A twelve lead ECG will not record without a patient identifier typed in, and the numbers are issued by the hospital's own system in a sequence. The trace stored at 03:40 carries a number the system has never issued, and it is a real recording of a real heart.
I am a cardiac physiologist and the twelve lead machines are mine to look after.
What the machine needs before it will do anything.
You enter a patient identifier, which on our system is a seven digit number, and it validates against the hospital's patient administration system over the network. If the number is not recognised the machine will not proceed. That is a deliberate control, because an ECG filed against the wrong patient is a dangerous document.
Then you attach ten electrodes, six across the chest and one on each limb, and you record ten seconds, and the machine stores the trace and sends a copy to the cardiology system.
It also records the impedance at each electrode, because a poorly attached electrode gives you a noisy trace, and it will tell you off if a lead is off.
Every few weeks there is a trace in the store from about twenty to four in the morning.
The identifier on it is seven digits and it is not one the hospital has issued. Not retired, not a deceased patient, not another trust's. Never issued.
That should be impossible before you get anywhere near what the trace is, because the machine checks the number against the system before it will record.
I have tried. I have typed in numbers that do not exist, on both machines, and it refuses, and it says so on the screen.
What is on the trace.
It is a recording. It is not noise, it is not a flat line, and it is not the fifty hertz interference you get from a lead lying on a bed.
Twelve leads, all with plausible morphology, impedance in range on all ten electrodes, which means ten electrodes were attached to something with the electrical properties of skin.
The rate is between forty four and forty eight. The rhythm is sinus. There is nothing dramatic about it: a cardiologist looking at it without being told anything would say it was a slow but normal trace, probably somebody fit, probably somebody asleep.
I have shown one to a consultant in that spirit, without saying where it came from, and that is what she said.
What has been done.
Both machines have been swapped. The one that did it is now on a different site and has not done it there in two years.
The network validation has been checked by our informatics team, who confirmed that the machine cannot proceed without a positive response from the patient system, and who then found in the audit trail that no query was made for that number on that night.
So the machine did not ask, and the machine will not proceed unless it asks.
The cardiology system's own logs show the trace arriving from the machine at twenty to four with its identifier.
Where the machine is at that hour.
On its trolley, plugged into a wall socket to charge, in the alcove opposite the lifts on the second floor, which is where it lives at night.
The electrode leads are coiled on the hook on the side of it. The electrodes themselves are single use and they come in a sealed packet, and the packets are counted because they are stock.
The count does not move on those nights. I checked that for a year, because if ten electrodes had been used I would have expected the box to go down, and it does not.
Which means ten electrodes with correct impedance, and no electrodes.
What is done about it.
The traces are in the cardiology system with an unmatched identifier and they sit in a reconciliation queue that somebody in medical records looks at monthly. There are about thirty in there now.
They cannot be deleted, because nothing in a clinical record can be deleted, and they cannot be matched to anybody, because there is nobody.
I raised it as an incident in 2022. It was closed as an equipment fault with no patient impact, and that is a fair conclusion from where the person closing it sat.
What I do.
I service the machines, I do the calibration checks, I keep the stock.
The alcove opposite the lifts is where the trolley lives and I have not moved it, because there is nowhere better and because moving it would be doing something about a thing I have decided not to be frightened of.
What I have changed is that I do not go up there for it at night.
If a ward wants that machine after midnight they come and get it, which is what they mostly did anyway.
In April 2022 I went up at about half past three for a request from the assessment unit, and the lifts were still and the corridor lights were on the night setting, and the trolley was where it always is.
The leads were not on the hook. They were laid out on the top of the trolley in the order you lay them out, V1 to V6 left to right and the four limb leads set out at the corners, which is a thing you do when you are about to put them on somebody, and which takes about twenty seconds, and which nobody in this hospital does except me and the two people I trained.