What happened
James Reilly, 60, was serving 15 years for robbery at HMP Stoke Heath. He had type 2 diabetes and cirrhosis of the liver, had been assessed as frail, and used a wheelchair. There was no care plan for the fluid retention and cellulitis his liver disease caused.
At 12.35pm on 25 September 2022 he pressed his cell bell. Officers found blood on the floor and called a code red, the alarm for severe blood loss. He was dizzy, with low blood pressure and a high pulse, and had vomited a large amount of blood. An ambulance was requested two minutes later but not prioritised; paramedics arrived at 2.00pm.
He was restrained with an escort chain for the journey, recorded as a low escape risk and a high risk to the public. The chain stayed on in the hospital, through an anaesthetic, and through an endoscopy.
After the procedure he deteriorated and haemorrhaged. The chain came off only when hospital staff said the situation was life-threatening. He was placed in a coma and died the following day, 26 September 2022, of a variceal bleed.
The outcome
The Prisons and Probation Ombudsman reported in May 2023. It found the escort chain undignified and unacceptable for a man with mobility problems who used a wheelchair and was vomiting blood, and found no care plan for his condition. It found there had been opportunity to involve healthcare staff in the risk assessment before the escort, and that the ambulance took over an hour.
It recommended care plans for prisoners with long-term conditions, and escort assessments that reflect the law on restraints and the risk the prisoner actually presents. The Prison Service said both were done.
The inquest concluded natural causes. No prosecution is recorded.
What went wrong
The restraint outlived every reason for it. A man with a diagnosed bleed, a wheelchair and low blood pressure was chained to an escort, and nobody revisited the chain when he was anaesthetised, during the procedure, or when he came round worse.
The trigger for removing it was not a clinical judgement about his condition. It was a hospital statement that he was dying.
There is a second failure, earlier: the assessment recorded a high risk to staff and a low escape risk, which is a judgement about danger rather than flight. Nobody joined that up with the fact that the man being escorted could not stand.
What it teaches
Escort restraint is a clinical decision as much as a security one. The test is continuing necessity, reviewed at the hospital door, in the anaesthetic room and on the ward, not the assessment written at the gate.
A person too ill to walk is not a flight risk, whatever the form says. And a chain that stays on through an endoscopy is not neutral: it tells the patient and the clinical staff what the service expects them to work around.
Questions for the room
- He was vomiting blood in a wheelchair. What did the escort risk assessment need to say that it did not?
- You are the escort officer in the endoscopy suite. Who do you ask, and what do you need to hear, before the chain comes off?
Evidence and limits
Confidence: confirmed. The Ombudsman’s report and a BBC News report were both read and agree on sequence, timings, recommendations and the inquest conclusion.
The record carries no date for the inquest conclusion, so none is given here. That the decision was lawful but inhumane is the Ombudsman’s wording, not a court’s.
He died of a variceal bleed, not of the restraint. The case is here because of the Ombudsman’s finding about the chain, not because the chain killed him.
Our restraint reduction training covers positional risk, duration limits and the monitoring that has to go with any hold.
Sources
Reviewed 20 October 2026.