Pricing assumed we knew the value. We usually don't.

Three doors: no, blind yes, or a shared bet.

Faced with a promising-but-uncertain drug, a payer has three doors, and only three.

Door one: say no. Refuse to fund it until the uncertainty resolves. Safe for the budget, but patients who might benefit get nothing, possibly for years, and some can't wait. A real cost, just an invisible one.

Door two: say yes, at full price. Fund it now, betting the promise holds. If it does, patients win and you paid a fair price. If it doesn't (the survival benefit evaporates, or five times as many patients show up as expected) you've committed a fortune to something that didn't deliver, and (as we'll see) you can't easily undo it.

Door three: share the bet. Don't stake everything on a single yes or no made blind. Instead, structure the decision so that payment is tied to how things actually turn out: pay something now, and make the rest conditional on the volume, or the results, that emerge. Fund the drug, but not unconditionally: split the risk between the payer and the manufacturer.

That third door is the managed-entry agreement (MEA), also called a risk-sharing scheme (RSS). It's not a discount and it's not a favour; it's a contract that reallocates risk, letting a payer say "yes, conditionally" where the only alternatives were a blind yes or a flat no. And the entire craft of designing one comes down to a single question: what, exactly, is the risk you're trying to share?

Two kinds of risk.

Because there isn't one risk: there are two, fundamentally different, and almost every MEA exists to address one or the other. Getting the diagnosis right is the whole game.

Financial (budget) risk. Here you're reasonably confident the drug works and is good value per patient, but you're unsure what it will cost you in total. Maybe the eligible population is hard to predict, maybe usage will creep beyond the licensed indication, maybe the sheer volume could blow the budget even at a fair per-patient price. The uncertainty is about how much, not whether it works. This is the Module 10 problem: cost-effective, but is it affordable?

Outcomes (performance) risk. Here the worry is different and deeper: you're not sure the drug delivers what it promises. The effect size is uncertain, the long-term benefit unproven, the trial too short or too clean to trust for real-world results. The uncertainty is about whether it works, or how well: the Module 9 (value of information) and Module 11 (real-world evidence) problem.

These are not variations on a theme: they're different diseases, and they need different medicines. An agreement that brilliantly controls your budget does nothing to tell you whether the drug works. An agreement that pays only for patients who respond does nothing to cap a runaway budget if it works in everyone. Before you design any deal, you must diagnose which risk you actually face, because the tools split along exactly this line.

Financial-based agreements.

Start with the tools for financial risk: the ones that protect the budget without ever asking whether the drug works (because, for this risk, you're not worried about that).

• Simple (confidential) discount. The bluntest tool: a straight price cut, usually confidential to protect the list price for reference-pricing purposes (Module 10's list-versus-effective-price point). Lowers cost per patient; does nothing about volume.
• Price-volume agreement. The price per unit falls as volume rises. The logic: a bigger market should command a lower unit price, and it protects the payer if usage turns out higher than expected, more patients, but each one cheaper.
• Spend cap (budget cap). A hard ceiling on total spend: beyond an agreed amount, the manufacturer refunds the excess or supplies the drug free. Whatever the volume, the payer's total exposure is bounded. The purest answer to "what if far more patients show up than we planned?"
• Dose or duration cap. The payer pays up to a defined dose or treatment length; beyond that, the manufacturer covers the cost. Useful when a minority of patients stay on the drug far longer than the average, driving disproportionate cost.

Notice the common thread: every one of these protects the budget and none of them cares whether the drug works. That's not a flaw: it's the correct design when the risk is financial. You've already accepted the drug delivers value; you're only managing the size of the bill. Point any of these at a drug whose effectiveness is in doubt, though, and they miss entirely.

Diagnose the risk, choose the deal.

Each scenario below hinges on a specific uncertainty. First diagnose the kind of risk: financial or outcomes. Then pick a deal to share it. Watch whether your deal actually addresses the risk, or misses it: a good tool aimed at the wrong risk still fails.

Every scenario had a good tool and a wrong tool, and the wrong tool wasn't a bad instrument, it was a good instrument pointed at the wrong risk. A spend cap is excellent engineering; it just can't tell you whether a drug works. Payment-by-results is elegant; it does nothing to cap a budget. This is the entire discipline of MEA design in one move: diagnose the risk first, choose the mechanism second. Reach for a favourite tool before diagnosing, and you'll confidently solve the wrong problem.

Now you.

Outcomes-based agreements (and why they're seductive).

Now the tools for outcomes risk: where you're unsure the drug delivers, and no budget cap can answer that. Here the deal has to tie money to what actually happens.

Payment-by-results (performance-based). You pay only for the patients in whom the drug works, by a pre-agreed definition of "works." The tumour responds: you pay; it doesn't: you don't, or you're refunded. On paper it's the perfect alignment: the manufacturer, confident in their drug, stakes payment on it performing; the payer stops paying for failures. Risk sits squarely with the party who claims to know the drug is good.

Coverage with evidence development (from Module 11). Rather than paying per individual result, you fund the drug provisionally while systematically collecting real-world evidence, then revisit the decision once the evidence matures: confirming, renegotiating, or withdrawing. It's the mechanism that turns "too uncertain to decide" into "decide, then learn," and it's how the value-of-information logic of Module 9 becomes an actual contract.

