A service evaluation of restraint, seclusion and escalation in inpatient mental health care: a retrospective observational study in England
This retrospective observational study of over 41,000 incidents in English inpatient mental health services reveals that restrictive practices operate as a two-stage system where service configuration determines restraint frequency and ward-level context influences escalation to seclusion, suggesting that reduction strategies must target both incident rates and the processes governing escalation.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Imagine the mental health hospital as a giant, bustling school with different wings: some are for students who need extra security (the "forensic" or secure units), and others are for students in acute crisis (the "non-forensic" or regular wards). For a whole year, from June 2024 to May 2025, researchers acted like detectives, looking at 41,685 "incident reports" from one massive NHS Trust in England. They weren't just counting how many times students got into trouble; they were trying to solve a mystery: Why do some trouble spots turn into a simple timeout (restraint), while others spiral into a locked room (seclusion)?
Here is what the clues revealed.
The Two-Stage Game: Where and How
The researchers found that restrictive practices (like physical holding or locking someone away) work like a two-stage video game.
Stage 1: The Map (Where the trouble starts)
First, the "map" of the hospital decides where the trouble is most likely to happen. It's not just about how "dangerous" a patient is; it's about the specific room they are in.
- The "Low Secure" Wing: Surprisingly, this area had the highest rate of physical holding, at 12.86%.
- The "Medium Secure" Wing: Close behind at 11.81%.
- The "High Secure" Wing: Even here, where you'd expect the most security, the rate was 11.45%.
- The "Older Adult" Wing: In the non-secure section, the older adult wards had the highest rate at 12.93%.
The paper suggests that the type of ward and the security level act like a filter, deciding where the initial "holding" happens. It's not a straight line where "more security = more holding." In fact, the "Low Secure" units had more holding than the "High Secure" ones.
Stage 2: The Spin-Off (Does it get worse?)
This is the big discovery. Once a patient is physically held (restrained), does it automatically turn into a locked-room seclusion? No. The paper argues against the idea that this is an inevitable chain reaction. Instead, it depends on the "ward context"—basically, the local culture and how the staff handles the situation right after the hold.
- The "Spin-Off" Champions: In the Medium Secure Units, 6.22% of restraint incidents escalated to seclusion. In regular Acute Wards, 4.72% escalated.
- The "Spin-Off" Stoppers: In High Secure Services, only 0.26% of restraints turned into seclusion. In Older Adult wards, the escalation was minimal.
The authors suggest that the environment acts like a shock absorber. In some wards, the staff or the routine successfully stops the situation from getting worse. In others, it keeps spinning out of control.
The "What If" Simulation
The researchers didn't just look at what happened; they ran a computer simulation (a "counterfactual" analysis) to see what might happen if things changed.
- The Finding: In high-risk settings, the simulation suggested that if you could stop the escalation process (the jump from holding to locking up), you would reduce seclusion rates more effectively than just trying to reduce the number of holding incidents alone.
- The Caveat: This is a simulation, not a proven fact. The paper says this suggests a new way to think about fixing the problem, but it hasn't been tested in the real world yet.
What the Paper Rules Out
The paper explicitly argues against the idea that seclusion is just a natural, unavoidable result of a patient being "high risk" or that it happens the same way in every hospital.
- It rules out the idea that security level is the only thing that matters. If it were, High Secure wards would have the most seclusion, but they actually have the least escalation.
- It rules out the idea that restraint always leads to seclusion. The data shows that in many places, a restraint happens, and then... nothing else happens. The situation resolves.
The Numbers Game
Let's look at the specific numbers the paper gives us, because they tell a story of their own:
- Total Incidents: 41,685.
- Total Restraints: About 3,390 incidents (which is 8.13% of all incidents).
- Total Seclusions: Only 165 incidents (0.40% of all incidents).
- The Link: If a restraint happened, the odds of it turning into seclusion were 5.72 times higher than if no restraint happened. But remember, that link is weak in some wards and strong in others.
- Ward Differences: In non-secure wards, female wards had a higher rate of restraint (9.47%) compared to male wards (6.73%).
The Takeaway: It's About the "After"
The paper concludes that fixing this problem isn't just about trying to stop the first "hold" (restraint). That's important, but it's only half the battle. The other half is fixing the "aftermath."
Imagine a car crash. You can try to prevent the crash (reduce restraint), but you also need to make sure the airbags work and the emergency team doesn't make the injury worse (prevent escalation to seclusion). The authors suggest that we need to look at the "ward-level processes"—the specific habits, staff responses, and routines that decide whether a scary moment stays scary or gets even worse.
They propose that the most dangerous places might not be the ones with the most restraints, but the ones where restraints are most likely to turn into seclusion. By studying the wards that stop the escalation (like the High Secure and Older Adult wards), we might learn how to build better "shock absorbers" for everyone else.
But remember, the paper is careful to say this is a service evaluation based on past data. It suggests these patterns and simulates potential improvements, but it doesn't claim to have solved the problem or proved that changing one thing will fix everything. It's a map for where to look next, not the destination itself.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.