Policy conversations often use the upstream–downstream metaphor because it captures a familiar problem: Debates tend to focus downstream, where the problem is visible — the flooded street, the blocked road, the failure that finally reaches the public. But many consequential decisions happen upstream before the visible problem appears.
Drug access follows the same upstream–downstream pattern. Manufacturers often criticize health plans and PBMs for blocking patient access to needed therapies, and sometimes that criticism is fair. But payer restrictions come at the end of a longer chain. Manufacturers decide upstream which products to advance, which indications to pursue, which markets to enter, and which products to keep on the market. Those decisions are less visible than payer denials, but they can be more absolute: If a company decides not to develop, launch, or continue a therapy, patients and prescribers never reach the point where a payer can say yes or no.
When “clinical” also means “commercial”
Not every abandoned drug should have moved forward. Many products fail for good clinical reasons. Some are ineffective. Some are unsafe. Some are too uncertain to justify further testing. But the line between clinical and economic reasons is less clean than the industry’s public language suggests.
A product may be described as lacking sufficient efficacy, lacking superiority, or having limited differentiation. Those may be valid clinical judgments. But they often carry commercial implications. A drug that works but does not work much better than available alternatives may have less pricing power, less expected market share, tougher reimbursement prospects, and weaker profit potential. “Insufficient superiority” sits right at the clinical–commercial boundary: It may mean the product is good enough for patients, but it may also mean it is not good enough for investors, payers, or the company’s own commercial teams. In practice, “not clinically differentiated enough” often means “not commercially attractive enough.”
Judgments about a product’s comparative clinical value are generally based on population-level evidence. A product that adds little benefit on average may still be the best option for some patients because of differences in response, tolerability, or the alternatives available to them. A product of modest average benefit may justify restrictions for a payer and abandonment for a manufacturer — but the two decisions leave patients in different positions. A restricted product can still be reached through an exception. A product that was never commercialized cannot. For a patient who would have done best on it, that is a missed opportunity and, in some cases, a worse outcome.
Blending clinical and commercial judgment is not an aberration. It is how pharmaceutical R&D is designed to work. Companies are not public-health agencies. They allocate capital based on expected return. The Congressional Budget Office has described pharmaceutical R&D investment as dependent on expected global revenues, development costs, and policies that affect demand for therapies and the supply of new drugs. That is how the business model works, not a moral accusation.
Older products show the economics most clearly. A drug can remain safe, effective, and clinically useful, yet lose its place in a portfolio once it comes off patent and margins thin. The clinical case for the drug has not changed; the economics have.
What the evidence shows
In a 2016 JAMA Internal Medicine analysis by Hwang and colleagues of 640 novel therapeutics entering phase 3 or other pivotal trials from 1998 through 2008, 344 failed in clinical development. Of those failed programs, 57% were attributed to inadequate efficacy, 17% to safety, and 22% to commercial reasons; for the remainder, the reason was unknown. That commercial category is best read as a floor, not a ceiling, because weak efficacy, uncertain superiority, and limited differentiation can also reduce expected revenue.
The same mix of clinical and commercial reasons shows up in the research on shelved drugs. A 2022 BMC Health Services Research systematic review by Krishnamurthy and colleagues found that promising drugs are commonly shelved because of insufficient efficacy or superiority to existing therapies and strategic business reasons, often tied to market prospects or industry consolidation.
The “killer acquisitions” literature ties the discontinuation of drug projects to economics even more directly. In a 2021 study, Cunningham and colleagues argued that established manufacturers may acquire smaller innovators in order to discontinue projects that could become future competitors. In pharmaceutical data, acquired drug projects were less likely to be developed when they overlapped with the acquirer’s existing portfolio, and the authors estimated that 5.3% to 7.4% of acquisitions in their sample were killer acquisitions. That does not prove that every abandoned asset was a suppressed breakthrough. It does show that shelving a drug can be a rational business decision, not just the result of clinical failure. When capital follows expected return, a therapy that works can still fail to earn a place in the market.
The no-go decisions we cannot see
Manufacturer no-go decisions leave little public record. A therapy that is never advanced, launched, or continued shows up only as an absence, and absences are hard to examine.
FDA can identify approved products that are not marketed or moved to the discontinued section of the Orange Book. Companies must give FDA reasons for not marketing a drug after approval, but those reasons are not public. So researchers can often see the outcome — an approved product is unavailable — without knowing whether the reason was demand, manufacturing, licensing, portfolio strategy, or something else.
Launch geography is equally hard to see from the outside. A therapy can be approved in one market yet unavailable in another because the manufacturer decides the expected price, volume, or launch cost does not justify entry.
The access debate rarely treats payer restrictions and manufacturer no-go decisions symmetrically. Payers ration access in the open through coverage rules that produce denial letters. Manufacturers ration access upstream, in development, launch, and discontinuation decisions that never reach a patient, a prescriber, a pharmacy counter, or a prior authorization request.
Follow the access debate upstream
Access is shaped by a chain of decisions, and manufacturers control many of the earliest links in that chain. The remedy is not to force manufacturers to commercialize every product or to pretend payer restrictions are harmless. Payer rationing and manufacturer rationing are not independent. Payer restrictiveness is an input to the manufacturer’s forecast. If a class is crowded with formulary exclusions, onerous prior authorization criteria, and step therapy, the expected return on a marginal molecule falls. Some products are left upstream before they ever reach a coverage decision.
If we are going to scrutinize payer behavior, we should also scrutinize manufacturer noncommercialization. Which approved products are not marketed? Which development programs or indications were stopped despite some evidence of benefit? Which products were discontinued, shelved, or never launched for reasons unrelated to safety or efficacy?
There is a concrete first step toward upstream transparency. The reasons companies give FDA for not marketing an approved drug already exist in writing; they are simply not disclosed. Some of that secrecy has a legitimate basis: A company may intend to out-license or sell a shelved asset later, and a public account of why it was set aside could weaken the product’s value in that deal. But that argues for limiting the detail of disclosure, not for withholding it entirely. FDA could report that a product was withheld for commercial rather than clinical reasons without publishing the underlying strategy. Standardized categories or aggregate reporting — alongside the Orange Book’s discontinued list — would let researchers and policymakers see how often approved therapies are withheld for reasons unrelated to whether they work. Disclosures of that kind would force no company to sell anything, or to reveal a specific business plan.
The access debate should look upstream as well as downstream. Payers can block access to products that reach the market. Manufacturers decide which products reach the market at all. Both decisions matter. One plays out on the public record. The other often comes with a quiet decision not to proceed.