The Noise Floor Problem
I’ve spent years designing clinical systems to solve alert fatigue. I built the exact same failure on my wrist.
I’ve spent years designing clinical products to solve alert fatigue. I built the exact same problem on my wrist without noticing.
Alert fatigue has a precise mechanism. It isn’t about bad alerts. It’s about volume exceeding threshold. When the noise floor rises high enough, the brain stops processing individual signals and starts treating the entire channel as background. Physicians don’t dismiss drug-drug interaction warnings because they don’t care. They dismiss them because the warning fires forty times before it fires once about something that matters. The signal degrades. The receiver adapts.
In clinical product design, the fix isn’t better alerts. It’s architectural. Reduce the noise floor. Tighten specificity. Route signals to the receiver who can actually act on them. The goal is a system where an alert interrupts because it’s rare. Rarity restores meaning.
I’ve understood this for years. I’ve applied it in EHR implementations, clinical decision support builds, and the pre-visit and post-visit workflow architecture we run at General Medicine. It did not occur to me until recently that I’d failed to apply it to the six-inch square around my wrist.
The Case Study Is My Apple Watch
I’ve been an Apple loyalist for a long time. Mac for clinical work, iPad for NerdMDs and advisory projects, iPhone Air as the daily driver. The ecosystem earns it. The Apple Watch felt like the natural extension of that stack.
What actually happened: I spent two years slowly disabling it.
App notifications went first. Then stand reminders. Then mindfulness prompts. Then I filtered contacts, built custom Do Not Disturb windows, and created exceptions for exceptions. Every time a buzz irritated me, I went into settings and removed another category. I was doing manual alert threshold management — reactively, iteratively — exactly the way a poorly configured EHR forces clinicians to manage their own noise floor.
By the time I finished, the watch was a step counter with a heart rate sensor and a nightly charging requirement.
The hardware was never the problem. The architecture was. I’d attached a high-volume notification surface to a stack already generating maximum ambient noise — three Apple devices, all talking to each other, all surfacing alerts constantly. The watch didn’t extend the ecosystem. It added a fourth alert channel with no triage layer between the input and my attention.
That’s not an Apple Watch problem. That’s a system design problem I created and then spent two years patching at the settings level.
Designing Out of It
On our last cruise, the muster check-in station happened to be inside the ship’s watch shop. Not planned. Just where the line ended up. A particular watch caught my attention — good weight, clean dial, the kind of thing that works dressed up or on a working dock. I was looking at it when the salesperson mentioned, almost in passing, that the captain of the ship wears that exact watch.
We bought it. The perk that came with the purchase was a tour of the bridge for the whole family.
Standing on that bridge, I got to see the captain in his actual operating environment. The watch made sense there. Analog. Clean dial. Nothing initiating contact. One job.
The Oura Ring I’m switching to operates on the same principle. No screen. No notification surface. It collects sleep staging, HRV, resting heart rate, temperature drift, and readiness data and feeds all of it silently into Apple Health. The device never initiates. You pull the data when you want it. The ring has no opinion about when that should be.
That’s a different sensing architecture. Passive collection, user-driven retrieval. The channel stays quiet because nothing is pushing through it.
I made a similar observation a few weeks ago writing about Eight Sleep — The Signal Is In Your Bed. The Care Pathway Isn’t. Ambient passive sensing is structurally different from a device that pushes. Eight Sleep monitors your sleep whether you’re thinking about your health that night or not. Oura does the same thing on your finger. Neither one asks for your attention. The data accumulates until you choose to check.
That’s the architecture I want on my body. Not more screens. A quieter channel with a cleaner signal.
Sizing kit is on its way. Full report once I’ve worn it through a real week.
The Apple Watch Doesn’t Get Retired
It gets a job description.
There are aluminum enclosures on AliExpress that convert an Apple Watch into what looks like a classic color-screen iPod — mechanical click scroll wheel included. I ordered one. When it arrives, I’ll strip the bands, drop the watch into the enclosure, and use it as a purpose-built dumb phone. Small enough to leave the house without the iPhone Air. Still capable of receiving a message or taking a call. Bluetooth for AirPods. No app ecosystem running in the background, no ambient notification pressure.
One device. One job. I’ll document the full conversion when the shipment clears.
The principle that took me years to operationalize in clinical product design applies everywhere: you cannot solve an alert fatigue problem at the settings level. You solve it at the architecture level.
The EHR version is choosing not to surface a low-specificity alert at all rather than making it easier to click through. The wearable version is choosing a ring with no screen over a watch with a configurable notification profile. The outcome you’re designing for isn’t a quieter alert. It’s a channel that stays quiet until the signal is worth the interruption.
I’ve been building that principle into clinical systems for years. It’s obvious in retrospect that the same logic applies to every device touching my body.
Subscribe for the Oura first-week report and the full AliExpress iPod conversion. Both coming as the hardware arrives.
If you’re managing alert fatigue — in your EHR, your wearable stack, or somewhere else entirely — drop it in the comments. Curious where others are drawing the line between signal and noise.
i have no Apple fatigue problem. Mostly cos i dint wear a watch....