Publication — not a case

The count of 1,983 deaths that cannot show restraint

According to INQUEST, 1,983 people have died in and following police contact in England and Wales since 1990, and restraint is not one of the things its table counts.

Police custody · 2020s

What this is

A statistics page maintained by INQUEST, the charity that supports bereaved families after state-related deaths, updated on 14 October 2026. It is a cumulative count of deaths in and following police contact in England and Wales since 1990, broken down by year and by force, and it draws on Independent Office for Police Conduct figures and on INQUEST’s own casework.

It is not a National Statistic and not a regulator’s return. INQUEST is a campaign and casework organisation with no enforcement power.

What it records

The table records 1,983 deaths since 1990. According to INQUEST, 1,242 of them were in custody, 89 were fatal police shootings, 484 were pursuit deaths and 168 were road traffic incident deaths. The Metropolitan Police accounts for 434 of the total and other forces for 1,549.

There is no restraint category in the table, no sub-count of deaths in which restraint was used, and no published restraint definition.

Separately, in its response to the 2025/26 figures published on 30 July 2026, INQUEST states that 10 of the 12 people subject to use of force in police custody were physically restrained, and that five of the 12 were subject to leg restraints.

How to read it

The absence is the page. The organisation most associated with counting deaths in custody, doing it for more than three decades, cannot supply a restraint figure, because its four categories record where a death happened and not how the person died. A death in custody is counted once, under its category, and restraint is invisible inside all four.

So do not attribute a restraint death count to INQUEST. Anyone quoting its total as a restraint figure is quoting a number that does not contain one.

The 10-of-12 statement is a different kind of claim from the table. It is a campaign organisation’s rendering of IOPC data for one period, and according to INQUEST it is neither a rate nor a national estimate, nor is it derived from the series.

Where it stops

England and Wales, from 1990 to October 2026, one organisation’s compiled series. It carries no finding about any death, and no legal or regulatory outcome attaches to it.

Where INQUEST’s rendering of IOPC figures differs from the IOPC’s own edition, the IOPC is primary. This page cites INQUEST only for what INQUEST publishes.

The table is a running total, so figures move as deaths are added and reclassified. A number quoted from this page is a number as at 14 October 2026, and should be dated when it is used.

One row of the published table is incomplete, so the shootings figure is carried as INQUEST states it.

Sources

Where a restraint is being reviewed after the event, our physical interventions training covers positional risk and the monitoring that has to continue after the hold ends.

Reviewed 20 October 2026.

All police custody cases  ·  Restraint Casebook

Book in a 1-2-1 Meeting

Find Your Solution

What do you need? Use our simple tool to find the most relevant training and support options for your role, sector and challenges.