In the first quarter of 2024, the American Society of Health-System Pharmacists counted 323 active drug shortages in the United States. It was the highest number since ASHP began tracking in 2001, edging past the previous record of 320 set in 2014. Counts have eased since then, but not by enough to call the problem solved.
The drugs on those lists are rarely exotic. They are chemotherapy agents that have been off patent for decades, antibiotics, and the emergency medicines kept in hospital crash carts. Many are generic sterile injectables. In 2023, the United States went through an acute shortage of cisplatin and carboplatin, workhorse cancer drugs. The trigger was a single Intas Pharmaceuticals plant near Ahmedabad, India. FDA inspectors who visited in late 2022 documented serious data-integrity failures, including, as the Washington Post reported, bags of shredded quality records, and the agency later issued a warning letter. Intas halted production there in 2023. According to the Post, the plant had supplied more than half of the US market for both drugs. By May 2023, a National Comprehensive Cancer Network survey found 93 percent of responding cancer centres short of carboplatin and roughly 70 percent short of cisplatin. Oncologists rationed, delayed and reordered treatment.
The usual explanation is a fragile supply chain: too much production offshore, too few suppliers, too little inventory. All true. But the thesis here is that those are symptoms. The underlying cause is a market where the cheapest reliable drug and the cheapest drug are priced the same, so nobody pays for reliability, and eventually nobody provides it.
Why old drugs are the ones that vanish
Look at what disappears. Branded drugs, with healthy margins, rarely run short for long. Their makers can afford redundant plants, safety stock and strong quality systems, and a shortage costs them real money in lost sales and lost reputation.
Old generics are different. Once several manufacturers compete, prices fall toward the cost of production. That is the point of generic competition, and it has saved health systems enormous sums. But when margins get thin enough, a manufacturer facing a costly quality upgrade may simply stop making the product. When a large plant fails an inspection, the remaining suppliers often lack spare capacity to fill the gap, because spare capacity is a cost nobody was paying for.
The cisplatin episode is a textbook case. After Intas went dark, the other makers of the drug could not ramp up quickly. Sterile injectable lines take time to qualify, and nobody had been paid to keep idle ones warm. The FDA ended up allowing temporary imports of unapproved cisplatin from abroad, a measure that tells you how little slack was left in the system.
The mechanism, in one uncomfortable sentence
A buyer who cannot tell a reliable supplier from an unreliable one will pay both the same price, and at that price only the unreliable one can afford to stay.
This is a quality problem that economists would recognise from used-car markets. Sterile manufacturing quality is largely invisible to a hospital pharmacist. The vial from a plant with a mature quality system looks exactly like the vial from a plant cutting corners. Both carry FDA approval. Both appear on the group purchasing contract as the same molecule, same strength, same fill. So the contract goes to the lowest bid, and the lowest bid tends to come from whoever is spending least on the things the buyer cannot see: backup equipment, extra quality staff, a second source of active ingredient, a few months of finished inventory.
Over years, that selection pressure squeezes reliability out of the market. Nobody decides to make the supply chain brittle. It just happens, one tender at a time.
A back-of-envelope case for paying more
Numbers help here, so consider an illustrative hospital. The figures below are assumptions for the sake of the arithmetic, not data from any real institution.
Suppose a mid-sized hospital spends $2 million a year on a basket of 50 critical generic injectables. It decides to pay a 10 percent premium to split each award between two qualified suppliers and to hold an extra month of stock. That costs roughly $200,000 a year in higher prices, plus the carrying cost of the inventory, call it another $30,000.
Now price a shortage. When a critical drug goes short, pharmacy staff spend hours each week finding alternatives, rewriting order sets and chasing grey-market offers. Clinicians switch patients to substitutes that may be costlier, less familiar or less effective. Procedures get postponed. If one serious shortage a year consumes the equivalent of two full-time pharmacy staff for six months, forces purchases of alternatives at several times the usual price, and delays even a handful of oncology cycles, the bill can plausibly pass the $230,000 premium. And that ignores the clinical harm, which is the part that should matter most.
The point of the exercise is not the exact figure. It is that the premium appears on a line item, while the shortage cost is scattered across pharmacy labour, clinical workarounds and patient outcomes that no budget report sums up. A hospital that runs this calculation honestly often finds redundancy is cheaper than it looks.
The buyer’s share of the blame
Here is the uncomfortable part for health systems. Hospitals and their purchasing groups are not only victims of shortages. Their buying behaviour helps create them.
For years, the dominant purchasing strategy for generic injectables has been to secure the lowest unit price, often through contracts that award most of the volume to one supplier. That approach is rational for any single hospital. Across the whole market, it rewards the lowest-cost producer, who is often the one investing least in redundancy and quality, and it starves the second and third suppliers of the volume they need to stay in the market.
Just-in-time inventory completes the picture. Hospitals holding a few days of stock have no buffer when a supplier falters. When a shortage is announced, many then order heavily at once, which turns a tight supply into an empty one. It is a bank run, conducted in vials.
None of this is irrational. It is what happens when everyone optimises for their own price.
