Jackson Hammond is a Senior Policy Analyst at Paragon Health Institute. He has been active in the federal and state health policy space since 2017.
Prior to joining Paragon, Jackson was a health care policy analyst for American Action Forum (AAF). While at AAF, his work focused on payer issues including private insurance, Medicare, and Medicare Advantage. Furthermore, Jackson wrote extensively about the 340B Program and contributed to AAF’s research on a variety of drug pricing issues.
On December 19, the Innovation Center at the Centers for Medicare and Medicaid Services (CMS) proposed two new mandatory drug pricing models: the Global Benchmark for Efficient Drug Pricing (GLOBE, for Part B drugs) model and the Guarding U.S. Medicare Against Rising Drug Costs (GUARD, for Part D drugs) model. These models would implement a drug pricing system that pegs the prices Medicare pays for drugs to those of select foreign nations, a policy referred to as a Most-Favored-Nation (MFN) system. Paragon has previously written about MFN models and advocated for a prospective method that would only apply to new drugs, rather than altering prices of drugs already on the market. The GLOBE and GUARD models use a retrospective method that applies to existing drugs and would collectively apply to the drug purchases of approximately 42 percent of Medicare enrollees. GLOBE and GUARD come on the heels of a voluntary MFN model that applied to Medicaid, the GENErating cost Reductions fOr U.S. Medicaid (GENEROUS) model. Below, I will cover how GLOBE and GUARD will operate, their expected impact, and how they compare to a prospective MFN model.
GLOBE
The GLOBE model will calculate an alternative rebate to the current Part B prescription drug inflation rebate for purchases of specified Part B drugs by enrollees with traditional fee-for-service (FFS) Medicare. Part B covers drugs that are typically administered at facilities such as physicians’ offices or hospital outpatient departments, as opposed to those dispensed from retail pharmacies. CMS will apply the larger of the inflation rebate or the GLOBE model rebate (explained in the appendix) to the specified drug.
CMS will determine enrollee cost-sharing by taking the lesser of 20 percent of either the standard payment amount or the adjusted GLOBE model benchmark (see appendix).
CMS will apply GLOBE model prices and cost-sharing to 25 percent of the traditionally enrolled Medicare Part B population. Enrollees who do not have Part B as their primary payer for these drugs and enrollees with Medicare Advantage (MA) are excluded from the model. Drugs meeting all four of the following criteria are included:
In selecting a pool of comparison nations, CMS will choose those nations that have at least 60 percent of the U.S.’ GDP, adjusted for purchasing power parity (PPP) and an annual real GDP of at least $400 billion. Per CMS, the countries meeting those requirements as of October 1, 2025, are: Australia, Austria, Belgium, Canada, Czech Republic, Denmark, France, Germany, Ireland, Israel, Italy, Japan, the Netherlands, Norway, South Korea, Spain, Sweden, Switzerland, and the United Kingdom.
The GLOBE model will run for five years starting on October 1, 2026, with two additional years for CMS to finalize and reconcile payment calculations. CMS estimates that the GLOBE model’s approach to determining the list of drugs included may encompass 55 percent of annual Part B FFS drug spending and will save the federal government $13 billion and seniors $4 billion over the course of the model.
GUARD
The GUARD model is very similar to GLOBE. It will test an alternative to the Part D prescription drug inflation rebate for purchases of specified drugs by Part D enrollees and assess how that alternative rebate reduces spending. Part D covers drugs that are typically dispensed at retail pharmacies. Like GLOBE, CMS will apply the larger of the inflation rebate or the GUARD model rebate (see appendix). If the net price of a drug including the inflation rebate exceeds the selected GUARD benchmark, manufacturers will be required to pay a rebate to Medicare equal to the price difference.
The following sole-source drug and sole-source biological product therapeutic categories are included: analgesics; anticonvulsants; antidepressants; antimigraine agents; antineoplastics; antipsychotics; antivirals; bipolar agents; blood glucose regulators; cardiovascular agents; central nervous system agents; gastrointestinal agents; genetic or enzyme or protein disorder (replacement or modifiers or treatment); immunological agents; metabolic bone disease agents; ophthalmic agents; and respiratory tract/pulmonary agents. Drugs will be excluded if they:
The GUARD model will run for seven years starting on January 1, 2027, through December 31, 2033, with an additional two years for CMS to finalize and reconcile payments. It will use the same reference country criteria as the GLOBE model. The GUARD model will cover 25 percent of Part D enrollees. CMS projects that the GUARD model will save the government $14 billion in reduced Medicare Part D spending over the course of the demonstration.
