hh-health-ai/healthcare-equity/biosimilar-erosion

Biosimilar erosion curve

Turn a biologic's exclusivity position and competitive set into a dated erosion curve that a revenue model can consume.

Workflow

1. Establish the reference product's position from the Purple Book

(scripts/purple_book.py): BLA number, first licensure date, reference-product exclusivity expiry, and any listed biosimilars with their licensure dates and interchangeability status.

1. Count and qualify the entrants. Not all approvals launch. Check for: launch

announcements, settlement-driven launch dates, manufacturing capacity, and whether the entrant is a large player with contracting muscle or a small one that will discount without winning share.

1. Determine the substitution mechanism, which drives everything:

• Interchangeable, pharmacy-dispensed → fastest erosion.
• Not interchangeable, physician-administered, buy-and-bill → slowest; erosion

depends on ASP dynamics and the physician's reimbursement spread, not on the pharmacy.

• Payer-forced switching → step-shaped erosion at formulary dates (1 January is the

usual step), not a smooth curve.

1. Model price and volume separately. Biosimilar competition is initially price-led:

the reference product often holds volume while conceding net price through rebates. A model that assumes volume loss will misread the first two years and then be surprised by the cliff when contracts finally flip.

1. Build the curve. Quarterly, from the entry date, with the shape justified by the

substitution mechanism and by two named analogue molecules with similar mechanics. Analogues do more work here than theory.

1. Watch the defence. Reference-product sponsors deploy: authorised biosimilars,

next-generation formulations (higher-concentration, subcutaneous, longer-acting), device improvements, and aggressive contracting that trades price for volume retention. Each changes the curve; name which ones are in play.

1. Instrument the curve. Once entry happens, replace assumption with observation

via rx-utilization → sdud-trx-proxy and Medicare Part B/D spending dashboards. A curve that is not being checked against data quarterly is a guess with a chart.

1. Emit the brief.

Not-automatic

A biosimilar approval does not license an erosion assumption. Approval without launch, without interchangeability, or without contracting capacity produces very little erosion, and several approved biosimilars have never meaningfully launched.

Reference: references/erosion-analogues.md. Contract: ../../references/evidence-brief.md.