Negotiation Outcomes Calculator

This model estimates the effects of altering the Medicare drug price negotiation provision in the Inflation Reduction Act (IRA). Impacts of policy changes are estimated for 2026-2030 and include the following measures:

• Federal savings from negotiation

• The number of drugs that would be selected for Medicare price negotiation each year:

  • In total

  • That are small molecules

  • Having an orphan indication

The model is designed to allow users to compare results from its base case (negotiation under current law) with different scenarios based on the changes they select from the drop-down menu options.

Users can interact with the model by changing the parameters for Medicare drug price negotiations, including:

• The number of drugs to be selected for negotiation

• The number of years before a drug can be negotiated

• Requirements for exempting from negotiation drugs with orphan indications

Methods

The model’s base case follows the provisions of the IRA and their interpretation by CMS under current law. It projects the characteristics of drugs selected for negotiation, their negotiated prices (maximum fair prices, or MFPs), and federal spending with and without these MFPs in place.

The base case (rules under current law) reflected in the model is as follows:

1. Selection

Each year, drugs qualify for selection if they meet the following criteria, beginning in 2024:

• Having at least $200 million in gross spending in Medicare Parts B or D

• Active ingredient or moiety approved by FDA at least 7 years ago if a small molecule or 11 years ago if a biologic

• Do not have an approved and marketed generic or biosimilar

Drugs are excluded if they meet criteria for categories including:

• Pure orphan drug with a single designation: having one orphan drug designation and indications only within this designation

• Small biotech drug (2026-2028 only):

  • a drug that accounted for less than 1% of Medicare Part B or D spending in 2021, and

  • that accounts for more than 80% of Part B or D expenditures attributable to its manufacturer

Drugs that meet these criteria are sorted according to their gross spending in Medicare. Drugs at the top of the list are then selected for negotiation, with negotiated prices going into effect according to the following schedule:

• 2026: first 10 drugs (Part D only, already announced)

• 2027: 15 drugs (Part D only, announced on Feb 1, 2025)

• 2028: 15 drugs (Part B and Part D)

• 2029 and onwards: 20 drugs each year (Part B and Part D)

2. Negotiated Prices

For each selected drug, the model estimates the statutory offer ceiling (the maximum price Medicare is allowed to offer to pay under current law) for the negotiation process according to their time on the market once negotiated prices go into effect. This amount is the lower of:

• The weighted average Part D net price of a drug, and a fixed discount based on the time the drug has spent on the market:

  • Short monopoly drugs (fewer than 12 years): 75% of non-federal average manufacturer price (NFAMP)

  • Long monopoly drugs (at least 16 years): 40% of NFAMP

  • Extended monopoly drugs (at least 12 years and fewer than 16 years; first goes into effect in 2030, negotiated in 2028): 65% of NFAMP

Then, the model estimates the Maximum Fair Price (MFP) that is likely to result from the negotiation process using the lowest of the following:

• The offer ceiling

• An additional discount of 10% relative to net price if the statutory offer ceiling is set by the drug’s net price

3. Application of Negotiated Prices Through Generic / Biosimilar Entry

The model assumes that a selected drug will remain subject to its negotiated price for 3 years. By 3 years, the model assumes that a generic or biosimilar of the drug will enter the market, disqualifying the drug negotiated prices under current law.

4. Estimated Federal Spending and Savings from Negotiation

The model projects gross and net federal spending on negotiation eligible drugs through 2030. For drugs selected for negotiation, the model estimates the difference between net Medicare spending with and without negotiation, with the difference accruing each year as federal savings.

Model Data and Parameters

The model relies on parameters describing key characteristics of drugs eligible for negotiation, as well as their spending and spending growth projections. These elements and their sources are summarized in Table 1.

Drugs included in the model

The model relies on a proprietary dataset of 645 negotiation-eligible branded drugs that combines data from the CMS spending dashboard with information from Evaluate Pharma, Drugs@FDA and other sources listed in Table 1. Drugs are excluded from this dataset if at least one generic or biosimilar competitor is already on the market for at least one of the drug’s formulations, which disqualifies the branded drug from negotiation under the IRA.

In addition, products are excluded if they are marketed under Emergency Use Authorization (EUA), are plasma-derived blood products, or no longer marketed in the U.S., as these products would not meet the criteria to be selected for negotiation.

Different formulations of the same active ingredients or moieties produced by the same manufacturer (or authorized generic manufacturer) are considered one drug. To establish whether a drug qualifies for negotiation, its time on the market is calculated based on the earliest FDA approval date for the active ingredients or moieties that it contains.

Drug classifications

Small molecule vs. biologic

Drugs included in the model were classified as small molecule or biologic based on whether they were approved by FDA under a New Drug Application (NDA) or Biologic Licensing Application (BLA).