The strongest counterargument
The best case against this thesis goes like this. Paying more does not guarantee reliability. A premium handed to a manufacturer may simply become margin, with no new line, no second site and no better quality system. Some shortages come from genuine accidents: a tornado hitting a plant, a sterility failure at a well-run facility, a sudden surge in demand. And raising generic prices across the board would cost health systems and patients billions, much of it wasted on drugs that never go short.
That argument partly holds. A blanket price rise would be a poor policy. Most of the thousands of generics on the market are made by enough suppliers, in enough places, that they rarely fail. And money without conditions buys nothing in particular.
But the counterargument attacks a version of the thesis nobody should hold. The case is not for paying more for everything. It is for paying more, on conditions, for a short list of critical, low-margin products with concentrated supply. Pay for a second qualified source. Pay for a verified inventory commitment. Pay for measured quality. That is a targeted purchase of insurance, not a subsidy, and it can be audited.
Second-order effects worth anticipating
Fixes to shortages have their own side effects, and a serious buyer should expect them.
Split awards and longer contracts reduce price competition in the short term. Some savings go away, and finance teams will notice before anyone notices the shortages that did not happen.
Reshoring, a popular political answer, raises costs and does not by itself solve concentration. A single domestic plant is just as able to fail an inspection as a single foreign one. What matters is the number of independent sources, not the flag over the factory.
Stockpiles can create their own shocks. If every hospital tries to build a three-month buffer at the same time, the rush itself can trigger the shortage it was meant to prevent. Buffers should be built slowly, by agreement with suppliers.
And better quality ratings, once published, will concentrate demand on the top-rated plants. That is the goal, but it could also overload them unless contracts give them the volume certainty to expand.
What each actor can actually do
Hospital chief executives and chief pharmacists can identify their twenty or thirty most critical generics, map how many plants actually supply each one, and move those products to dual-source contracts with inventory commitments. The premium should be reported to the board alongside the shortages avoided, so it is not quietly cut in the next budget round.
Group purchasing organisations can build reliability into their scoring, awarding points for redundant manufacturing, disclosed sourcing and inventory guarantees, instead of treating price as the only differentiator.
Payers and government insurers can recognise that a hospital paying a reliability premium is not being wasteful. Payment rules that reimburse generics at the lowest available cost make it harder for providers to buy anything else.
Regulators can finish building a quality rating that buyers can actually use. The FDA has been developing a quality management maturity program to assess how robust manufacturers’ quality systems are. A rating kept inside the agency changes little. A rating that purchasers can see and contract on changes incentives.
Manufacturers and founders looking at sterile injectables should notice that the business model only works if some buyers will sign long, volume-guaranteed contracts. A new entrant without such commitments is making a bet that the market has repeatedly punished.
Ministries of health outside the United States should not assume this is an American problem. Many countries buy generics through national tenders with the same winner-takes-all logic, and several have seen shortages of the same old antibiotics and cancer drugs. The fix translates: split tenders, reward reliability, and know where your critical drugs are made.
Lessons that hold up
Since the 2024 peak, a lot of proposals have circulated: tariffs, reshoring, stockpiles, export restrictions. Some help at the margins. A few lessons seem more fundamental. Reward quality, not only price, which requires a visible quality signal and a buyer willing to pay for it. Split awards on purpose, because contracting with two or three suppliers, even at a small premium, keeps backup capacity alive. Commit to volume for longer, since multi-year contracts give manufacturers a reason to invest where one-year tenders at the lowest bid do the opposite. Hold modest buffers for the critical few, meaning the crash-cart drugs and the core chemotherapy agents rather than everything. And demand transparency about where critical drugs are made and how many plants supply them, so buyers can spot concentration before it fails.
Why this is hard to fix
Paying more for generics is politically awkward. Every stakeholder has spent a generation celebrating low generic prices. A hospital chief pharmacist who pays a premium for a second supplier has to explain the higher bill every year, while the benefit, a shortage that never happens, is invisible.
That asymmetry is the heart of the problem. Shortages are expensive, but the cost is diffuse: delayed treatments, substitutions, staff time spent hunting for supplies. The savings from cheap purchasing are concentrated and show up in the budget. Until purchasers can count the cost of unreliability as clearly as they count unit price, the market will keep delivering the same result.
What to watch over the next two years
Three signals will show whether anything has changed. First, whether large purchasing groups publish contracts that explicitly pay for redundancy or inventory, and whether those contracts are renewed once the memory of the last crisis fades. Second, whether a manufacturing quality rating moves from a regulatory pilot into something buyers can see and use in tenders. Third, whether the number of sole-source critical injectables falls. The headline shortage count is a lagging indicator. The number of products that depend on one plant is a leading one.
If those three move, the next Intas-style failure will be a problem. If they do not, it will be another crisis.
The takeaway
The record set in early 2024 was not caused by one tornado, one plant or one country. It was caused by a market design that treats all suppliers of an old drug as interchangeable and pays them accordingly. Fixing shortages will cost money, paid deliberately for redundancy and quality. The alternative is paying the same money, or more, in the chaos of the next shortage.
Has your hospital or purchasing group changed its contracting to reward reliability? What did it cost, and was it worth it? We’d like to hear.
