Health supply chains failed in 2020 not because buyers were careless but because they had been optimised for a different objective. Cost-efficiency and resilience pull against each other: the same consolidation that lowers unit price concentrates risk in fewer places. Procurement reform has to resolve that trade-off, and largely it has not.
That makes the problem structural rather than administrative. It cannot be fixed by better contract management or a more competent buyer. It is a design choice about what a supply system is for — made deliberately, for defensible reasons, long before anyone needed to defend it.
What actually failed in 2020 — and what did not
The common account of the pandemic supply crisis is that it was a procurement scandal. Some of it was. But the scandal and the structural failure are separable, and conflating them implies cleaner contracting alone would have changed the outcome.
The structural failure came first and was global. On 3 March 2020 — before most national lockdowns — the World Health Organization reported that “surgical masks have seen a sixfold increase, N95 respirators have trebled and gowns have doubled”, and estimated that “industry must increase manufacturing by 40 per cent.” Its Director-General’s warning was about the supply system, not about buyers: “Without secure supply chains, the risk to healthcare workers around the world is real.”
The UK experience is documented in unusual detail. The National Audit Office’s The supply of personal protective equipment (PPE) during the COVID-19 pandemic (25 November 2020) found that stockpiles “were intended for an influenza pandemic” and “did not hold all the equipment that proved to be required during the COVID-19 pandemic, such as gowns.” It described a market of “desperate customers competing against each other, pushing up prices”. Unit price rises against 2019 ranged “from a 166% increase for respirator masks to a 1310% increase for body bags”.
One figure deserves more attention than it gets. Of 32 billion PPE items procured between February and July 2020, only 2.6 billion were delivered to front-line organisations in that period. The binding constraint was not the decision to buy. It was the physical time between ordering and arrival — the lead time a just-in-time system has no slack to absorb.
The Public Accounts Committee agreed in COVID-19: Government procurement and supply of Personal Protective Equipment (February 2021): the plans, stockpile and distribution arrangements “were inadequate for a coronavirus pandemic.”
What did not fail is worth naming. Distribution logistics, once stock existed, largely worked. The failure concentrated where a system with no buffer met a shock every other buyer in the world faced simultaneously.
The trade-off that reform keeps running into
Modern healthcare procurement rests on a coherent theory: aggregate demand, standardise specifications, tender competitively, hold minimal inventory. Each element is sound in isolation, and together they have removed very large amounts of cost from health systems.
Why consolidation genuinely works on price
Centralised purchasing converts a fragmented buyer base into a single large customer. Suppliers can plan production and amortise setup costs over longer runs. Buyers get lower prices, fewer contracting overheads and better assurance, because a central function can afford regulatory and quality expertise no individual clinic can. These are real gains, not accounting illusions.
They are harder to bank than the theory suggests. The NAO’s January 2024 report NHS Supply Chain and efficiencies in procurement examined the centralised English purchasing body, which oversees a catalogue of more than 600,000 products. Against a target of 80% of relevant NHS spending routed through it by 2023-24, reported mid-year market share in September 2023 was 57%. The £2.4 billion savings target had been calculated by a method that “re-counted savings from previous years” and “did not take inflation into account”. Between June 2022 and March 2023, “on average more than one quarter (26.9%) of eDirect orders were delivered late, by an average of 22 days.”
That matters specifically for resilience. If centralisation is not reliably delivering the savings that justify concentrating risk, a system may be carrying the fragility without collecting the discount.
Why the same mechanism concentrates risk
The mechanism producing the saving is the mechanism producing the fragility. Tendering on price selects the lowest-cost producer, which over successive rounds tends to mean the largest producer in the lowest-cost location. Standardisation means every buyer wants the identical item. Minimal inventory means no buffer between disruption and stockout.
Run that for twenty years and the supply base becomes efficient and correlated. Correlation is the operative word: buyers hold contracts with several suppliers and believe themselves diversified, but if those suppliers draw on the same manufacturers or input producers, the diversification is nominal.
Low-margin categories are the most exposed. The OECD working paper Shortages of medicines in OECD countries (2022) found shortage notifications rose “by 60% over the period 2017 to 2019” across 14 OECD countries, and that “shortages affected predominantly older, off-patent molecules.” It noted that “regulation and reimbursement policies, such as those that favour unsustainably low prices, may influence commercial decisions, putting supply at risk”, and that failure can follow “reliance on a sole active pharmaceutical ingredient (API) supplier”.
The Strategic Report of the Critical Medicines Alliance (February 2025) states the causal chain plainly: procurement practices “with a particular focus on price as the most important, or even the only meaningful procurement criterion – have been a significant driver of the current market dynamics for mature medicines. These market dynamics have led to consolidation of suppliers and outsourcing to other jurisdictions.”
The visibility problem: nobody can see past the first tier
Beneath the trade-off sits a more basic obstacle. Most buyers do not know where their products actually come from.
A procurement function contracts with a supplier, who is accountable for delivery. But that supplier’s own inputs — components, sterile packaging, active ingredients, the plant where an item is finished — are commercially confidential, and standard contracts rarely require disclosure. The buyer sees tier one clearly and everything below it dimly.
