The containment is up. Whether it did what it was supposed to do is a different question — and it is the question no guidance helps you with.
What closes that gap is measurement during the works rather than an assessment afterwards. And it takes three measurement positions, not two: one in the works zone, one in the adjacent critical area, and one reference point elsewhere in the building. Without that third point you cannot attribute a rise observed next to the works, and your data is unusable in the only conversation that matters.
Why this comes up
When construction starts inside a working hospital, containment is installed. Screening goes up, the works zone is placed under a ventilation regime, the adjacent critical areas are protected.
All of that is visible. Whether it held is not.
That is the structural problem. Containment is specified as a set of measures and verified as an impression — a walk-round, and the contractor's assurance. Both are given in good faith. Neither produces anything when a complaint arrives from the ward next door three weeks later, by which time the works have moved on and the screening has come down.
Which guidance applies, and which function inside the hospital carries the responsibility, differ from one country to the next. The distance between what is prescribed and what can be evidenced does not.
The common structure — and the common gap
National guidance on infection control during construction differs in detail, in legal weight and in which body issues it. What it very largely shares is a structure, and that structure has a well-documented ancestor: the Infection Control Risk Assessment (ICRA) approach, published in its most widely used form by the American Society for Health Care Engineering and adapted into national and institutional guidance in many countries.
The logic is the same wherever it appears:
| Step | What it classifies |
|---|---|
| 1 | The construction activity, in four types — from non-invasive inspection, through small-scale short-duration work producing minimal dust, to moderate-to-high dust generation, to major demolition and construction |
| 2 | The vulnerability of the patients on the other side of the wall, in four groups — from office and non-clinical areas through to intensive care, operating theatres, oncology, transplant and protective isolation |
| 3 | The combination, which yields a class of precautions — typically four, from least to most restrictive |
The heaviest classes are where the substantial measures live: dust-tight screening from floor to structural ceiling, a defined ventilation regime for the works zone, and protection of the surrounding areas where the most vulnerable patients are.
And here is the finding that makes this an international question rather than a local one. The ICRA matrix itself contains no air sampling, no particle counting, no measurement method, no threshold value and no monitoring frequency to verify that the containment worked. It specifies procedures; it does not specify how you would know whether they were effective.
That absence travels with the framework. Where national guidance adopts the matrix — and it very commonly does — it tends to adopt the gap along with it. The measures are prescribed in detail. The verification is not prescribed at all.
A 2025 review in the Journal of Hospital Infection, surveying national and international guidance on reducing exposure to Aspergillus in healthcare settings, reached a conclusion in the same direction: that more consistent and comprehensive recommendations are needed.
What the guidance does ask for, and why it is not enough
Where verification appears in guidance and in hospitals' own implementing procedures, it takes a consistent form: site visits during the works, an inspection at the end, and a requirement that the physical integrity of the screening be checked regularly.
Those are inspections. They tell you what should be looked at. They do not tell you what it should satisfy, because there is no criterion attached to them.
The imbalance inside a single document is often striking. Guidance can be precise enough to require that an opening in a wall be repaired or temporarily sealed within a defined number of hours, and entirely silent on how you would demonstrate that the same wall held for the other seven hundred hours of the month.
This is not a criticism of the hospitals implementing it. Read a set of institutional procedures side by side and you find faithful implementations: the risk matrix is there, the guidance is cited, the measures are correctly translated. The gap sits in the guidance, not in its execution — which is exactly why it is so persistent.
Why two measurement points give you no answer
The obvious arrangement is one instrument in the works zone and one in the adjacent critical area. Reasonable enough: that is where it is released, and that is where you do not want it.
Now suppose the value in the adjacent area rises. What do you know?
Three explanations look identical on two instruments.
- The screening is leaking. Dust is passing from the works into the adjacent area.
- The whole building went up. It was a bad day outside, and that carries indoors.
- Something happened in that room itself. A door left open, a cleaning round, a bed being made, a movement.
Only the first concerns the works. You cannot charge the other two to the contractor, and the contractor knows it. A chart with two lines is not evidence in that conversation; it is an invitation to a discussion.
