In almost every industry, the replacement supplier is a question of price and availability. In the healthcare industry it is a question of quality and approval: The active ingredient, the material, the manufacturing site and the supplier are named in the regulatory dossier, and whoever swaps them changes the product in the legal sense. Every disruption therefore hits twice, once as an outage and once as waiting time for the replacement to be cleared.
The reported numbers give the scale: At the start of 2026, Germany's BfArM had 546 ongoing drug shortages on its register, and in the year before it received 1,514 notifications covering 1,041 shortages.1 That is the normal state. For antipsychotics, the average shortage lasted 230 days in 2025.2 For comparison: In the automotive industry, a chip-related production stop is counted in weeks, and the restart needs no authority.
Four segments under one umbrella term
The term healthcare industry covers four businesses that have little in common beyond the buyer. They differ in what actually matters for risk: How long it takes until a replacement supplier is allowed to deliver. The bottleneck sits in qualification: Procurement negotiates a new contract in weeks, while the release behind it follows deadlines of its own.
Exhibit 1 How long a supplier switch really takes
The gap between the fastest and the slowest segment runs to several years, and it does not follow technical complexity. A syringe is simpler than an endoscope, yet switching it is only faster as long as the sterilisation slot and batch release cooperate.
What follows from this: A group-wide supplier strategy that treats all four segments alike is wrong in at least one of them. For syringes and wound dressings, a framework agreement with a second vendor is enough. For a radionuclide with a six-hour half-life, the same agreement changes nothing at all. And because a switch takes months to years depending on the segment, early detection is what counts: A disruption that only surfaces as a crisis has already used up that lead time.
Five points where the chain breaks
The public debate looks at missing medicines at the pharmacy counter, which is the end of the chain. The causes sit further upstream, at five points that resemble each other structurally: A market the buyer does not move, high barriers to entry, no short-term substitute.
Exhibit 2 The chain and its bottlenecks
Traced along a single antibiotic tablet
Raw material The precursors come mostly from China 
Active ingredient and component Synthesis mostly in India or China 
Manufacturing Europe presses the tablet 
Packaging Blister, leaflet, batch release 
Distribution Wholesalers supply the pharmacies 
Hospital and pharmacy Only here does the shortage show A substance ban hits a few dozen substances, but thousands of products
Fluoropolymers such as PTFE sit in catheters, seals and coatings because they stay chemically inert inside the body. That same property makes them a target for regulation. The proposed twelve-year derogation is not decided, and the committee opinion is due by the end of 2026.
A raw material that cannot be stockpiled
The most widely used radionuclide in nuclear medicine decays faster than a supply chain can be replanned. It is produced from molybdenum-99, which in Europe comes from four research reactors: Petten, Mol, Řež and Świerk. The oldest site has been operating since 1957, the youngest reactor dates from 1974. France has been out since OSIRIS was shut down in 2015, and Germany does not produce at all.
The precursor comes from two countries
Europe formulates and packages, while the chemical synthesis of the precursors moved to Asia over two decades. For individual antibiotics the concentration is higher still. Without the active ingredient you cannot press the tablet, however modern the European plant may be.
Medical devices buy electronics in a market they do not move
An imaging system needs the same components as those two buyers, but orders volumes that earn no priority. Once allocation starts, medical technology waits at the back, and swapping a component triggers a fresh assessment because of the risk class.
Two methods, few facilities, no switching between them
Gamma irradiation covers another 40 to 45 percent, and other methods barely matter. Both depend on a handful of facilities with long permitting procedures, and moving from one method to the other means a fresh validation, not a rebooking. The US industry association expected 30 to 50 percent less sterilisation capacity under strict emission limits, and without sterilisation no sterile product can be shipped.
From here on, the chain is broadly supplied
Packaging, wholesale and dispensing have many suppliers and short switching paths, so a disruption there stays local. The bottleneck is created early in the chain: It only becomes visible here, when the shelf stays empty.
Two of these five are the ones we consider most critical, because the usual remedies fail on them. For active ingredients, diversification works in principle, except that the market is so concentrated that the second vendor often sits in the same industrial park as the first. For radionuclides it does not work at all: Technetium-99m, with its six-hour half-life, cannot be stockpiled, and a replacement reactor does not appear within a planning period.9
That removes the standard tool for the most critical positions. What remains is lead time: Start pre-qualifying the second supplier before the first one fails, and keep the regulatory documentation in a state where the switch is an activation rather than the start of a project.
What follows from this: In the healthcare industry, a delivery delay grows through the approval waiting time into a shortage of months. Weak points only show early if the chain behind the direct suppliers is visible. At several of the five points there is barely room to diversify, and there this visibility of critical dependencies is often the only lever.
The regulatory multiplier
After the technical restart, the chain waits for an authority: A notified body or regulator assesses the replacement, and that assessment carries waiting times of its own that have nothing to do with the disruption.
