The External Hard Drive: Why Doctors Still Write on Their Hands

Walk onto any acute ward during morning rounds and you will see the same thing. A resident with a telephone wedged between ear and shoulder, copying a potassium value onto a scrap of paper, or onto the back of their own hand. We tend to read this as carelessness: doctors with bad handwriting who cannot be bothered with the record. The truth is closer to the opposite. They are writing because they are afraid of forgetting.

When one person is carrying twenty complex patients across a shift, the brain runs out of working memory. Writing things down is not sloppiness; it is a cognitive necessity. The note becomes an external hard drive, the thing that keeps patient A's potassium trend from bleeding into patient B's arrhythmia. It forces a little structure onto a great deal of chaos.

Which means the problem was never that doctors have to document. The problem is that the tool we gave them is so cumbersome that a ballpoint pen on skin is still the faster instrument for thinking. We built a record they have to escape in order to do their jobs.

Writing to think, writing to defend

We should be honest about what happens when that doctor finally sits at a terminal to transcribe the notes from their hand. The purpose quietly changes. They stop writing to reason and start writing to be defensible — for the audit, for the Medizinischer Dienst, for the colleague who may one day ask why a decision was made.

Most progress notes are now written for a reviewer years in the future rather than for the colleague taking over the patient in five hours. To build that defensive layer, the physician pads the record with redundancy: pasting in lab values that are already stored, just to prove they were seen, adding the same boilerplate to every paragraph. In the flood of compliance text, the one thing that mattered — the clinical reasoning — is drowned. We have made the record legally robust and clinically unreadable.

The machine takes dictation; the doctor edits

The way out is an unsentimental division of labour. A clinical note is doing two jobs at once, and only one of them needs a human being. The first is commodity work: vital signs, medication lists, the standard safety checks. No clinician should be transcribing this by hand. A system can capture it in the background — through ambient capture or integration with the existing record — and keep the audit trail intact without ever touching the doctor's attention.

The second job is the one that cannot be delegated: synthesis. “I am holding the beta-blocker because the patient is symptomatic, despite reaching the rate-control target.” That sentence is the reason the doctor is in the building. Freed from data entry, the note returns to its original purpose — a place to think, not a place to file.

A quality mechanism, not a convenience

It helps to drop the idea that automated documentation is a comfort for lazy doctors. When a physician writes on their hand, they are trying to stay safe. When they paste junk into the record, they are trying to be defensible. Both are rational responses to a system that has confused logging the facts with reasoning about them.

So the task for anyone who owns the technology is narrower than it sounds. Automate the bureaucracy, so that the external hard drive in a clinician's head can be spent on solving the case rather than on remembering to record it. The point is not to make doctors write less. It is to give them back the part of the record that was ever worth writing.