Small biotechnology

Certain drugs may be excluded from the Medicare negotiation under current law if their manufacturers qualify as a small biotechnology company. Drugs that represented at least 80% of their manufacturer’s total revenue from Part B or D in 2021 and that did not constitute more than 1% of Part B or D spending are considered to qualify for this exclusion.

Orphan indications

Drugs in the dataset are categorized as single-designation pure orphan drugs if they have one orphan designation and all indications are within that designation. Drugs with multiple orphan designations and all of whose indications were within those orphan designations are considered multiple-designation pure orphan drugs. In addition, drugs are classified on whether they have any orphan indication.

Medicare spending

Data for the latest available year of Medicare drug spending in Part D and Part B, 2022, was obtained from the CMS Drug spending dashboard. To estimate future drug spending, 2022 spending data was projected through 2030 using compound annual growth rates (CAGRs) estimated from revenue projections in Evaluate Pharma, where available, for each drug. Drugs with no revenue projections in Evaluate were assumed to maintain current levels of spending, with no additional growth.

FAQ

Why are model results different from those of the Congressional Budget Office (CBO)?

CBO initially estimated that the federal government would save $99 billion through 2031. The base case of our model estimates savings of $83 billion through 2030.

While the estimates are similar in magnitude, there are important differences between them. CBO benchmarks its analysis to baseline spending estimates. This “top-down” approach means that other trends may factor into overall results. Our estimates come from a “bottom-up” approach that aggregates results across individual drugs used by Medicare beneficiaries, but doesn’t take into consideration interactions with other programs and budgets.

Another difference is that our model doesn’t consider biosimilar and generic entry before selection, and assumes that all drugs exit the negotiation program after 3 years due to generic or biosimilar competition. When we alter the model to simulate random entry of generics or biosimilars before or after selection based on published data, estimates of federal savings decline by about 30% (read more about this below in the FAQ labeled "How does the model handle generic and biosimilar entry?"). It is likely that our models use different sources to model spending growth, generic entry, and the levels of reduction in prices. Over time, even small differences can compound into large ones.

Finally, the purposes of the two models are also different. This model is designed to examine the direction and magnitude of effects from changing different aspects of negotiation based on the current stock of branded drugs used in Medicare. Results from changing policies in this model are best compared with its base case, and not the results of other models.  

How does the model estimate changes for the first two years, since 25 drugs have already selected?

This model doesn’t use the first two selected cohorts as fixed inputs. This is to provide an “apples-to-apples” comparison of how changes in policy would affect outcomes in all years through 2030.

How well does the model predict the drugs already selected?

There are some differences in what’s selected for 2026 and 2027. Our model relies on projected estimates of spending using CMS 2022 dashboard data and analyst revenue growth estimates sourced from Evaluate Pharma. CMS uses more recent and granular data for its selection. This likely contributes to differences in spending levels that lead to selection. It also appears to make our model more likely to identify drugs as being made by a small biotech company, which exempts them from selection in the first years of the program.

Of the 25 drugs selected by CMS, the model selects 19 in the same year as CMS. Some drugs are selected by both CMS and the model in different years. Not taking year into account, the model identifies 21 of 25 drugs. Disregarding our model’s small biotech identification further increases concordance to 23 of 25 drugs over the two-year period.

How does the model handle generic and biosimilar entry?

The model uses a simplifying assumption that branded drugs currently on the market will have no biosimilar or generic entry before 2030, and that they remain selected for three years before they exit the program due to generic or biosimilar entries.

We estimated the effect of generic or biosimilar entry occurring before or after selection by drawing 1,000 simulations from a distribution of time to generic or biosimilar entry from Berger et al.  This change resulted in a -30.38% reduction in federal savings (95% CI -49.47% - 25.67%).

What happens if you alter assumptions about net prices negotiated by Part D plans?

The model assumes that Part D plans achieve similar net prices to the rest of the market. However, this may not be the case. Using published estimates from Hernandez et al and net price data from SSR Health, we estimate that Part D plans net prices are higher than those of the overall market, amounting to 71% of the market average reduction relative to list prices. Adopting this assumption increases federal savings by 7%.  

The level of negotiated prices relative to Part D net prices for drugs in competitive therapeutic classes is particularly uncertain. The model assumes a reduction of 10% relative to prevailing Part D net prices. Altering this assumption by 10% in either direction results in 9% increases and decreases in federal savings.  

How sensitive are results to uncertainty in spending growth projections?

The model bases spending growth rates on analyst revenue projections for individual drugs sourced from Evaluate Pharma. We estimated the effects of uncertainty in spending projections by running a 1,000 iteration simulation in which growth rates were randomly adjusted based on forecast error estimates in a study by Laroque. Under this approach, federal savings were on average 0.30% lower (95% CI -1.02% - 0.43%).