The number of notified bodies is the hardest figure in this row, because it describes a capacity limit rather than processing quality. Fifty-two notified bodies handle certification for an entire continent, and that capacity cannot be expanded at short notice, because the assessors themselves have to be designated and monitored. One more application in the system means a longer queue for everyone else.
Exhibit 3 The same disruption, two timelines
- Outage detected
- Replacement qualified
- In production
- Outage detected
- Assessment and clearance
- Replacement released
- In production
The multiplier stretches the disruption and at the same time makes precaution more expensive: A second qualified supplier costs more than a contract, because it is a regulatory position of its own that has to be maintained. That arithmetic explains why the second source so often exists only in the contract and not in the dossier: The dossier entry costs money continuously, the contract does not. But a second source without a dossier entry shortens the assessment time by not a single day.
On top of that comes regulation that turns into a risk itself: With fluoropolymers, the PFAS restriction covers a group of substances for which there is so far no equivalent substitute in catheters and coatings. ECHA proposes a limited twelve-year derogation for parts of medical technology; it has not been decided, and the opinion of the responsible committee is due by the end of 2026.10 A manufacturer therefore has to take material decisions today for products whose approvability will only be settled years from now.
What follows from this: The authority's assessment time can only be moved forward. A second source that exists only in the contract does not help when it matters; what shortens the wait is the alternative carried in the dossier and already assessed. Regulatory work therefore belongs in the preparation.
What policymakers are doing about it
The EU has recognised the problem and, with the Critical Medicines Act, created its first instrument that addresses the dependency itself rather than only its symptoms. In May 2026, Parliament and member states agreed on a compromise intended to secure supply of around 270 active substances classified as critical, among other things through joint procurement and incentives for European production. The regulation has not been formally adopted yet: Confirmation by the plenary and the Council is still outstanding at the end of August 2026, after which its rules take effect in stages from entry into force.3
It makes origin a procurement criterion and thereby creates, for the first time, a price premium for security of supply that European manufacturers need against Asian cost structures.
It does not build a plant. An active ingredient synthesis takes years from permit to validated production, and the breaking points are already in effect today, not once the regulation is fully in force.
Until then the risk stays where it sits today. What counts for the coming years is therefore above all which of the five breaking points a company can even see inside its own operations.
What actually helps in practice
From the five breaking points we derive three approaches. We have sorted them by what they change about the duration of a disruption, not by effort.
Pre-qualify replacement suppliers for critical materials
The only lever that shortens the regulatory section of the timeline is preparatory work: A replacement supplier or a second manufacturing site that has already been assessed and is carried in the dossier. It costs money continuously and saves months when it matters. Whether that pays off for a position is decided by the risk calculation.
Know the chain down to the precursor
Two suppliers drawing the same precursor from the same region are one supplier with two billing addresses: The diversification exists only on paper, because risks correlate along the chain and a disruption at the precursor hits both sources at once. Concentration in active ingredients makes this the rule, and it only becomes visible if the chain has been mapped beyond the direct supply relationship.
Track regulatory deadlines as risk dates
The PFAS decision, the end of the MDR transition periods and the entry into force of the Critical Medicines Act all carry dates. Unlike geopolitical ruptures, they can be planned as appointments, and anyone tracking them in the same register as supplier risk spots collisions early enough to act.
The US data series shows how urgent this has become: An average drug shortage lasted around two years in 2019, 4.3 years in 2024 and more than five in 2025.13 The direction is clear. Preparation targets exactly that duration, and for your own positions it can be estimated:
Calculation model How long is your restart?
Open calculation model
How long is your restart?
Model, not measurement. Origin does not enter the number of months, it tests its precondition: A second source only shortens the calculation if it does not depend on the same precursor as the first. The finding follows from the selected combination and from an industry observation that Indian active ingredient manufacturers source a substantial share of their starting materials from China.4 The time calculation is based on the ranges from Exhibit 1, split into sourcing (25 percent), qualification (35 percent) and regulatory assessment (40 percent). The scope of the change applies a factor of 0.7 to 1.5, in three tiers: Same substance with the same supplier at a new site, same substance with a new supplier, new substance with a new supplier. A second source that has already been assessed cuts the assessment share to one sixth. The numbers do not replace a case-by-case review, they show the order of magnitude and where the time sits.
Conclusion
The healthcare industry carries the same risk profile as other sectors with concentrated precursors, it simply repairs more slowly, because every replacement has to pass quality and regulatory requirements. That shifts the centre of gravity of precaution from reaction to preparation: What decides the length of a shortage is not how fast a company reacts when it happens, but how much regulatory work was done beforehand.
The five breaking points themselves remain. Only which of them comes under pressure first will shift: The next fixed date is the ECHA opinion on PFAS at the end of 2026. The five points therefore belong in continuous monitoring rather than a one-off survey.
Where production will sit in future is the next decision. It comes more easily once it is clear which positions can be moved at all.