Prospective MFN Is a Better Way Forward
American patients have long subsidized the pharmaceutical innovation that other developed nations enjoy; this situation is no longer tenable, either financially or politically. The goal of the GLOBE and GUARD models is to save money for patients and taxpayers, as well as to equalize prices between the U.S. and other developed nations. However, both models—as well as the GENEROUS model—use a retrospective MFN design. As I and my colleagues at Paragon have previously written, a retrospective design will disrupt existing agreed-upon pricing arrangements and harm U.S. pharmaceutical innovation by lowering rates of return on existing products. Manufacturers base investments in research and development on these returns. Importing previously agreed-upon foreign government-set prices will result in significantly fewer resources for research and development for manufacturers in the immediate term, while offering no mechanism to incentivize other developed nations to raise their prices. Moreover, if manufacturers cannot trust the government not to interfere with the prices of drugs already on the market, manufacturers are likely to invest less—and more cautiously—in researching and developing new drugs.
This is not to say that U.S. patients and taxpayers should pay whatever price pharmaceutical manufacturers charge. While MFN models inherently reduce the revenue that manufacturers receive from Americans, a prospective system would mitigate the loss of U.S. revenue with higher revenue abroad. In a prospective MFN system, the U.S. would only apply MFN pricing to future drugs or future contracts between manufacturers and other nations for existing drugs. This would give manufacturers both time and leverage to negotiate with other nations; manufacturers could credibly threaten to reduce or end access to their products for nations that do not agree to high enough prices.
Another concern with GLOBE and GUARD is that some of the included countries are significantly poorer than the U.S. In Paragon’s prospective MFN brief, we proposed a PPP-adjusted inclusion baseline of 90 percent of U.S. per capita GDP. The GLOBE and GUARD models’ decision to include nations with PPP-adjusted GDPs as low as 60 percent of the U.S. lumps tiny nations like the Czech Republic in with Germany. This will likely result in significantly lower payments than if the models had used a higher GDP baseline that is economically similar to the U.S. The potential set of countries for the mandatory GLOBE and GUARD models markedly differed from the voluntary GENEROUS model. GENEROUS had selected a much smaller set of more economically-similar countries—specifically the non-U.S. G-7 countries (the United Kingdom, France, Germany, Italy, Canada, and Japan), as well as Denmark and Switzerland. CMS should consider finalizing the model with the GENEROUS country set.
In addition, while the combined $28 billion in savings over the next seven to eight years is a relatively small amount compared to the roughly $330 billion Medicare spent on drugs in 2023 alone, both models state the potential for further inclusion of additional therapeutic categories and more enrollees. As such, the models have the potential to expand and impact more of the pharmaceutical landscape.
Conclusion
Ending the disparity between prices in the U.S. and other developed countries is an important policy goal. However, retrospective MFN systems are far too risky to the U.S.’s already-tenuous position at the forefront of pharmaceutical innovation, with China threatening to overtake us. CMS should strongly consider implementing prospective MFN models in place of the current retrospective models, as well as adopting the more appropriate list of nations used in the GENEROUS model.
Appendix
GLOBE Benchmark Calculation
The GLOBE rebate will generally be determined by subtracting a calculated benchmark (see below) from 106 percent of the average sales price (which is typically what Medicare pays for Part B drugs) of the drug. The benchmark will be determined by the greater result of the following two methods:
CMS will then adjust the benchmark with a modest increase to account for market differences (population size, volume of sales, etc.) between the included nations and the U.S., as well as an additional increase that will typically equal 6 percent in order to mirror the standard 6 percent add-on percentage to average sales price that Medicare pays for Part B drugs.
GUARD Benchmark Calculation
The GUARD model will also use the greater of the amounts produced by Methods I and II listed above when determining the benchmark for a given drug.
After selecting the greater of the results from Method I (known as the default international benchmark in GUARD) or Method II (known as the updated international benchmark), CMS will adjust the selected price upward to account for market differences between nations. The default international benchmark will be adjusted upward by 2 percent, and the updated international benchmark will be adjusted upward by 5 percent.
The GUARD model will then compare the benchmark to the standard Medicare net price (determined by subtracting manufacturer rebates and discounts from the wholesale acquisition cost of the drug).
Subscribe
Sign up now for your health policy updates.
