Prior authorization decides whether a therapy starts. Step therapy decides the order. Both do their job at the front door. But after the approval clears, a high-cost chronic therapy can run for years — escalating in dose, accumulating refills, drifting past the clinical picture that justified it — and nothing decides whether it should still be running.
That is the gap. It is structural, not occasional — every population that has looked has found it. The diagram below is the whole thesis: the front door is guarded, and the long corridor after it is not.
Monitoring and flag-rate figures: the DURA study, drawn from the NIH All of Us Research Program cohort (n=30,734). Spend anchor is the claims-weighted average therapy cost in the commercial-payer cohort.
Cadence runs a defined cohort — minimum 2,500 members, scoped by therapy class, never by cost. Seven clinical triggers fire on explicit criteria. No random sampling, no cost-threshold cherry-picking. The triggers produce the flag rate: in the self-funded employer cohort, 26.4% of the book met at least one.
The seven triggers are Section 2 of the Cadence Governance Standard v1.1, which is published in full and open to inspection.
Every flagged case is read by externally credentialed PharmD or MD reviewers — at least two per cohort, with no payer employees in the review chain. Each reviewed case returns exactly one of four advisory outcomes. The finding is advisory; no prescription changes without prescriber action.
A structured second look found grounds to act — dose, taper, or switch — in roughly 60% of reviewed cases across the studied cohorts. The rest were confirmed appropriate to continue, which is itself a documented governance result.
Pipeline figures are the self-funded employer cohort (9,500 members). Reviewer Influence Rate (RIR) = (Adjust + Taper + Switch) ÷ Reviewed. Outcome mix is illustrative of the reviewed population.
Generated Savings Value (GSV) is computed from the reviewed outcomes with fixed Therapy Adjustment Factors — Adjust 0.25, Taper 0.50, Switch 0.30 — applied to the claims-weighted average therapy cost. The formula is open. Nothing is taken on faith.
Cadence labels every figure by confidence. Cohort findings — flag rates, review counts, influence rates — are Measured. The dollar value is Derived from the published formula. The color tells you which is which everywhere on the site.
Exact GSV values: $8,640,430 (Cohort 1) · $5,611,568 (Cohort 2) · $14,251,998 combined. The DURA cohort (30,734 patients) is flag-rate evidence and is not included in GSV. The evidence base is 65,234 patients across all three cohorts.
At cycle close, the complete governance artifact is sealed. No record, outcome, or trigger configuration can be altered afterward. The cycle ran under a versioned configuration fingerprint locked before the first review — so the method that produced the result is fixed and inspectable. The audit trail is the product.
And the gap does not close itself. When the same population was governed a second time, the persistence rate was 50% — the drift keeps happening because nothing upstream changed. Continuation governance is a recurring cycle, not a one-time cleanup.
Cycle requirements and the seal are Sections 6 through 8 of the Cadence Governance Standard v1.1. Governance persistence is measured cycle-over-cycle on the same population.
A defined cohort, looked at again, by independent reviewers, against a published standard — and a sealed record that says what was found.
That is what a buyer gets: not a denial engine, not another prior-auth gate, but the first instrument that governs continuation. The standard is open, every number carries its confidence tier, and you can model your own book before you commit it.