The Conveyor-Belt Fallacy: When Velocity Destroys Value

Walk into almost any hospital management meeting and the screen on the wall will be measuring time. Door-to-doctor minutes. Average length of stay. Throughput per bed. We run the place by the clock, on a quiet assumption borrowed from the factory floor: that if we can move the unit faster through the process, we have made it better. The patient is the unit. The pathway is the process. Efficiency is velocity.

So we buy technology to act as a lubricant. We want the admission that takes one click instead of ten, the order set that populates itself, the pathway with every grain of latency sanded out of it. The pitch is always the same, and it is seductive: we will remove the friction between the symptom and the prescription.

But a hospital is not a factory. On an assembly line, a pause is a defect to be designed out. In evidence-based medicine, the pause is the product. The moment a clinician stops, looks at the trend, and asks whether the standard answer fits this particular patient is not dead time in the process. It is the thing the patient is paying for.

Reflex instead of reflection

When you smooth a pathway far enough, you change what the clinician is actually doing. The pre-filled bundle for community-acquired pneumonia appears, and the button to accept it is large and inviting. The route to review the underlying evidence or adjust the regimen is three sub-menus deep. The interface has made it effortless to do the standard thing and laborious to think about whether the standard thing applies.

The result is a quiet drift from interrogation to confirmation. The clinician stops asking whether the protocol is right for the person in front of them and starts accepting it the way the rest of us accept terms of service: by clicking through without reading. We have not made the decision better. We have made it faster, and then mistaken the speed for quality.

The bill arrives later

Systems thinking has a name for what this produces: failure demand. It is the work created by not doing something properly the first time, the demand a system manufactures for itself. Frictionless interfaces are very good at generating it, because the savings are visible immediately and the costs arrive weeks later, on a different ledger.

Consider the trade in concrete terms. A physician uses a one-click macro to start a broad-spectrum antibiotic, saving perhaps three minutes on a crowded round. The contraindication that the pause would have surfaced goes unseen. A fortnight later the patient is back with a Clostridioides difficile infection and stays an extra week. We banked three minutes of clinical time and spent it many times over in bed-days, suffering and cost. The transaction was optimised. The outcome was not.

Friction worth keeping

The uncomfortable conclusion is that not all friction is waste, and that good design has to tell the two apart. The friction of a slow login, a buried menu, a demographic form retyped for the third time is pure bureaucracy and should be hunted down without mercy. But the friction of a system that calculates a dose and then declines to activate the sign button until the prescriber has looked at the renal-function trend is something else entirely. It is a safety feature wearing the costume of an inconvenience.

This is the audit worth running on every workflow and every tool a vendor brings to the table: where has the design preserved thinking space, and where has it quietly removed it? When a system promises to cut diagnostic time in half, the figure tells you nothing on its own. Ask how it did it. Did it surface the relevant data faster, so the same judgement happens sooner? Or did it shave off the minutes by making it awkward to second-guess the model?

The two look identical on the dashboard and could hardly be more different at the bedside. A pizza can be ordered without a second thought; a course of chemotherapy cannot. Where the design removes the pause, what looks like a faster hospital is only a hospital that has learned to make its mistakes more efficiently.