluded that if the police and coroner had moved “with reasonable expedition,” the last three “would probably have been saved.”
Could a chart have caught him sooner? In work funded by the inquiry, researchers at Imperial College applied cumulative sum charts, a tool from statistical process control, to yearly deaths among the patients of GPs and practices in five English health authorities from 1993 to 1999. Of 1,009 GPs, 33, Shipman among them, crossed an alarm threshold designed to detect a rise of two standard deviations in standardised mortality. The authors wrote that poor data quality, and factors such as how many of a doctor’s patients were in nursing homes or hospices, were likely explanations for most of the other alarms. A 2008 study noted that Shipman had not been the most extreme outlier. Its own simulations concluded that, at best, routine monitoring could act as “a backstop.”