Walk your own building the way we would.
This is the walk we do on a first visit, written down so you can do it yourself — outside-in, the way you already move through the place. It names what makes each area hard and what to look at. It does not tell you what to buy. A couple of things will land that hadn’t before; that is the whole point of a first step.
Nothing here is gated. No email, no login, no form. Read it, print it, mark it up, hand it to whoever should see it.
Which walk are you actually being sold?
“A walkthrough” is three different things wearing one word, and it is worth knowing which one is in front of you. A risk assessment starts with your documents and your incident history and produces a finding only after analysis — days later, not in the parking lot. A spatial walk — this book — is a pass through the building, layer by layer, to see what the place is doing. An integrator’s site survey starts at the equipment already on the walls and often produces a parts list and a proposal before the visitor has left.
None of those is wrong. But an output that appears before the surveyor is out of the building is shaped by what it is for: a finding takes analysis the walk itself cannot hold, and a quote written on the spot has been addressed to the one person who happened to be standing there — not to facilities, risk, clinical leadership, IT, and finance, who all have a hand in the decision and none of whom were on the walk.
What this book is, and what it isn’t.
This book is the spatial walk. It is a genuine first step, and it is one station in a larger method. As far as we could find, no published standard defines a single canonical hospital security walk — but the one reasoned professional method we know of does not start at the fence. It starts at your documents: the incident log, the drawings, the scope of what you are actually protecting. This book walks the building because that is how you think about it; a full assessment starts before the walk, at the paperwork.
So use this to find the two or three things worth a real conversation — then have that conversation with someone who can start at the documents.
The property
What you’re looking at. the first thing anyone meets and the last place your coverage is dense. Open sky, cars, and the distance between a parked car and a staffed door, at every hour you operate.
What makes it hard. distance and thin crossing. The lot is far from staffed space and lightly populated, so whatever happens there happens with few people near and a long way to a door — and darkness is a condition that changes by the hour, not a fixed feature you can design around once.
What to look at
The farthest space from a staffed door, and how long the walk from it actually takes. Which parts of the lot go dark first at dusk. Whether the busiest exit and the thinnest-covered corner are the same place at shift change.
A good answer sounds like
“We know our lowest-traffic hour and our longest walk, and they don’t overlap by accident.”
A thin answer sounds like
“It’s lit and there are cameras” — coverage described, not the isolation that makes the lot hard.
If the answer is “I don’t know.” the walk has never been timed with a security lens, which usually means the lot was planned for parking and inherited security after.
Who else has a hand in this: facilities, grounds and lighting, the parking operator, and often risk management.
The parking structure
What you’re looking at. levels with no single vantage, enclosed stairs and elevators that are the only way through on foot, and light that changes by floor and by hour. Treat it as its own building, because it behaves like one.
What makes it hard. the geometry works against being seen: verticality with no place to watch the whole, blind corners and columns, and enclosed routes that are the least-observable spaces yet the only pedestrian paths. And a physical fact most spec sheets never mention — concrete scatters sound and confuses the systems that listen for it. Equipment that locates a sound by when it arrives across several microphones depends on a clean signal; a reverberant concrete structure is exactly the environment that muddies it, so a component validated in an open or corridor-like space can behave differently here, and the datasheet won’t tell you where that line is.
What to look at
Stand on a middle level and count the blind corners from one spot. Take the enclosed stair from a car to the door the way anyone on foot would, and notice who could see you. Clap once, or let a car door slam, and listen to how the sound smears.
A good answer sounds like
“We’ve walked it at 6pm, we know which level goes dark first, and we know the stair is the least-seen space we own.”
A thin answer sounds like
“It’s the same as the lot, just covered” — the tell that the structure’s own geometry hasn’t been walked.
If the answer is “I don’t know.” the structure has been treated as parking with cameras added, and its acoustics and sightlines have never been tested against what the equipment assumes.
Who else has a hand in this: the parking operator, facilities, and whoever specified the equipment now inside it.
