Why 900 MHz RTLS Is Failing Hospitals: Coverage Gaps, Latency, and the Life-Safety Problem

By the MGM Solutions team — life-safety RTLS specialists since 2002.

Every few years a new radio band gets marketed to hospitals as the answer to real-time location.
In the 2010s it was Wi-Fi. More recently it has been Bluetooth Low Energy. In between, a number of
facilities were sold on 900 MHz RTLS — sub-GHz active RFID promising long range,
fewer readers, and a lower installed cost than anything else on the market.

The range claim is real. The problem is what hospitals actually bought with it. Long range and
precise location are in direct tension, and 900 MHz systems resolve that tension by giving up
precision. For inventory that is an acceptable trade. For patient elopement prevention,
staff duress, and infant security — where the answer has to be a specific door at a specific
second — it is not.

This article walks through the five failure modes we see most often when facilities call us to
replace or supplement a 900 MHz deployment, and what to require instead if life safety is on the
line.

What 900 MHz RTLS Actually Is

900 MHz systems operate in the sub-GHz unlicensed ISM band. Tags transmit a beacon; fixed readers
hear it; software estimates where the tag is based on which readers heard it and how strongly. Because
lower-frequency signals travel farther and penetrate construction better than 2.4 GHz signals, a single
900 MHz reader can cover a very large area — sometimes an entire wing.

That is exactly why the reader count is low and the quoted price is attractive. It is also exactly
why the location answer is coarse. If one reader covers a wing, then “this reader heard the tag” means
“the tag is somewhere in the wing.” Vendors close that gap with signal-strength math — RSSI
trilateration, fingerprinting, machine-learning models trained on your building. All of these are
inference. None of them is certainty.

Failure Mode 1: Zone-Level Answers Sold as Room-Level

The single most common complaint we hear is that the system’s location accuracy was demonstrated in
an empty conference room and never reproduced on a live unit. RSSI-based positioning is sensitive to
everything a working hospital contains: metal bed frames, lead-lined imaging rooms, crash carts, fluid
bags, elevator shafts, people. Signal strength varies with all of them, and the estimate drifts.

A drift of ten or fifteen feet does not matter when you are looking for an IV pump. It matters
enormously when the question is “is the patient in room 412 or in the stairwell across the hall,”
because those two answers trigger completely different responses. Staff learn quickly which alerts to
trust. Once they stop trusting the map, the system stops being a safety system regardless of what the
dashboard says.

Failure Mode 2: The Band Is Crowded and Unprotected

The 900 MHz ISM band is unlicensed and busy. Utility meters, industrial telemetry, cordless
equipment, logistics tags, and other RTLS deployments all share it. Hospitals are also dense RF
environments by nature. When the band gets noisy, beacons are lost, and a lost beacon looks exactly
like a tag that has not moved.

Unlicensed does not mean unusable — every practical RTLS technology, including ours, uses
unlicensed spectrum. What matters is whether the system’s life-safety decisions depend on
clean RF. A design that infers position purely from signal strength is fragile in a noisy band. A
design that confirms position with a second, non-RF signal is not.

Failure Mode 3: Latency at the Only Moment That Counts

Battery life on any active tag is a function of how often it transmits. Long-range sub-GHz
transmissions are relatively power-hungry, so vendors extend battery life by lengthening the beacon
interval — sometimes to thirty seconds or several minutes for stationary assets.

Now do the elopement math. A person walking at a normal pace covers roughly 250 to 300 feet per
minute. If a tag beacons every thirty seconds and the software needs two or three consistent readings
before it declares a location change, the patient can be out the door and across the lot before the
first alert fires. The system is not wrong; it is late. In wander management, late and wrong are the
same outcome.

This is why we build chokepoint detection on Low Frequency (LF) field triggers at
doorways and elevators rather than on beacon-interval polling. The tag is woken by the LF field the
instant it enters it, and the exit event is reported immediately — independent of how often the
tag happens to be beaconing.

Failure Mode 4: Infrastructure Debt and the Upgrade Trap

Facilities buy 900 MHz for the low reader count, then discover that improving accuracy means adding
readers — and adding readers in a hospital means conduit, ceiling access, infection-control
permits, and after-hours labor on occupied units. The cheap system becomes an expensive one, one
change order at a time, and the accuracy ceiling is still set by the underlying method.

Worse, most of these platforms are single-purpose. The wander system does not talk to the duress
system, which does not talk to the asset system, so each new use case means another vendor, another
tag, another head-end server, and another integration to nurse call. See our breakdown of
why Wi-Fi, BLE, and 900 MHz fall
short for life-safety RTLS for the full architectural comparison.

Failure Mode 5: Protection Stops at the Building Line

Almost every wander-management deployment we are asked to review protects doors and stops there.
The alert fires as the patient crosses the threshold, and then the map goes dark. Staff are told
someone left; they are not told where the person went.

That gap is the reason MGM repositioned around campus-wide coverage.
SecurTRAK delivers elopement prevention, staff duress, infant security, and asset tracking
both indoors and outdoors on a single infrastructure — from the patient room to the
parking lot, one tag, one map, one alerting pipeline. We cover this in depth in
Campus-Wide
Elopement Prevention: Why Indoor-Only Systems Leave Patients at Risk.

900 MHz vs. Wi-Fi vs. BLE vs. Triple-Technology RTLS

The distinction that matters in this table is not frequency — it is certainty versus
inference. Infrared cannot pass through a wall. When an IR receiver in room 412 reads a tag,
the tag is in room 412; there is no probability attached. The 433 MHz link then carries that confirmed
read building-wide, through walls and floors, and LF handles doorways and chokepoints. Three
technologies, one tag, each doing the job it is physically good at. That is the whole design argument,
and it is why we are comfortable putting our platform behind
wander and elopement
management and wireless staff
duress rather than asset tracking alone.

What to Require in Your Next RTLS RFP

1. Demand a live, occupied-unit demo. Empty-floor accuracy is not accuracy. Walk the
   route a patient would actually take, at walking speed, during a shift change.
2. Ask how room-level location is determined. If the answer contains the words
   “estimate,” “algorithm,” or “trained model,” you are buying inference. Ask what happens when the model
   is wrong and who notices.
3. Specify time-to-alert, not update interval. Write an acceptance criterion in
   seconds, measured from crossing a door to an alert appearing on the responder’s device.
4. Ask what happens outside. Get the coverage boundary drawn on a site map, including
   parking areas, courtyards, and the route between buildings.
5. Count the tags. If patients, infants, staff, and equipment each need a different
   tag and a different head-end, you are buying four systems and paying to integrate them.
6. Confirm the integration list in writing. Nurse call, access control, elevators,
   CCTV, and paging — named products and versions, not “open API.”

If you want a second opinion on a 900 MHz proposal already on your desk, we will read it with you.

Frequently Asked Questions

Related reading:
BLE vs 433 MHz for
Hospital RTLS  • 
The $543 Million Lesson:
Why Wi-Fi-Based RTLS Failed at the VA  • 
How MGM Solutions Is Different