The problem with international reference pricing

The Department of Health and Human Services (HHS) recently put forth a proposal to change the reimbursement rules for drugs administered by physicians under Medicare Part B. These include drugs which are commonly administered in physicians’ offices, such as chemotherapy and many injectable and infused therapies. Currently, these drugs are reimbursed based on existing sales prices. The Centers for Medicare and Medicaid Services pays physicians who administer these drugs the Average Sales Price (ASP) in the US – defined as quarterly US sales, net of rebates, divided by total number of doses sold – plus a 6% dispensing fee.

The proposed rule would replace the ASP with a benchmark (Target Price) based on international prices. The exact formula for the Target Price is yet to be determined, though it will be a function of prices in a currently unspecified basket of countries. After the rule is implemented, reimbursement for drugs whose ASP exceeds this average would be adjusted downward.

With the publication of the “International Pricing Index Model for Medicare Part B Drugs” proposal, the Trump administration takes a step towards fulfilling one of the key promises of its “American Patients First” blueprint: lowering the price of prescription drugs. In doing so, it also seeks to address another of its recurring themes: the imbalance in drug prices between the US and the rest of the world.

While lowering drug prices is a valuable pursuit, there is a significant chance that this proposed new approach would achieve minimal savings, while also drastically limiting foreign access to prescription drugs.

The International Pricing Index Model is a form of External Reference Pricing (ERP). Governments use ERP to limit the ability of manufacturers to charge different prices across countries, a strategy referred to as price discrimination. Many governments in Europe use some form of ERP.

The evidence from Europe shows why the ERP model can be problematic. To minimize the effects of ERP pharmaceutical companies, charge higher-than-expected prices in lower income countries, and sometimes even delay entry. For example, Abacavir, an antiretroviral therapy for HIV that is part of the WHO list of essential medications was introduced in all countries in Western Europe within three years of receiving marketing approval, in 1999. The only Eastern European country where Abacavir was available at the time was Poland, where it cost the same amount as in Italy. Other Eastern European countries had to wait at least an additional two years before the drug became available. In Slovakia, Abacavir was launched in 2008, at an initial launch price that was comparable to Germany’s, the country with the highest price levels in Europe.

 This reaction produces two consequences. First, patients in lower-income countries find it more difficult to access prescription medication, both because these drugs only become available several years after first receiving marketing approval, and because when they do become available, they cost more. Second, governments of higher-income countries find it more difficult to use ERP effectively to lower prices because prices in lower-income countries are set higher and later.

Because the US pharmaceutical market has the highest prices and generates the largest revenue among all countries in the world, adopting ERP in the US would lead to much stronger incentives for manufacturers to resist low prices in foreign countries. Consider the hypothetical example of a drug manufacturer with a new drug approved for the US and Canada. As a whole, the US market is about ten times bigger than the Canadian market. Thus, if Canadian prices were just 10% lower than US prices, the manufacturer would be better off forgoing sales in Canada entirely and selling only in the US than risk having the price in the US fall by more than 10%. In practice, Canadian drug prices are currently about 60% lower than US drug prices. If reference pricing is adopted, manufacturers would likely hold out for a doubling of Canadian prices or threaten (credibly) to not sell there at all.

HHS estimates that the International Pricing Index Model would save Medicare more than $8 billion dollars from 20 drugs only. However, this analysis assumes that prices abroad will remain the same, which is implausible. Instead, manufacturers will almost surely raise their prices abroad, and, if raising prices proves too difficult, delay entry to prevent the US government from observing a lower price or simply introduce different versions of the drug in the US and abroad. Either reaction would imply that the Target Price will approach the initial US price level, which in turn means that Medicare spending would not decline very much, or at all.

Rather than trying to impose foreign prices, HHS would be better off importing foreign wisdom. Foreign prices for pharmaceuticals are lower because most governments can bargain directly with pharmaceutical companies, which gives them sufficient leverage to obtain better deals. Moreover, these governments use a variety of tools that CMS is currently unable or unwilling to use, such as cost-effectiveness evaluations, risk-sharing arrangements, and public tendering. Rather than outsourcing price negotiations to other countries, it would be much more fruitful for the US government to start negotiating itself.