Key Takeaways
- De-escalation training, an insurance policy and a response system act at three different points in an incident. They can’t stand in for each other, and price is the wrong way to sort them.
- Each one has a single thing it depends on, and knowing what that is tells you exactly where that route leaves you.
- Nobody has measured whether any of the three lowers incidents at a business your size, including the one this company sells.
These three don’t compete. De-escalation training works on the minutes before a situation turns, an insurance policy works on the months after it’s over, and a response system is the only one that touches the minutes while it’s happening. All three get filed under workplace violence prevention methods, and only one of them is prevention.
Lined up as prices they look like three versions of the same purchase. They’re three answers to three different questions, and the useful one is which of those moments your business currently answers with nothing. The free decisions come first and this page assumes you’ve made them, so what follows is what you spend money on afterward.
Which workplace violence prevention methods are you actually choosing between?
Three things that act at three different points in an incident. They come back under one heading in every list, alongside cameras, better lighting and a policy document nobody reads. Sorting that list by price tells you which one is cheapest and nothing about which gap you’re filling.
Four questions do the sorting, and the same four apply to each option in the same order.
- When it acts. Before the situation turns, during it, or after it’s over.
- What it changes. The outcome it moves, as distinct from the outcome you’re hoping it moves.
- What has to be true for it to work. Each of the three rests on one condition, and that condition is where it fails you.
- What’s known about it at your size. Whether anybody has measured it in a business like yours.
The fourth question is the one most comparisons skip, and it’s the one that keeps this honest. Once you’ve asked it, the other three change job: you’re choosing on function, and function is knowable without a trial.
What does de-escalation training actually change?
It changes what a person does in the window before a situation turns physical. That window is real, and it’s narrower than the brochure suggests. When a 2025 study went back through hospital violence incidents as nursing staff had written them up, a verbal threat or insult came first in 30.8% of them. In the rest, nothing in the record announced it. The record is the thing to hold onto. That study analyzed what staff wrote down afterward rather than what happened, and every incident in it was in a hospital.
The strongest case anyone has published for de-escalation training comes from psychiatric wards. A cluster randomized trial reported in Frontiers in Psychiatry covered all six psychiatric hospitals in Slovenia. It gave staff on some wards 16 hours of training, left the others alone, and counted aggressive incidents over 6,401 treatment episodes. The trained wards ran at 27% of the rate on the untrained ones. That’s a big effect, measured on psychiatric wards with colleagues inside calling distance.
Widen the lens and what training reliably moves is something quieter. A 2025 meta-analysis in International Nursing Review, pooling four controlled studies of 821 healthcare professionals, found the clearest gain in how confident staff felt handling violence. Confidence is worth buying, and it’s also far easier to measure than an incident rate.
So the condition is presence. Training works where somebody is there, reading the room, with room to back away. One person behind a register at nine at night has been trained for a conversation they might not get to have.
What does insurance actually change?
It changes who writes the check, and when. Workers’ compensation covers medical costs and part of lost wages once a claim is filed, which is real money and the reason to carry it. It changes nothing about whether the incident happens, and less about the eventual cost than owners expect.
The part that surprises people is what a claim does to the price of the next few years. Premiums get adjusted by an experience modification, a factor built from your own claim history. NCCI writes the rating plan most states use. It weights how often claims happen more heavily than how badly they turn out, and losses count hardest below a per-claim threshold each state now sets for itself.
Read that back into a small business and it says something specific. Three modest claims move your price further than one severe one, and the event you insured against comes back as a bill spread thin across renewals.
The condition here is filing. A policy pays on the claims that get made, and it goes on pricing you whether or not you make them. What it can’t return is the shift that went uncovered or the person who decided not to come back.
If you’d like a second pair of eyes on which of those three moments your business leaves open, tell us how your shifts run.
What does a response system actually change?
It changes whether anybody knows while it’s happening. That’s the only one of the three moments where the situation is still live, and the other two options leave it alone entirely. Whether it changes anything beyond that depends on whether the person it’s for will carry it.