The entrance
What you’re looking at. the one opening the whole building funnels through, and the one place where “secure” and “welcoming” pull hardest against each other.
What makes it hard. openness itself is not the problem — a hospital has to be enterable. The hard part is whether the openness is a decision someone made and can account for, or a default that accumulated. Count the ways in and how many are attended; the number is usually higher than the one people say.
What to look at
How many doors actually open to the inside, and how many are screened or attended. Who decided a given side door stays open through the day — and whether that decision is written down anywhere.
A good answer sounds like
“Every open door is open on purpose, and I can show you where that was decided.”
A thin answer sounds like
“That door’s always been open” — a habit, not a decision, and it will read as one when someone asks.
If the answer is “I don’t know.” nobody owns the front-door decision day to day, which is common and is itself the finding.
Who else has a hand in this: registration and front-of-house, nursing, patient experience, security, and facilities.
The public interior
What you’re looking at. the space a visitor moves through before anyone has decided whether they should be there — and the routes that quietly lead from public areas toward places that aren’t.
What makes it hard. the interior is designed to help a stranger find their way, which is the same thing as helping a stranger move unchallenged. Verification here is a human habit, not a device, and habits vary by who is at the desk.
What to look at
Walk in the way a visitor would and notice whether anything asks who you are or where you’re going. Ask someone at the desk how they’d check a person they didn’t recognize.
A good answer sounds like
“There’s a way we verify someone, the desk knows it, and it doesn’t depend on who’s working.”
A thin answer sounds like
“People know each other here” — recognition is not verification, and it fails exactly when it matters.
If the answer is “I don’t know.” verification lives only in what the day shift happens to know — which means it isn’t a process yet.
Who else has a hand in this: registration, patient experience, nursing, and security.
The departments
What you’re looking at. five areas that each break a different way. The ED’s screening carve-out for critical patients is correct and is also a permanent unscreened corridor. Behavioral health is the one place a security device can itself become the hazard. Maternity’s risk is social, not a defeated door. The pharmacy’s access layer says yes to the authorized. And elopement turns on a judgment made hours earlier, elsewhere.
What makes it hard. each area’s control points a different direction, so a single mental model fails. The thing that makes the ED hard is not the thing that makes behavioral health hard, and treating them alike is how a gap hides.
What to look at
In the ED, the door the screening policy exempts. On the behavioral-health unit, whether the mounted hardware itself could be a fixed point — no minimum height. In maternity, whether the tag system was last tested or only installed. At the pharmacy, whether the record reconciles across systems. For the wanderer, whether the exit knows the risk flag.
A good answer sounds like
“Each of these has its own answer, and I can tell you what’s different about all five.”
A thin answer sounds like
“They’re all locked down” — one model applied to five different problems.
If the answer is “I don’t know.” one of the five has never been walked on its own terms — each has a dedicated room on our site for exactly that reason.
Who else has a hand in this: clinical leadership for each unit, pharmacy, nursing, risk, and security — never security alone.
The staff’s day
What you’re looking at. the building as the people who work in it experience it — where they can summon help, whether it reaches them where they actually are, and the trip from a far parking space to the door at the end of a shift.
What makes it hard. the fixed panic button is in the wrong place the moment a nurse is inside a room, and the call for help is only as good as its weakest link — it has to reach a person, tell them where, and leave a record. The walk to the car is the property’s isolation, felt by a specific tired person at the worst hour.
What to look at
Where duress lives relative to where staff actually stand. Whether a call reaches a person carrying a device or a screen nobody watches. Whether the longest walk to a car crosses the thinnest coverage at shift change.
A good answer sounds like
“A call for help reaches a named person with a location, and we know our worst walk.”
A thin answer sounds like
“There are buttons at the stations” — coverage at the desk, not where the work happens.
If the answer is “I don’t know.” the response chain has never been tested end to end, only device by device.
Who else has a hand in this: nursing, security, facilities, and HR — the program is theirs; the physical layer is what we build.
