Industry: Pharmaceutical Product: ModelRisk Application: Drug pricing strategies
The list price on a 30-day pack of a new chronic-therapy oral is set at $1,200. After commercial rebates, Medicare Part D coverage-gap mechanics, Medicaid Best Price + AMP rebates, 340B ceiling pricing, and patient-assistance write-downs, the blended net realised price is approximately $562 — a 47% gross-to-net ratio, with a 10%/90% range of $432 to $689. Every dollar of list-price discussion at the launch committee was a 47-cent discussion of revenue.
A multinational launching its first US specialty product rebuilt the price-to-net realisation in ModelRisk and added a parallel European HTA model to find the launch price at which NICE acceptance probability crossed 80%. The two together produced the global pricing committee's first defensible cross-market launch price — one priced in distributions, not in single rebate-percentage assumptions.
The blended net mean is $562 on the $1,200 list — a GTN of 47%. The P10/P90 band spans $432 to $689, a $258 range that is 21% of list price. A finance model using a single point GTN ratio understates the operational planning risk by roughly that 21% in either direction — which is why the list-price decision can only be read against the distribution that produces net, not against list itself.
US gross-to-net is not one number but a weighted average across distinct payer channels, each with its own rebate distribution:
Drawing per-channel rebates jointly with a Dirichlet on channel shares (because the mix itself moves with payer coverage decisions) gives the blended GTN distribution. The Beta distributions are chosen over plain Normal because rebate percentages are bounded on [0, 1] and historically left-skewed — Normal would generate impossible negative rebates and miss the long left tail toward higher rebates that drives the worst-case net price.
Channel-by-channel:
Medicaid nets about $327 per pack and 340B about $400. Commercial nets at $720. The blended net is dominated by the commercial and Medicare channels (73% of mix combined) but constrained by the regulated Medicaid + 340B tail. A 2-percentage-point shift in channel mix toward Medicaid moves blended net by approximately $8 — small per pack, but not nothing for a multi-hundred-million-dollar revenue line.
The Medicaid rebate percentage swings net by ±$55, the largest single driver — and the one farthest outside the manufacturer's control. Commercial-rebate negotiations are second; Medicare Part D rebates third. Channel-mix shift to 340B is fourth and trending upward year-on-year as covered entities expand. The pricing committee's leverage is concentrated in commercial-channel negotiation strategy and in the timing of list-price escalators against contractual rebate ceilings.
The European launch decision is structurally different. Net realisation is largely set by national reference pricing once a price is approved, so the binding constraint is HTA acceptance probability. For UK NICE submission, the incremental cost-effectiveness ratio at threshold:
\[ \mathrm{ICER} \;=\; \frac{C_\text{new} - C_\text{comparator}}{\mathrm{QALY}_\text{new} - \mathrm{QALY}_\text{comparator}} \;<\; \pounds 30{,}000/\mathrm{QALY} \]
With comparator cost £5,200/yr and QALY gain ~ Triangular(0.18, 0.32, 0.48), the simulation produces the acceptance frontier:
The price at which P(NICE acceptance) crosses 80% is approximately £13,300/year at the standard £30k/QALY threshold — well below what the indication's commercial team had assumed achievable. At the £50k/QALY end-of-life threshold the same probability shifts to roughly £18,700/year, but the indication does not qualify for that threshold under current criteria. The deterministic ICER calculation with QALY-gain point estimate of 0.32 would have said £14,800 — close to the simulation answer, but with no information about how much acceptance probability tightened around that price.
A $1,200 list price is a $562 net price with a $99 standard deviation and a $258 P10–P90 band. Treating list as revenue, or treating GTN as a single number, miscalibrates every downstream forecast that depends on it. The simulation in ModelRisk replaces a deterministic GTN ratio with a distribution shaped by the channel-rebate structure that actually produces it — and the pricing decisions that emerge from that distribution are the ones the launch committee can defend at the first contract review.