Both are genuinely clever, and both point payment at reality instead of at a promise. This is the intellectually satisfying end of the MEA world: deals that seem to dissolve the very uncertainty that made the decision hard. Which is exactly why the next screen matters, because "seems to" is doing a lot of work in that sentence.

Why the clever deal often fails.

Here's the hard-won truth that separates the textbook from the trenches: the more elegant the deal, the more likely it is to break in practice. Outcomes-based agreements are beautiful on paper and notoriously painful in reality, for reasons that are worth knowing before you propose one.

• They need an outcome you can actually measure: fast, cheaply, and unambiguously. Payment-by-results requires a clear signal: did it work for this patient, yes or no, soon enough to matter. But most of medicine isn't like that. The benefit is often survival measured in years, or a slow reduction in risk, not a clean "responded / didn't respond" you can read in three months. If you can't define and measure the outcome crisply, payment-by-results has nothing to hang payment on.
• They cost a fortune to run. Tracking every patient's outcome, adjudicating disputes, verifying who responded and who didn't, moving money accordingly, this is heavy administrative machinery. The transaction cost can eat much of the value the clever deal was supposed to create, and both sides quietly come to resent it.
• They ratchet. As Lesson 3 warned, once a drug is in use, stopping is politically brutal: patients are on it, clinicians prescribe it. A deal that says "we'll withdraw if the evidence disappoints" often can't be enforced when the moment comes. The condition is easy to write and hard to pull.

So the counter-intuitive lesson: a boring financial deal (a flat confidential discount, a simple spend cap) frequently outperforms a sophisticated outcomes-based one, precisely because it has no moving parts. It doesn't need a measurable endpoint, an audit system, or the will to withdraw. The sophisticated deal wins only when its demanding conditions are genuinely met: a fast, clean, cheap-to-measure outcome (like Scenario C in the interactive) and the operational capacity to run it. Match the deal's complexity to your ability to execute it, not just to the elegance of the idea. Sometimes the smartest deal is the dumb one.

What's the best critique?

Why this matters for HTA

Managed-entry agreements are where an assessment becomes a deal, and designing or critiquing them well is a defining practical skill in market access:

• Diagnose the risk before reaching for a tool. The single most common MEA failure is a well-built mechanism solving the wrong problem: an outcomes deal on a budget risk, a spend cap on an effectiveness question. Always name the risk first: is it how much it'll cost or whether it works? Everything else follows from that answer, and getting it wrong wastes the whole instrument.
• Respect the execution cost, not just the design. An agreement is only as good as your ability to run it. Before proposing an outcomes-based scheme, ask whether the outcome is genuinely measurable, fast, and cheap to verify, and whether the withdrawal condition could ever actually be enforced. If not, a simpler financial deal will usually deliver more, more reliably. Elegance on paper is not performance in practice.
• Read a proposed deal as a claim about uncertainty. When a manufacturer offers a particular structure, it reveals what they think the risk is, and sometimes what they're trying to keep hidden (a confidential discount protecting a list price, an outcomes deal betting on a soft endpoint). The structure of the deal is itself evidence; read it.

A risk-sharing deal isn't a discount or a favour: it's a contract that splits a specific uncertainty between the two parties. Name the uncertainty correctly and the right deal is almost obvious; name it wrongly and no amount of cleverness in the contract will save you.

Risk-sharing & managed-entry agreements, in one breath.

• Pricing assumes we know a drug's value; often we don't. A managed-entry agreement (MEA / risk-sharing scheme) is the "third door": neither a blind yes at full price nor a flat no, but a contract that shares a specific risk, letting a payer say "yes, conditionally."
• The whole craft is diagnosing which risk you face. Financial (budget) risk: the drug works, but how many patients / how much total? Outcomes (performance) risk: does it actually deliver what it promises? Different risks need different deals.
• Financial tools protect the budget without asking whether the drug works: simple/confidential discount, price-volume (price falls with volume), spend cap (hard total ceiling), dose/duration cap. Outcomes tools tie money to reality: payment-by-results (pay per success) and coverage with evidence (fund, collect real-world data, revisit).
• The counter-intuitive truth: the more elegant the deal, the more it tends to fail in practice. Outcomes-based deals need a fast, clean, measurable outcome (rare), cost a fortune to administer, and ratchet (hard to withdraw). A boring financial deal often wins because it has no moving parts. Match the deal's complexity to your ability to execute it.

Every risk-sharing deal answers one question: what, exactly, don't we know: the size of the bill, or the truth of the promise? Get that diagnosis right and the contract almost designs itself; get it wrong and you'll engineer a beautiful solution to a problem you never had.

We've now managed uncertainty with clever contracts, but every tool in this lesson quietly assumed you can estimate the value, the population, the outcome, even if imperfectly. There's a corner of medicine where even that assumption collapses: drugs for diseases so rare that the trial has forty patients, the QALY estimate is a guess, the budget is enormous per patient, and the whole apparatus of HTA (thresholds, ICERs, risk-sharing) strains to the breaking point. How health systems handle orphan drugs and rare diseases is the final lesson of this module.