Divergence over Thresholds
Conventional clinical alerting fires when a value crosses a population-level cutoff. Divergence-based detection fires when a signal departs from what was expected for this particular person on their particular trajectory — which is often much earlier, and much more meaningful.
The same blood pressure reading means different things depending on where a person has been. A threshold cannot see that; a trajectory can.
Consider someone with well-controlled type 2 diabetes, A1c steady around 7.0. A pharmacy fill gap appears, and their glucose pattern shifts over three days. No threshold has been crossed. Nothing would fire. But the trajectory has diverged, and a low-intensity check-in now — we noticed a change, is everything all right with your medication? — costs almost nothing and may prevent the episode that a threshold would eventually catch three months and one hospitalization later.
Divergence detection requires something thresholds do not: a maintained expectation of where this person was headed. That is what a Clock 1 representation is for.