Short answer
An isolation room provides containment only while its pressure differential is maintained, typically negative by at least 2.5 Pa for airborne containment or positive for protective isolation. A failed damper, a loaded filter, a change in demand elsewhere on the air handling unit or a door held open removes that containment without any indication in the room. Monitoring has to treat pressure as a relationship between room pairs, logged continuously, rather than as a gauge someone reads.
Containment is a relationship, not a reading
This is the conceptual mistake that shapes everything else. People talk about the pressure in an isolation room as though it were a property of that room, like its temperature. It is not. It is a difference between that room and what is next to it, and it exists only in comparison.
That means three things. A single sensor cannot tell you whether containment holds. The relevant relationship may involve an anteroom, a corridor and sometimes an adjacent ward, forming a cascade rather than a pair. And a change in any of those spaces changes the relationship, even when nothing in the isolation room itself has changed at all.
The failure modes, and why none of them alarm
01
Filter loading on the serving AHU
As filters load, delivered airflow falls. The room may hold its temperature perfectly while losing the airflow margin that produced the differential, because temperature control and pressure control are being delivered by the same fan.
What it costsThe room feels correct, reads correct on temperature, and is no longer containing.
02
Demand changes elsewhere on the system
Where an isolation room shares an air handling unit with other spaces, a change in demand elsewhere redistributes airflow. The cause of the containment loss is in a different room and possibly on a different floor.
What it costsDiagnosis is nearly impossible without seeing the whole system at once.
03
Doors, and ordinary clinical work
A door held open during a transfer collapses the differential immediately. This is routine, unavoidable and usually brief, which is exactly why it is never recorded and why it is hard to distinguish a normal transfer from a door left propped.
What it costsWithout duration data you cannot tell a two minute transfer from a forty minute lapse.
04
Damper and actuator drift
Control dampers stick, actuators lose calibration and balancing drifts from where the commissioning report says it is. This is slow and there is no fault code for a damper that is 15 percent off.
What it costsThe room quietly leaves the condition it was commissioned into, and the commissioning report still says it is fine.
What continuous containment monitoring looks like
- Model the relationship, not the room. Configure room pairs and cascades so the system knows what each space is meant to be positive or negative against.
- Log continuously. The value of the record is that it has no gaps, which is what lets you state whether containment held for a given patient episode.
- Alert on duration, not just threshold. A momentary collapse during a transfer is normal. The same collapse for forty minutes is not. Only duration distinguishes them, so the alerting logic has to include it.
- Correlate with the AHU. Pressure loss caused by filter loading looks identical in the room to pressure loss caused by a stuck damper. Seeing the serving unit alongside the room separates them.
- Include door events where available. A differential that collapses at the moment a door opens needs no investigation. One that collapses with the door shut needs it urgently.
Why it matters beyond the accreditation file
It is easy to frame this as a compliance exercise, and continuous logs certainly make accreditation evidence straightforward. But the reason to do it is narrower and more serious than that.
An isolation room exists to protect people who are not in it. When containment fails, the harm lands on other patients, on staff and on visitors, and it lands invisibly. The record is not primarily for the surveyor. It is so that the question of whether containment held during a particular admission has an answer at all.
What pressure differential does an isolation room need?
For airborne infection isolation the room is held negative relative to adjacent spaces, commonly expressed as 2.5 Pa or more. Protective isolation for immunocompromised patients reverses the relationship and holds the room positive. Exact values follow the hospital protocol and the applicable standard for the room class.
Why is a pressure gauge outside the room not sufficient?
It shows the differential at the moment somebody looks at it. It has no memory, does not alarm and cannot show what happened overnight or during a specific patient episode.
For continuous evidence, and for alerting while a lapse can still be corrected, the differential needs to be logged rather than displayed.
How do you avoid constant false alarms from doors opening?
Alert on duration rather than on the threshold alone. A collapse lasting seconds during a transfer is normal clinical activity. The same collapse persisting for tens of minutes is a real condition.
Where door state is available, correlating it with the differential removes almost all remaining noise, because a collapse with the door shut is immediately distinguishable from one caused by a transfer.
