Piped oxygen is metered into every zone, and the tank outside is weighed continuously, because running out is a disaster with a public inquiry attached. Ward 4 draws forty litres a minute that no mask, no ventilator and no leak accounts for.
I am an authorised person for medical gases at a district general hospital. Eleven years in the role and twenty six in hospital estates.
How the oxygen gets to a patient, because almost nobody in the building knows.
Outside there is a vacuum insulated evaporator, which is a tall steel vessel of liquid oxygen standing on load cells. It is weighed continuously and the telemetry goes to the gas company, who deliver on that figure.
From there it goes through pressure regulation into a ring main, and off that into zones, one per ward or department, each with an isolation valve in a box with a glass front and a zonal flow meter.
Demand is not a mystery. A nasal cannula is one to four litres a minute. A simple mask is five to ten. A non rebreathe is fifteen. A ventilated patient is worked out on the circuit. You can walk a ward with a clipboard, add up every device in use, and you will land within a few litres a minute of what the zone meter says.
I have done that with a student engineer as a teaching exercise about thirty times.
Ward 4 is a twenty eight bed medical ward on the second floor.
Add up the devices and you get a number. The zone meter reads that number plus about forty litres a minute.
All day, every day, for the eleven years I have had the job, and according to the logbook for eight years before that.
Forty litres a minute is a lot. It is three patients on a non rebreathe mask, or about two and a half cubic metres an hour, or the thick end of twenty thousand pounds a year.
The hospital's total is about nine hundred litres a minute, so it is not a number that shows up on a bill as alarming. It shows up as one ward being busy.
What a leak of that size should do, which I have to explain to every new person who looks at this.
A leak is the right first answer. Pipework leaks, valves pass, and a hospital has kilometres of copper in ceilings.
You find one with a pressure decay test. You isolate a section, bring it to working pressure, shut the valve, and watch the gauge. A section that holds is sound. That ward has been tested section by section under permit nine times in my tenure, twice by a specialist contractor with a calibrated test set, and every section holds for the required period with no measurable drop.
You also find leaks by smell and by sound and by soapy water on a joint, and there is a third way that is better than all of them: oxygen enrichment. Forty litres a minute of oxygen leaking into a ceiling void raises the oxygen concentration in that void, and an enriched atmosphere is a fire risk, so we carry monitors for exactly this.
Twenty point nine percent in every void on that ward. Which is air.
Where the meter says it goes.
I fitted a second meter in 2019, downstream of the zone valve, a different make, calibrated, because a suspect meter is the cheapest explanation and I wanted to kill it.
Both meters read the same.
Then I did the one thing that makes this story the thing I think about on the drive home.
I closed the zone valve.
Not in use. A planned shutdown at two in the morning, on a Sunday, ward cleared to sixteen patients, every one of them assessed as not requiring oxygen, two anaesthetists and a matron present, cylinders at every bed as backup, the whole thing under permit with a signed risk assessment, which is how you do it and the only way you do it.
With the zone valve shut, the downstream meter reads zero, because it must.
The upstream meter, between the ring main and the valve, read thirty eight litres a minute for the forty minutes the valve was closed.
What happened to that finding.
I wrote it up and it went to the medical gas committee, which includes a consultant anaesthetist, the head of estates, pharmacy and the gas company's engineer.
The conclusion, minuted, is that the upstream meter is faulty and should be replaced.
It was replaced. The new one reads the same.
The minute from the next meeting records that the matter is closed, that there is no clinical risk, and that the reported consumption is not material against the trust's total usage.
Both of those things are true. That is why it is closed.
What I do about it now.
I do the statutory testing, I keep the permits, I sign the certificates, and the figure goes on the annual return in the column for that zone, and nobody has ever queried it.
The gas company delivers on the weight of the tank, and the tank knows. That is the part I find hardest. Forty litres a minute is about fifty eight cubic metres a day, and the tank outside goes down by it, and a lorry comes and puts it back.
So whatever it is, it is breathing.
I have changed one thing. The permit for any work on that zone used to allow a single authorised person to supervise, which is me, and I have had the procedure amended so it requires two.
The wording I used in the amendment is that the zone has a history of anomalous flow readings which may complicate a shutdown. That is true and it passed without comment.
The reason is that when I closed that valve at two in the morning in 2019 I was standing at the valve box in the corridor with the matron twenty feet away, and the thing I have never written down is that the hissing in the wall behind the box did not stop when the flow did.