What this is
The Independent Office for Police Conduct’s annual statistics on deaths during or following police contact, twenty-first edition, published on 31 July 2025 and covering 1 April 2024 to 31 March 2025 in England and Wales.
The restraint figures sit in chapter 5 of the report. It is the IOPC’s own count, published each year on a consistent basis, and it is the closest thing England and Wales has to a national series on this question.
What it records
Seventeen deaths in or following police custody in the year.
Five of the 17 had some force used against them by officers or by members of the public before their death, and all five were physically restrained by police (IOPC, 31 July 2025, chapter 5). One of those deaths also involved force by security officers. Three of the five were White and two were Black. Three of the five incidents also involved leg restraints. Fifteen of the 17 deaths were under independent IOPC investigation.
The IOPC’s stated position is that restraint or other force did not necessarily contribute to the deaths.
How to read it
Five of 17 is not a rate and not a trend. The same measure gave 14 of 24 for 2023/24, and 10 for 2025/26. A series that runs 14, then five, then ten is a sub-characteristic of a small annual cohort, and quoting any single year of it, high or low, misrepresents how the figure behaves.
The denominator is not fixed either. This report’s trend table gives 17 for 2024/25 and marks 2023/24 as 25, revised, while the 2023/24 report stated 24. The restrained count is unaffected by that; the base it is divided into is not.
The reading rule that travels with every edition is about cause. Five deaths in which restraint occurred is not five deaths caused by restraint, and the IOPC says so itself. Each custody death has its own investigation, and where relevant its own inquest. That is where causation gets decided, by someone with the evidence, and it is not decided by this table.
Where it stops
One year, England and Wales, deaths in or following police custody. Deaths after police contact where the person was not in custody are counted in a separate IOPC series, and the two must not be added together.
The report gives no breakdown of hold, duration or position for the five, and no count of how many restraints police carried out in the year. No rate can be calculated from it, and this page calculates none.
The series is the IOPC’s own, and no second body publishes an equivalent count. That is why the revision of last year’s total matters in practice: two briefings written a year apart may quote two different numbers for the same year, and the later one may not be the one on the slide.
Sources
Where a death follows restraint in custody, our physical interventions training covers positional risk and what has to happen in the minutes after a hold ends.
Reviewed 20 October 2026.