Probabilistic prevention
Prevention used to wait for certainty. Now systems act on the odds, stepping in before a problem is confirmed — and where you set the threshold quietly decides who gets help and who waits.
Change driver · Updated July 2026
The shift ahead
From proof to probability
For most of history, action waited on proof. The change is that systems no longer wait — they move on the likelihood and sort out the certainty later.
A tax authority lets a fraud-risk score decide who gets investigated and who gets cut off. A railway acts on the probability that a part is about to fail and replaces it before it does. A medical panel redraws the line between being at risk and being sick, and who gets treated moves with it. In each case, a threshold no one voted on is deciding who gets acted on.
The shift is not earlier detection alone. It is the movement of action into probabilistic territory — where a likelihood, a trajectory or a risk threshold is treated as reason enough to act, long before proof arrives.
Why it matters
Acting on proof means you are rarely wrong. Acting on odds means you will sometimes be wrong on purpose.
But being wrong is harder to pin down than it sounds. When prevention works, the thing you prevented never happens — so you rarely see the disaster you averted, and almost never see the person you acted on who would have been fine anyway. The error rate is partly unknowable, which means the real argument was never about accuracy. It is about where the line is drawn.
And someone has to draw it. Set the threshold to act at a 50% risk and the person sitting at 48% gets nothing — no treatment, no coverage, no attention — until the day the risk becomes a diagnosis. Set it lower and you commit to acting on many people who were never going to need it. Each threshold prices the cost of acting against the cost of waiting, and decides who is worth moving early for. The unsettled question is who sets that number, and whether the person just below it ever gets a say.
Possible futures this could enable
-
01
Prevention moves upstream of the problem
Systems act before anything breaks or shows — and when it works, the crisis simply never arrives, which is exactly why the win is so easy to miss.
Early signalEurope’s big rail operators, Deutsche Bahn and SNCF among them, now run predictive maintenance that reads wear in real time and fixes parts before they fail — the International Union of Railways reports it cuts equipment failures by roughly 30%.
-
02
The line between at-risk and sick keeps moving
Where the boundary between a risk and a diagnosis is drawn decides who counts as sick — and who qualifies for treatment because of it.
Early signalIn January 2025 a Lancet commission split obesity into ‘pre-clinical’ and ‘clinical’ stages — redrawing the line between being at risk and being sick, a line that in many systems decides who qualifies for care.
-
03
The threshold decides who gets acted on
Everything hinges on where the line is drawn — and when it is drawn on the wrong inputs, with no way to appeal, the people on the wrong side of it pay.
Early signalThe Dutch tax authority let a fraud-risk algorithm decide who to pursue, and it wrongly branded roughly 26,000 families as fraudsters — using traits like nationality as risk factors, clawing back tens of thousands of euros and helping force the government to resign in 2021.
Right now, the line between acting on proof and acting on odds is being crossed in dozens of places at once, rarely with any announcement.
Some of this is mature and well-governed: credit decisions, fraud checks, industrial maintenance, established screening programs. Some of it is racing ahead of any agreement about where the thresholds should sit, who should set them and whether anyone affected can push back. The power to score almost anything now exists. The rules for acting on those scores do not.
The line that matters is the line between prediction that serves the person and prediction that is done to them. The stronger version acts early with a threshold set in the open, a way to contest the call and someone accountable for both kinds of error — acting when it should not have, and failing to act when it should. The weaker version treats a number as a verdict, hides who chose it and leaves the person on the wrong side of the line with nowhere to turn.
Watch where the threshold is set, and who gets to set it.
The driver strengthens as scores start triggering consequences that used to require proof — money withheld, access denied, treatment begun, a person investigated. It strengthens each time the cost of waiting for certainty is judged higher than the cost of acting too soon, and someone has to decide, in advance, where that balance tips.
The question is not whether to act on the odds. Often we should. The question is who draws the threshold — who prices acting against waiting — and whether the person just under the line is ever part of that decision.
This is one force among many.
We track the ones that will reshape your field, and what to do about them.