What this is
The Care Quality Commission’s statutory report on the Mental Health Act for 2024 to 2025, published on 29 January 2026. Section 120 of the Mental Health Act 1983 requires the regulator to monitor how the Act is used and to report annually. Part of that duty is recording the deaths of detained patients that providers notify to the CQC.
It is one notification return, from one publisher, covering England. This page covers the seclusion figure within it; the restraint figure has its own entry.
What it records
The return covers 314 notified deaths for 2024/25: 253 detained patients and 61 patients on community treatment orders.
Seven detained patients died during or within seven days of seclusion or time out, in the report’s own framing.
In the same return, fourteen detained patients died “where it was recorded that the patient died within 7 days of restraint”.
Forty-one of the 314 deaths were recorded as unnatural in the same table.
The report does not disaggregate the seclusion deaths by cause beyond the overall natural and unnatural split, and it records no finding of causation for any of the seven.
How to read it
Both figures are proximity measures. A notification records that a death followed a restrictive practice. It does not record the observations carried out, the length of time in the room, whether a doctor attended, or what the post-incident review found. Seven is a count of deaths close in time to seclusion, and the report offers no denominator: nothing records how many seclusions took place in the year.
The pairing is what earns the page its place. Seclusion is often presented in training as the calmer alternative to a physical hold, the option a ward reaches for when it wants to reduce hands-on risk. The return does not support that framing: seven deaths close to seclusion against fourteen close to restraint, from the same 314 notifications, in the same twelve months.
Present them together. A trainer who quotes the restraint figure and leaves seclusion out is describing half the exposure, and is teaching seclusion as the safer option without evidence for it.
Where it stops
One year, one nation, one recording route. The count covers detained patients. Deaths of informal patients, and of people restrained or secluded in other settings, are not in it.
Provider notification practice varies between trusts, and the report does not claim completeness. Deaths of patients on community treatment orders are not compulsorily reportable, so that element may be an undercount, as the CQC states.
The figure rests on a single publisher’s published return and comes from the same return as the restraint count. It is a separate measure, not an independent source, and no second body corroborates it.
No causal claim is made, and this page makes none.
Sources
If you are reviewing how often restraint and seclusion are used in your service, and what happens after them, our restraint reduction training covers the recording, the review and the debrief.
Reviewed 20 October 2026.