Staff opt out of systems they don’t trust, and they do it quietly. The badge goes in a drawer at the start of a shift, the dashboard still shows every device enrolled and reporting, and no alert anywhere says the person who needs it left it there. A device that also records where somebody went and how long they took stops being a safety device the first time it’s read as a performance file.
That’s why the design question runs backwards from the buying question. If the person carrying it doesn’t feel safer for carrying it, nothing else matters. Coverage, network, alerting, all of it sits downstream of that.
There’s also a version of this that makes an incident worse. A study in the Israel Journal of Health Policy Research ran interviews and focus groups at one hospital on how violent episodes unfold. The security personnel themselves said being called in can escalate an episode instead of settling it. They’d sooner be summoned with judgment than by default. Speed is one part of a response. What arrives, and how it behaves, is the rest.
So the condition is reach. A response system works where the person can get to it and will use it, and the second half is where most of them fail. The honest position on evidence is that nobody has measured whether one lowers incidents at a business of ten people. No trial, no cohort, no state dataset. The place to test the idea is the moment nobody else can see, and two of the three options have nothing to say about it.
How do the three compare, side by side?
They separate cleanly on the first three questions and agree completely on the fourth. Nothing here says one of them wins, because they aren’t answering each other.
| De-escalation training | Insurance | Response system | |
|---|---|---|---|
| When it acts | Before | After | During |
| What it changes | What a person does in the window before it turns | Who pays, and when | Whether anyone knows while it’s still live |
| What has to be true | Somebody’s present, reading it, with room to back away | You file, and you stay insurable at a price you’ll pay | The person can reach it and is willing to use it |
| What’s known at your size | Nothing measured | Nothing measured; the pricing formula is published | Nothing measured |
With no outcome data on any of them, what’s left is what each one is for and what each one needs to work. Both of those you can check against your own week tonight.
When is each one the right answer?
Each of them is right somewhere, and the somewhere is specific. Training earns its money where the trouble is verbal, repeated, and involves people your staff will see again. Insurance earns its money always. A response system pays for itself only where somebody is genuinely on their own.
Take the cheap option seriously when it fits. If what you need is a noise, a basic panic button off the shelf is a good answer, and it’s quicker and cheaper than anything else here. What separates it from a system is whether help has to arrive in the right room, and whether the thing has to work every time. If a loud alarm covers the situation you’ve got, buy the alarm.
Training is the wrong purchase when your exposure is one person at a register at closing. Insurance is the wrong thing to feel protected by and the right thing to carry anyway. A response system in a business where nobody is ever alone is money spent on a problem you don’t have.
Whatever you land on, budget for the drills. Handing over a device is a delivery, and the practice around it is what makes it work when somebody’s hands are shaking. Ask any vendor what happens in the week after the box arrives, and who runs the drill.
You started with three prices and a decision that felt like picking a winner. What you’ve got is three moments, and a business that already answers two of them. If somebody threatens a member of your staff and walks out, you know roughly how that goes: a conversation, maybe a claim, a premium that moves next year.
The gap is almost always the same one. Between the moment a situation turns and the moment it’s over, most small businesses have nothing, because the two options that are easy to price both sit on the far side of it. It’s the shortest of the three windows and the only one where anything can still change.
Every list of workplace violence prevention methods is really a list of three different jobs. Work out which of the three moments your business currently answers with nothing. That’s the purchase, and everything else on the list is a preference.
COMPARING YOUR OPTIONS
Cover the Minutes While It’s Happening
If the moment your business answers with nothing is the one while it’s still live, that’s the moment we build for. A demo shows how a press on one of our silent buttons reaches the people set up to respond with who pressed it and where, and what the included training library gives you for the drills.
References
- Understanding the Experience of Workplace Violence in Hospitals as Documented by Nursing Staff, 2025
- Effectiveness of De-Escalation in Reducing Aggression and Coercion in Acute Psychiatric Units: A Cluster Randomized Study, Frontiers in Psychiatry, 2022
- The Effectiveness of Workplace Violence Prevention Education Training Programs on Healthcare Professionals’ Confidence, International Nursing Review, 2025
- NCCI, Experience Rating Plan Methodology Update FAQs
- Violence against physicians and nurses in a hospital: How does it happen? Israel Journal of Health Policy Research, 2017