What happened
David Bennett was 38 and had a long history of chronic schizophrenia, receiving care in mental health settings since the early 1980s. From the autumn of 1995 he was detained under the civil powers of the Mental Health Act 1983 at the Norvic Clinic in Norwich, a medium secure unit serving Norfolk, Suffolk and Cambridgeshire. His illness was refractory, and he was on atypical antipsychotic medication.
On the evening of Friday 30 October 1998, after an altercation with another detained patient in which blows were exchanged, he was transferred from one ward to another. There he assaulted a nurse, and the nursing team restrained him immediately. He was held on the floor in the prone position for at least 25 minutes. Towards the end of that period staff became aware that he had collapsed and was unconscious. Resuscitation failed, and he was pronounced dead shortly after midnight on 31 October 1998.
The outcome
The inquest was heard from 3 to 17 May 2001. Two forensic pathologists who had each conducted a post-mortem agreed the cause: myocardial dysrhythmia secondary to prolonged restraint in the prone position, and to the long-term administration of antipsychotic medication. The inquest was highly critical of the restraint.
An independent inquiry reported in 2004 with 22 recommendations, most on race, ethnicity and staff training, four on manual restraint, resuscitation and the response to a collapsed patient.
What went wrong
Duration. A man was held prone for at least 25 minutes, and the collapse was noticed only towards the end of it. The inquiry recommended that prone restraint be limited to three minutes. The trust’s policies, revised from May 2001, required prone restraint to be timed, recorded and the position changed as soon as practicable, but had not adopted the three-minute limit.
The panel also noted the absence of a resident doctor on call, saying a doctor should be immediately present, or attend within 20 minutes, wherever a patient is detained. Resuscitation training sat apart from restraint training until May 2001.
What it teaches
Put a clock on it. A prone hold is a timed intervention with an end point, and the person applying it is also the person who has to notice that the struggling has stopped. A policy that says “as soon as practicable” leaves the judgement to whoever is holding, in the middle of the incident. The team that holds someone is also the team that has to start work on them if they stop breathing, and until May 2001 those skills were taught apart in this trust.
Questions for the room
- The struggle stopped. What did that change about the risk, and who in that team was watching for it?
- You are leading a team and the patient is still resisting after three minutes. What is your next move, and what do you say out loud?
- The inquiry asked for a doctor within 20 minutes. What would that have changed here, and who commissions that cover?
Evidence and limits
Confidence: probable. The inquiry report and the inquest dates come from a published primary source, but the verdict itself was not checked against the inquisition, so this page states no verdict. Contemporaneous coverage reported accidental death.
The report is dated December 2003 in the record’s note and cited as 2004 in the journal article, so both years appear above. The pathologists also recorded an alternative mechanism, potassium release when pressure on a limb is released, which they did not resolve.
No member of staff is named, and no prosecution is recorded in the sources held.
Sources
- The Psychiatrist (H. N. Ball), 2005 · `01_source_text/pos-asphyxia-local/Death-in-restraint-lessons.txt`
- Police Complaints Authority, Safer Restraint conference report, April 2002 · `01_source_text/library-sudden-death-risk/PCA Safer Restraint Report.md`
Our restraint reduction training starts with duration: how long a hold may run, who times it, and who has the authority to end it.
Reviewed 20 October 2026.