A specialist does not fill their own diary. It fills because colleagues send them patients: a family doctor, another specialist, a hospital department. The practice therefore depends on a flow it does not control, and that flow rests on one thing — what goes back to whoever referred.

What goes back is the only link

The referring doctor never comes to check. They send a patient, then they wait for a letter. If that letter arrives quickly and says what it should, they will refer again. If it does not arrive, they will not complain: they will refer elsewhere, and the practice will never learn why its flow thinned out.

It is a silent loss, and a structural one. Nothing in a practice’s tooling records who refers, or what was sent back to them. The information does exist — scattered through the files, one patient at a time, and nowhere as a state anyone can read.

That letter is not a courtesy

It tends to be treated as professional politeness. The law treats it as something else.

The French code of medical ethics requires a doctor to keep a file for every patient in their care, and to pass on to the doctors taking part in that care the information and documents useful to continuity of care — at the patient’s request, or with their consent.

On discharge from a health establishment the requirement gets far more precise. A discharge letter, written by the establishment doctor who treated the patient, is handed to them on the day they leave. It is sent the same day to their family doctor and, where applicable, to the practitioner who referred them, through a secure health messaging service or any other means guaranteeing confidentiality, and it is filed in the shared medical record where one has been opened. It states the treatments prescribed on discharge and those stopped during the stay, announces results still pending, and sets out the follow-up to be put in place.

“The same day” is the part of the text a practice organisation handles worst.

A third rule frames the rest. Within a single care team, information strictly necessary to coordination and continuity of care may be shared: it is deemed entrusted by the patient to the team as a whole. Outside that team, sharing requires their prior consent.

So the question is never “how do we make this move faster”. It is: how do we make it move fast, to exactly the right recipient, through a channel that stays under the practitioner’s control.

Where the work actually goes

The letter itself does not take long to produce. The cost is the chain around it.

A dictation waiting to be typed up, because the person who types it is on the phone.
A letter ready but unsigned, sleeping until the practitioner next comes through.
A report held back for a result that has not come in, and that nobody is watching once the week has moved on.
A recipient whose details have changed, discovered the day the envelope comes back.

None of these is serious on its own. Together they produce an absence: right now, in this practice, nobody can say which letters are waiting, since when, or which one is late against an obligation rather than against a habit.

It is not the letter that is short of time. It is having someone whose job it is to know, without having to think about it, which one never went out.

The diary is not the loop

The booking platform — Doctolib in most cases — knows who came, and when. The practice software knows what was done. Neither holds the loop: who referred this patient, what went back to them, what is still owed.

These tools are not bad; they were not built for this — and they are closed. The problem is therefore not solved by waiting for an opening that will not come, but in the tool as it is, on a machine the practice controls.

The line nothing crosses

An agent does not write the clinical content of a report. It interprets no result, concludes nothing, never passes itself off as the practitioner, and nothing goes out in the practice’s name without explicit agreement given before it is sent.

What it holds is the state: what came in, what went out, what is waiting on a result, what has run past its deadline. It prepares what can be prepared from what the file already contains, and it flags discrepancies rather than settling them — an uncertain recipient, a referred patient nothing ever went back for.

What stays with the practitioner

The examination, the conclusion, the signature, and everything that carries professional liability: none of it moves, and none of it should be able to move.

What does move is keeping the list. Desk work that shows up in none of the practice’s figures — until the quarter when a colleague quietly stops referring and nobody saw it coming.

What an agent takes on in a practice, with no practitioner named, is set out in our resources. How it is done is discussed face to face, not written down here.

The letter back to the referring doctor