The measurement design: three positions
| Position | What it demonstrates |
|---|---|
| 1 — works zone | What is actually released during the works, and at which hours of the day |
| 2 — adjacent critical area | Whether the screening and the ventilation regime are holding. This is the room you are protecting |
| 3 — reference point elsewhere in the building | What lets you separate a works effect from a building or outdoor-air effect |
The third point carries the whole argument. If the reference rises at the same moment and by the same amount, it is not the works — explanation 2. If the reference stays flat while the adjacent area climbs, you have ruled out explanation 2 and two remain. Cross that against what the works-zone instrument was doing at the time, and usually one plausible explanation is left standing.
This is less than proof. Three measurement points do not prove a cause. They rule out a class of explanation. But that is precisely enough to hold a conversation that otherwise cannot be held, and it is immeasurably more than a walk-round produces.
That is why three is a measurement design rather than a sales figure. The third instrument does not measure more; it makes the other two interpretable.
What continuous measurement shows and a walk-round does not
The outdoor-air effect in explanation 2 is not a theoretical possibility. It is the reason we started placing the third point.
The observation below comes from a controlled room in a hospital in Belgium. No construction — a room with mechanical ventilation and an active air cleaner, where nobody expected anything. No complaints, no incidents. Thirty days, 9,862 measurements of particles ≥ 0.5 µm.
- 68.5 % of measurements sat above the reference line agreed with the customer — derived from ISO 14644-1 Class 7 for particles ≥ 0.5 µm, used as an operational reference and expressly not as a classification.
- The longest continuous exceedance lasted eight days.
- Of 227 elevated episodes, 18 carried a simultaneous CO₂ or VOC signal — the indicators of activity inside the room.
- Roughly two thirds of the episodes coincided with elevated outdoor air.
That last figure is the point. In a room with no works anywhere near it, the large majority of the rises followed the air outside the building. A building tracks its surroundings more closely than most people assume.
Set that beside a construction situation and the consequence is immediate: without a reference point, every one of those episodes would have looked, measured next to works, like a leak in the screening.
Two things belong here honestly. This is not construction data — it is a different room, with no works, and it proves nothing about what screening does. What it demonstrates is the method: no walk-round sees this, and without a third measurement point it is not distinguishable from a works effect. And it is one room over one month; it is an observation, not an established pattern.
What this means in practice
The conversation with the contractor shifts from impression to record. Today, dust control is a clause in the contract and compliance with it is a matter of trust. If dust reaches the ward next door, a discussion without evidence follows — and the hospital usually loses that discussion. With a measurement series you raise it with the contractor on the day, with a curve, rather than afterwards with an impression.
The requirement can go into the tender documents. This is the implication most often overlooked. Written into the contract as a requirement, the cost can sit with the contractor rather than with the hospital. That turns the question from an expenditure into a specification.
Complaints get an answer rather than a reassurance. Works inside a running hospital generate complaints. With a measurement series you answer them with data — and if something genuinely is wrong, you know before the complaint arrives.
And something remains afterwards. A series covering the full construction period is usable in the project file and in the final account. A walk-round leaves nothing behind.
Where this is not needed, and what we do not do
No adjacent critical area, no question. If nothing you are protecting borders the works, this does not apply. Nor does it at a low risk class: guidance conditions its heaviest measures on that classification for good reason, and measuring where even screening is not called for is a solution in search of a problem.
And no guidance obliges any of this. The frameworks in use are recommendations rather than requirements in most jurisdictions, and none of them asks for measurement. Anyone who suggests a standard imposes this on you is mistaken. The argument for measuring is not that it is required — it is that you otherwise carry a duty of oversight with no instrument to discharge it.
Beyond that, and we would rather say it in advance than afterwards:
- No ISO 14644 classification, no qualification certificate, no validation, no certificate. That remains with the party that qualifies your rooms, and it should stay there.
- No microbiology. We count particles, not organisms. Microbial counts and surface sampling are laboratory work.
- No statutory exposure measurement. That requires an accredited laboratory. Continuous monitoring complements it and does not replace it.
- And measuring does not prevent dust migration. A measurement produces evidence. The screening, the ventilation regime and the people maintaining them prevent the migration.
InsightAir provides continuous air quality monitoring in critical environments. What happens between two point measurements is usually the most interesting part.
If works are running, or one is planned next to a critical area, half an hour on site is the usual first step: what is under way, which rooms adjoin it, and whether this makes sense in your situation — including the conclusion that it does not.
