The behavioral health reimbursement system is broken for Rogers and broken for payors. Rogers sees roughly 20 of 230 Molina Medicaid eating-disorder members. The other 210 cycle through acute settings at enormous cost, with uneven outcomes and no measurement spine. The architecture below changes that math.
FHG's role is honest broker and measurement anchor. We don't have a preferred clinical lane, a competing network, or a payor contract to protect. What we have is the reimbursement architecture, the measurement platform, and the relationships to wire the pieces together.
Three-Part Payment Architecture
Existing FFS revenue cycle stays intact. No disruption to current billing, credentialing, or claims workflow. This is additive, not a replacement.
In advance — monthly or fully upfront. More than 2× the FFS episode rate. Payment flows before services are rendered, ending the prior-auth friction that governs the status quo.
On top of Parts 1 and 2. Target: 4–5× revenue per patient compared to FFS baseline. Savings are split against a pre-agreed medical-loss-ratio floor — no black box, no post-hoc negotiation.
F50 Economics — Eating Disorder as the First Lane
The growth path is 20 today to 60–100 — a Rogers capacity question, not a demand question. The capacity is ramping: Rogers opened two additional residential facilities at the Oconomowoc campus dedicated to ED and OCD in July 2025. The architecture converts payor waste into Rogers revenue and better patient outcomes simultaneously.
How the Pieces Wire Together
Solventum CRG is the scoreboard — real-time cohort reporting to both Rogers and Molina simultaneously, with no black box and no total-cost-of-care bleed into out-of-lane conditions.
Solventum CRG (formerly 3M, $42B in managed-care premiums prospectively risk-adjusted) provides the measurement rails. What makes it the right scoreboard: it was designed for behavioral-health value-based payment, it is publicly traded, and it carries real-time Cohort reporting to both sides at once. Rogers sees exactly what Molina sees. No surprises, no interpretation disputes after the fact.
Limbic AI is already live at Rogers (deployed December 2024) — 92% positive response rate, 3× admit rate, no headcount addition. It is the front door that turns the measurement architecture into a real-time patient-routing system.
Boundaries and Controls
The architecture works because it respects Rogers' operating constraints rather than trying to override them. Every structural decision flows from these four boundaries:
- Narrow network, not all-comers. Explicit inclusion and exclusion criteria on every Cohort. Rogers is paid for what it treats, not for volume targets it cannot responsibly meet.
- Full Rogers UM/UR control. Molina's denial machinery turns off for this population. Rogers sets the utilization and clinical criteria. Non-negotiable per the Milwaukee Day conversation.
- Metric-gated, dollar-backed. Six-month pilot with named achievements, ROI benchmarks, and real dollars. Phased: pilot → proof → expansion. Budgeted into next year's plan before signature.
- Shared savings on PMPM / MLR only. Out-of-lane conditions (conditions outside the named clinical population) do not hit Rogers' reconciliation. Rogers is not in the total-cost-of-care business.
Pilot Architecture
Concierge Pilot — ~100 Patients, Two Named Humans
Why Eating Disorders, Why Now, Why Rogers
Solventum's CRG architecture is the emerging national framework for behavioral-health reimbursement — designed to allow plans to step out of clinical pathway-setting while maintaining actuarial accountability. The window to become the standard-of-care rails that plans default to in behavioral health is approximately 18 months.
Rogers is positioned to take that lane because of three things that are genuinely rare in combination: the only IP eating-disorder program in the Midwest, a measurement spine already in place (PopulationManager + Cerner + Limbic AI), and executive leadership that understands the difference between a wish and a business case.