A patient you last saw in January is sitting down in front of you. They have brought a plastic bag. Inside it are eleven printouts, four of them photocopies, two of them the same report printed twice. They also have a phone, and on the phone are photographs of three more reports, taken at an angle, in a gallery containing nine hundred other pictures.
They are being helpful. This is genuinely everything they have.
The consultation is seven minutes long. Somewhere in the bag is the reason you changed their medication in January, and it is the one thing the bag cannot tell you.
The history exists. It is just not in an order.
It is worth being precise about what is actually wrong here, because it is not that information is missing. Almost all of it is present. The problem is that none of it is sequenced, sourced, or connected to what was said at the time.
Which one is current? Two reports, three weeks apart, same parameters, different values. One is a repeat because the first looked off; one is from a different lab because the first was inconvenient. Nothing on either page says which is which. The patient does not remember. The order in the bag is the order they came out of a drawer.
Why was the change made? You stopped something in January. The reason was spoken aloud in the room — a symptom they mentioned, a report you were looking at, a cost they raised. If it was not written down in the moment, it is now gone, and the note in the register says what was done rather than why.
What is not in the bag? This is the dangerous one, because absence does not announce itself. The report they did not bring looks exactly like the report that was never taken. You cannot tell the difference from across a desk, and the follow-up you asked for in January either happened or did not.
Whose report is this? Families share a phone and a chat. A photograph of a lipid profile arrives at the clinic with no name visible in the frame, sent from an account belonging to a son who is managing his mother's appointments and his father's medication. The front desk files it under whoever they think it is.
None of these are memory failures on the patient's part. They are the ordinary consequence of a record that lives across a bag, a gallery, a register and a chat, with no single place that holds the trail.
The reports arrive at the clinic and stop there
Most clinics have quietly become a WhatsApp practice without deciding to.
Reports come to the clinic number because that is the number the patient has. Whoever is on the front desk that month receives them, and does something sensible — saves the file, writes it in the register, tells you when the patient arrives. That works, and it keeps working right up until the person changes.
Then it stops working in a specific way. The photographs are on a device. The register has entries but not documents. The message that arrived with the report — "doctor said to send this before Tuesday", or the sentence describing a symptom that made the report worth taking — was never part of the record at all, because there was nowhere for a sentence to go.
Two things follow that are worth saying plainly.
The first is practical: the clinic's memory of a patient is only as long as the tenure of its longest-serving front-desk staff member. That is usually much shorter than the patient relationship.
The second is about confidentiality, and it gets very little attention. Patient documents accumulating in the camera roll of a personal phone are patient documents on a personal phone. When that person's employment ends, nobody retrieves them, because nobody has a list of what is there. This is not a hypothetical risk in Indian clinical practice; it is the default arrangement in a large number of otherwise well-run clinics.
A question worth asking about your own clinic
Set aside software for a moment. Pick a patient you have seen more than three times over more than a year.
If you had to produce their complete trail — every report, every visit note, every message they sent, in date order, with each document attached to the visit it belonged to — how long would that take, and how much of it would depend on one person being available and remembering?
Most clinics can produce the register. Fewer can produce the documents. Almost none can produce the sentence that came with each one.
That last gap is the expensive one, because the sentence is usually where the clinical context is. The report is a number. The message that accompanied it is why the number was taken.
What is actually fixable
Not the bag. The patient will keep bringing the bag, and asking them to stop is the same mistake as asking them to log into a portal — it puts the burden on the person with the least reason to carry it.
Three things are fixable, and they are all about where the record lives rather than how the patient behaves.
The document attaches to the patient, not to a device. A report that arrives on the clinic's number should end up in that patient's record, not in a gallery that belongs to whoever answered. This single change is what makes the clinic's memory outlast its staffing.
The message stays with the document. Whatever was written when the file was sent is part of the record. It is one line, it costs nothing to keep, and it is the difference between a number and a finding.
The trail is answerable, not just stored. Storage is not the goal. Being able to ask "when did we last change the dosage, and what was in front of us when we did" — and get the note back with the note shown — is the goal. A record you have to scroll is a bag with better lighting.
Seven minutes is the constraint, and it will not move
Everything above is really one point. The consultation is short and is not going to get longer. Any arrangement that requires you to search during those seven minutes has already failed, and any arrangement that requires the patient to have organised their own history has failed before they walked in.
What can change is whether the clinic holds a trail that answers a question in the time it takes to ask it — including in eight months, when the person who filed it has moved on and the patient arrives with a fresh bag.
Kognora gives a clinic one record per patient — reports, messages and visit notes held together — and answers questions from that record with the source document attached, so an answer about a January visit arrives with the January note. Each patient's record is isolated from every other, data is not used to train anyone's model, and nothing is sent to a patient without a person reading the draft and pressing send. Kognora does not diagnose, treat or advise. It retrieves the clinic's own records and shows them with their source; clinical judgement stays with the doctor. The plain version of what it does and refuses to do is written out here, and the same problem in a different profession is in where did the client send that document?