hh-health-ai/healthcare-equity/gross-to-net-bridge

Gross-to-net bridge

Assemble a defensible net-price path for a drug from open CMS sources, and mark clearly where the estimate stops being data and starts being assumption.

Workflow

1. Anchor list price. Pull WAC from the company's own disclosures or a public

price-transparency filing. Record the date of the last list increase — most large caps now take one modest increase per January.

1. Build the observable net floors.

• Medicaid: SDUD total reimbursed ÷ units gives a pre-rebate reimbursement level;

the statutory Medicaid rebate (23.1% of AMP for most brands, plus the CPI penalty) sets a hard floor well below it.

• Medicare: the CMS Medicare Part B and Part D Spending by Drug dashboards publish

average spending per dosage unit and annual change — the cleanest open series for a Medicare-weighted net trend.

• 340B: HRSA covered-entity growth in the relevant setting is a structural

GtN headwind. Size it qualitatively; the discount itself is not public per drug.

1. Layer the policy calendar. IRA maximum fair price effective dates, Part D

redesign (manufacturer liability in the catastrophic phase), and any inflation rebate exposure where list increases outran CPI. Pull dates from a reimbursement engine rather than re-deriving them here.

1. Bridge it. Gross → statutory rebates → commercial rebates and PBM fees →

340B and other mandated discounts → patient assistance and copay support → net. Only the statutory and Medicare components are observable. Label the commercial rebate line as the assumption it is, and show the net-revenue sensitivity to it at ±5 percentage points.

1. Sanity-check against reported results. Company-reported net revenue ÷ your

estimated volume gives an implied net price. If that implies a GtN outside reasonable bounds for the class, your volume estimate is wrong, not the company's.

1. Emit the brief, with the commercial-rebate assumption named in Not-automatic.

Class heuristics

GtN gaps are widest where PBM competition is fiercest — insulins, respiratory inhalers, TNF inhibitors, and increasingly the GLP-1 class. They are narrowest in buy-and-bill oncology and single-source rare disease. A model that applies one house GtN percentage across a diversified pharma's portfolio will misprice the mix shift, which is usually the actual driver of net-price surprise.

Not-automatic

An observable Medicaid or Medicare net level does not license a commercial net price. The commercial book is where the rebate war happens and it is not public.

Framework anchors: your own valuation framework; Kongstvedt on PBM and rebate mechanics; MedPAC and KFF for policy context. Contract: ../../references/evidence-brief.md.