The systems that watch it all
What you’re looking at. every camera, badge, and alarm as one chain rather than a shelf of devices — and the honest question of whether you’d know if one of them quietly stopped.
What makes it hard. the gaps live at the seams, invisible from inside any single system. Each device passes its own test; what nothing tests is whether the event that should have recorded reached the disk, whether the denied badge also raised a camera, whether the alarm at 3am was ever seen. A device that fails loudly gets fixed; one that fails silently waits until someone needs what it should have held.
What to look at
Pick last Tuesday and try to retrieve who opened a given door, on the same timeline as the video. Note how long it took. Find the door people prop, and ask what the workaround is working around.
A good answer sounds like
“I could reconstruct an event across systems in minutes, and I know which doors get propped and why.”
A thin answer sounds like
“Everything’s recording” — recording is not retrievable, and nobody has tried to pull it whole.
If the answer is “I don’t know.” the whole chain has never been tested together under real conditions — only the devices, one at a time.
Who else has a hand in this: IT, security, facilities, and every clinical team whose workaround routes around a control.
What’s about to change
What you’re looking at. any build or renovation on the horizon — because security infrastructure installs in a fixed order, and the cheap moments to decide it pass quietly.
What makes it hard. the drawings decide before the concrete does. Pathways are committed at construction documents and the trades build to them within weeks; the building’s geometry — adjacencies, stair placement, secure-versus-public zones — closes even earlier, at schematic design. Once a surface closes over a path that was never run, adding a device later is opening finished work, not making a purchase.
What to look at
What stage any current project is in right now. Whether space and power are reserved for a security head-end you haven’t selected. Whether anyone who knows what the system needs was in the room when the paths were drawn.
A good answer sounds like
“I know what stage we’re in and what it’s already closed behind us.”
A thin answer sounds like
“We’ll handle security at the end” — the end is after the cheap decisions are gone.
If the answer is “I don’t know.” no one has mapped the security decisions to the construction calendar — which is the most common answer, and the reason to ask now.
Who else has a hand in this: the architect, capital projects, the general contractor, the electrical engineer, IT, and the authority having jurisdiction.
After
What you’re looking at. the building on its worst day, read backward — not to relive it, but to ask whether you could produce a defensible account of what your systems saw and did.
What makes it hard. the questions after an incident are evidentiary, and they arrive late: an investigator, counsel, or the board asks for the timeline months later, when retention windows may already have closed. An account that exists in six unreconciled systems is not an account you can produce.
What to look at
Whether one authorized person can retrieve the full event — activation, location, access, video — time-synchronized, without stitching systems by hand. Whether video retention outlasts how long an investigation takes to begin.
A good answer sounds like
“One person could produce the whole timeline, and it would still exist when we were asked for it.”
A thin answer sounds like
“We’d pull the footage” — footage is one system; the account is all of them, aligned.
If the answer is “I don’t know.” the record has never been assembled end to end except during an actual incident — which is the worst time to discover the gaps.
Who else has a hand in this: risk, legal, security, IT, and the governing body the report goes to.
You walked it. Now the two or three that stuck.
If this worked, a couple of things landed that hadn’t before — a walk you’d never timed, a door nobody decided, a record you’re not sure you could produce. That is exactly what a first step is supposed to do: not settle the question, but show you which questions are worth a real answer.
The next step isn’t a purchase. It’s a proper assessment that starts where this book said the real method starts — at your documents and your incident history — and turns the two or three things you found into a finding you can act on. When you want that, the door’s open.
You’re the expert. What did we miss?
We do this work, but you live in your building every day, and the research only goes so far. So here is the honest ask: read this against your own place, and tell us what’s wrong with it. What’s missing? What would you add? Is it laid out the way you’d actually walk it? We’d rather have your correction than your email address.
There’s no form here on purpose. If something in this guide is wrong for a hospital like yours, tell us and we’ll fix it — the next version is better because you walked yours.
Illustrative demonstration guide on a demo site. On a live deployment this ships under the integrator’s brand and may carry that integrator’s own named, verified project results.