What this is
A joint investigation by BBC Radio 4’s File on 4 and the Bureau of Investigative Journalism, reported in January 2012, into deaths following police restraint in England and Wales. It is not a state publication but a dataset built by journalists: Freedom of Information requests asked the Independent Police Complaints Commission for the names behind its published statistics, then cross-referenced them against cases already in the public domain.
The period covered is 1998/99 to 2008/09.
What it records
Eighty-six people died in police custody in England and Wales in those eleven years after restraint was used. Of those, 16 deaths were categorised by the IPCC as directly “restraint-related” (File on 4 and TBIJ, January 2012).
The gap between the two numbers is the finding. The IPCC counted a death as restraint-related only where it met its own definition, and it defended that definition because a tight definition is what allows trends to be tracked over time. Everything else in the 86 is a death in custody in which restraint featured, without a formal finding that restraint caused it.
The investigation’s conclusion was that official figures understated the number of people who die after police restraint, because anyone who dies following restraint without being formally arrested falls outside the deaths in custody figures. The IPCC accepted that it held a separate list of deaths following police contact where the person had not been arrested or detained. It said it did not know how many of those were restraint-related, and that it was considering a new study. The chair of the Commons Home Affairs Committee called for a proper inquiry.
The IPCC then asked the Government Statistical Service to review its figures. That review found the cases correctly classified and the criticism unsupported, and recommended more be done to help users understand them.
How to read it
Sixteen is a count of cases meeting a definition. It is not a count of people who died after restraint, and the investigation exists because the two were being treated as the same thing.
The 16 also left out one of the country’s most high-profile restraint deaths: a man who died in 1999 after being restrained by eight police officers, and no officer was charged. If a definition produces a total that omits the case everyone remembers, the total measures the definition.
Treat the 86 as an investigative estimate built from case files, not a national statistic, and never add it to a count from another publisher.
Where it stops
Eleven years, one setting, England and Wales, and a dataset assembled by hand from FOI’d names. Nothing in it covers restraint in care homes, hospitals, schools or prisons.
Both published items reporting it are BBC coverage of the same investigation, so the two links below are one source and not two. The IPCC’s underlying case data was never published and could not be opened. The figures are as reported.
It is a document of 2012, describing what official counting looked like then. The IOPC now publishes restraint-involved custody deaths on a different basis.
Sources
- BBC News, restraint deaths review after BBC investigation, 31 January 2012
- BBC News, deaths in custody ‘understated’ (File on 4), undated
If you are briefing staff on why a death after restraint may not appear in any national figure, our vulnerable and volatile persons training covers the judgement calls before hands-on restraint, and the record that has to follow.
Reviewed 20 October 2026.