Managed Entry Agreements: Paying for a Drug Before the Evidence Is In

A managed entry agreement is a deal that lets a health system start paying for an expensive new medicine before anyone is sure it is worth the money. The simple versions just lower the price through a confidential discount or rebate. The ambitious versions link payment to whether the drug actually helps patients, and that link is the whole problem: proving who benefited turns out to be slow, costly, and often inconclusive.

Key points#

• A managed entry agreement trades certainty for access: coverage starts now, and the risk of being wrong is split between payer and manufacturer.
• Financial deals (discounts, rebates, budget caps) are common because the math does not depend on any single patient's result.
• Outcomes-based deals (payment by result, coverage with evidence development) sound fairer but require registries and monitoring that are expensive and hard to run.
• Italy's national data suggest these agreements recover only a small slice of drug spending, and the outcomes-based ones return less than the plain financial ones.
• Confidential pricing protects the discount but blocks anyone outside the negotiation from judging whether the deal was good.

The bind a new drug creates#

Picture the moment a costly therapy reaches the market. A trial may have followed a few hundred patients for a year and shown movement on a lab value that stands in for the outcome people actually care about. The payer is now asked to fund the same drug for a decade, in patients who look nothing like the trial population, at a price that can run into six figures per person. The evidence is thin exactly where the spending is thick.

From there a payer has two clumsy options. It can refuse to cover the drug until better data arrive, which stalls treatment for people who might genuinely benefit. Or it can pay the full sticker price and hope the money buys something real. Neither feels defensible. The managed entry agreement is the negotiated compromise between them. The OECD's 2019 review of these arrangements, drawn from a survey and interviews across twelve countries, describes them as a way to open access while managing the uncertainty around a product's value and its effect on the budget (Wenzl and Chapman).

Two kinds of deal, one much harder than the other#

Not all of these agreements ask the same thing of the people running them.

Financial agreements#

The financial kind is the everyday tool. A straight price cut, a cap on total spending, or a volume rebate is easy to write into a contract and easy to enforce, because a discount does not care whether any given patient responded. The OECD review reports that arrangements of this sort have been used in at least two-thirds of the OECD and EU countries surveyed. They are unglamorous and they work.

Outcomes-based agreements#

The outcomes-based kind is where the idea gets interesting and where it starts to strain. Two designs dominate. In payment by result, the manufacturer is reimbursed only for the patients who actually respond to the drug. In coverage with evidence development, the payer grants temporary reimbursement on the condition that fresh data are collected while the drug is in routine use, and revisits the decision once those data land.

Both promises are hard to argue with on paper: pay for results, not for volume. Keeping the promise is another matter. Someone has to define what counts as a response, measure it in every treated patient, settle the borderline cases, and then actually trigger the refund. Reviews of the wider European and United States experience note that most performance schemes concentrate in cancer care and tend to lapse or go unrenewed within a few years.

Why the measuring is the hard part#

The machinery of measurement is where these deals tend to come apart. To pay by result, a system needs a registry that identifies every eligible patient, records the agreed endpoint, and connects that endpoint to a payment. It needs staff to run the registry and clinicians willing to enter data during appointments that are already crowded.

The OECD authors are blunt about the strain. They flag a limited ability to genuinely reduce uncertainty, because the real-world data collected are often of doubtful quality or hard to interpret. They point to the difficulty of acting on results once a coverage decision has to be reopened. And they describe an administrative load that falls precisely on that data collection and analysis (Wenzl and Chapman). The version of the deal that sounds most rigorous is the version that demands the most invisible labor.

What the numbers actually show#

Italy is a useful test case because it built a national registry system specifically to run these agreements. A study of sixty-two medicines managed this way between 2019 and 2021 found that all the paybacks combined recovered roughly 327 million euros, which came to about 0.9 percent of what public health facilities spent on medicines over those three years. The outcomes-based agreements returned less money than the simpler financial ones, despite costing far more to administer, and the authors concluded there is limited evidence that these arrangements lower spending overall.

Put plainly: the design that asks the most of the system returned the least. That does not make the whole approach a failure, but it does puncture the assumption that tying payment to outcomes automatically protects the public purse.

The cost hidden in the confidentiality#

There is a second, less obvious price. Financial agreements usually work by keeping the true price secret. The published list price stays high while an undisclosed rebate flows back to the payer. That secrecy is what allows a manufacturer to give one country a steep discount without having to extend it everywhere, since many systems set their own prices by referencing the published prices of their neighbors.

The same secrecy that makes the discount possible also shuts out independent scrutiny. If what was paid and what was recovered stay inside the negotiating room, neither you nor any independent analyst can judge whether the deal was any good. The OECD review names this directly: confidentiality itself becomes a barrier to independent evaluation. A tool built to manage uncertainty can end up concealing the very evidence that would resolve it.

Reading a coverage announcement well#

When a health system announces that it will cover a costly new therapy, you can get past the press release with the same handful of questions every time. Is the deal financial or outcomes-based? If it claims to tie payment to results, what exactly counts as a result, and who verifies it? Is there a registry capable of measuring that, or is the monitoring more hope than plan? And can anyone outside the room ever check what was actually paid?

None of this makes managed entry agreements a bad idea. They are a reasonable answer to a real bind, and for a one-off, very expensive therapy they may be one of the few ways to square early access with affordability. But the more a deal promises to pay strictly for value, the more monitoring it silently requires, and the record so far suggests that promise is far easier to sign than to keep.