Throughput
When a system becomes faster, what else becomes thinner?
Roscoe is asked to do something simple: explain how a national cell-therapy program became dramatically faster so that other programs can copy it. What he finds instead is a chain of small, local, perfectly reasonable decisions that have removed weeks of waiting from a life-critical pathway. Patients are treated sooner. Bottlenecks disappear. No single person can claim the redesign, because no single person made it.
As Roscoe follows the holds that used to slow the program, the pattern becomes harder to dismiss. This is not sabotage. It is not negligence. It is success that exceeds the understanding of the people who produced it. Every change can be defended on its own terms. Taken together, they produce a result bigger than authorship.
Throughput turns the series toward medicine, patient time, and the uneasy truth that pauses, safeguards, and human friction may be the very things keeping efficiency from becoming something colder. The question is no longer whether optimization is happening. It is what remains when a system learns to treat delay as waste.