I recently returned to clinical practice after time away. I’ll leave the setting unnamed — because after comparing notes with a few physician friends, I’ve concluded the name doesn’t matter. What I found is everywhere. It has been widely described. And at this point it is best understood not as an acute problem but as a chronic one.
Call it click creep.
Ambient scribing — the AI that now listens to a visit and drafts the note — was supposed to make documentation easier. In some ways it does. But it turns out you can automate the note and still drown in clicks. The two problems were never the same one.
Here is how click creep happens, and why it is so hard to stop: every individual addition is reasonable. A prompt to ask one more screening question. A field to capture one more quality measure. A “next best action” that, on its own, is genuinely smart. No single one is the problem. The problem is the accumulation. Add enough individually defensible things and you arrive at a collectively indefensible whole — an unfathomable volume of extra fields that no one designed on purpose and no one can quite account for.
And the accumulation compounds. Click fatigue sets in. Clinicians begin clicking through rather than clicking thoughtfully. Which means the data we are so busy collecting isn’t very good or very reliable in the first place. People develop workarounds — the smart defaults, the copy-forwards, the paths of least resistance. We build ever more elaborate machinery to capture information, and the information we capture gets worse.
As an organizational leader, I’ll make a confession. The follow-through rate on actually using much of the data we ask people to collect is wildly low. We add the field. We rarely revisit whether anyone looks at what it produces. Somewhere between the good intention and the annual report, the loop quietly fails to close.
And lest this sound like only the clinician’s burden — the same thing is happening to patients. A few days ago I saw my own primary care doctor. Before the visit, I was asked to complete five pages of questions, many of them new. We then had a good conversation, on the topics that actually mattered. Almost none of it touched the five pages.
So how do we move forward?
I want to propose something deliberately blunt. A great purge. Call it the grand declicking.
But a purge is only a slogan unless you understand why the clutter accumulated in the first place — and click creep is not an accident. It is a ratchet. Every field in your record was added by someone with a plausible reason. Almost none were added by someone who also owned the job of taking one away. Addition is easy, and its cost is diffuse — spread in three-second increments across thousands of clinicians who never chose it. Deletion is hard, because it requires a specific, named person to stand up and say: *we are safe without this.* In most organizations, no one holds that job. So the ratchet turns only one way.
The grand declicking, then, is not mainly an act of deletion. It is an act of governance. It means giving someone the authority to remove, the way we have long given everyone the authority to add. It means treating every field as a cost until proven otherwise — default off, not default on. It means asking each existing question to justify its continued existence, and retiring the ones that cannot. And it means one simple, uncomfortable rule: if we are not using the data, we do not collect the data.
None of this is a technology problem. Ambient AI will keep getting better, and it should. But better tools laid over bad governance simply produce more sophisticated clutter. The system does not decide what to ask. People do. Which means people can decide to stop.
I think often, in the exam room, about what the click is for. Every field I complete that no one will ever read is a small theft — a few seconds taken from the one thing the visit was supposed to be about, which is the person in front of me. Multiply that theft across a career, across a workforce, across a country, and you get a health care system that documents more and attends less.
We did not get here through one bad decision. We got here through ten thousand reasonable ones. The grand declicking is simply the discipline to reverse a few thousand of them — and the humility to admit that collecting less might be how we finally come to know more.