The scale of the resulting blind spot has been measured. A 2023 Health Affairs study, Competition And Vulnerabilities In The Global Supply Chain For US Generic Active Pharmaceutical Ingredients, mapped 565 facilities producing 1,379 generic active pharmaceutical ingredients across 42 countries. It found that “about a third of APIs were manufactured by a single facility, and another third were manufactured by two or three facilities.” Most striking for procurement policy: “more than one in every five APIs reflected markets in which current Food and Drug Administration standards would have failed to detect low competition because there were three or fewer API manufacturers despite there being four or more manufacturers of finished generic drugs.”
That is the visibility problem stated precisely: competition at the tier a buyer can see actively conceals sole-sourcing one tier up. A health system can hold a dozen contracts with no way of establishing whether they converge on two factories — and will find out during a disruption.
Fixing this is harder than it sounds. The OECD paper Promoting Resilience and Preparedness in Supply Chains (2024) notes that “there currently exists little research on full supply chain visibility beyond the second tier supplier”, and that pursuing it “would require considerable investment of time, money, and organisational resources”. It also cautions that “creating information assets without a culture of responsiveness will not assure resilience.”
What reform has attempted — and what it has not settled
Stockpiles
The instinctive response to shortage is to hold more, and it is the weakest available answer. A stockpile answers a specified scenario and performs badly against an unspecified one. The OECD observes that pandemic-era “centralised inventory of medical equipment and supply generally failed to meet needs… the selection of items needed in a crisis is hard to predict.” Tellingly, stockpile performance “in rich and poor countries was not significantly different” — a failure of design rather than funding.
Domestic and near-shore capacity
Onshoring is the most politically attractive response and the most economically contested. Domestic capacity is more controllable in a crisis, particularly when other governments restrict exports. But it is usually more expensive — the Critical Medicines Alliance puts generic production costs in Asia at “20-40% of the corresponding cost in Europe” — so capacity that only earns its keep in emergencies must be subsidised in normal years or it will not exist. The EU’s proposed Critical Medicines Act pays that subsidy explicitly through “Strategic Projects”. Whether it produces durable capacity or a subsidised industry that still cannot compete is unresolved.
Rules that let buyers reward something other than price
The quieter and more consequential strand concerns what a buyer is permitted to reward. If rules and incentives push evaluators toward lowest compliant price, resilience cannot be bought however desirable, because there is no column for it in the scoring matrix. The Critical Medicines Alliance quantifies how rare the alternative remains: awarding on the Most Economically Advantageous Tender, which allows non-price criteria, “currently still accounts for a minority of public procurement procedures (24% of countries), and most are awarded based on price only (62% of countries).”
The UK’s Procurement Act 2023, whose main provisions came into force on 24 February 2025, restructured the framework for public contracting, including the objectives authorities must have regard to and transparency requirements. The EU proposal goes further, requiring procurers of critical medicines to include “a broader set of requirements… such as diversified sources of input material and monitoring of supply chains.”
The unresolved tension is that a criterion is only as strong as the willingness to lose money on it. Permitting a buyer to weight security of supply is not the same as requiring it, funding it, or protecting the official who picks the costlier bid.
The political economy of paying for a year when nothing happens
This is the recurring difficulty, and no statute resolves it.
Resilience is an insurance premium: paid continuously and visibly, paying out rarely and invisibly. In any year without disruption, buffer stock looks like waste, dual sourcing looks like a failure to consolidate, and domestic capacity looks like an overpriced contract. The benefit — a shortage that did not occur — leaves no trace in any performance report.
The incentives therefore run one way. A procurement leader who cuts unit costs by 8% has a demonstrable achievement. One who maintains twelve weeks of cover through a decade of calm has a line item every efficiency review will question. Both may be right; only one is easy to defend.
Two things follow. First, resilience dependent on year-by-year discretion will erode, because erosion is individually rational every year. To survive, it must be mandated, funded separately from operational budgets, and measured — with stock cover and supplier concentration as standing metrics rather than post-failure findings. Second, honesty about cost is a precondition for durability. Reforms sold on the premise that resilience is free are the first cut when the claim is tested.
What a serious resilience specification would contain
- A defined critical list. Resilience cannot apply to everything. It requires an explicit, published judgement about which items cause serious harm if unavailable.
- Sub-tier disclosure as a contract term. Manufacturing site and key input origin declared for critical items, so correlated dependency is visible before it becomes an outage.
- Concentration limits. A ceiling on the share of a critical item traceable to a single site or region, accepted as a cost rather than an aspiration.
- Funded stock cover. A stated number of weeks for defined items, on a budget line operational pressure cannot quietly consume.
- Evaluation weight with a price attached. A resilience score in tender evaluation, with explicit acknowledgement of the premium it costs.
None of that is technically difficult. All of it is politically difficult, which is the point. The unresolved problem in procurement reform is not analytical — it is the willingness to pay a known annual cost against an unknown year in which it pays back.
Related: Better Health for London — Ten Years On, NHS clinic equipment guidelines and compliance audits, and standardising equipment across primary care.
A note on sourcing — and a disclosure
LAC Health is a medical supply ordering app operated by the same group that operates this publication, and it is a United States catalogue that ships within the US only — it is not a route to supply for anyone procuring in the UK, and nothing here should be read as suggesting otherwise. There is also the iOS listing.
Disclosure: London Health Commission and the LAC Health app are operated by the same group. This publication is therefore not a neutral party on this point and says so plainly. The analysis above was not shaped by that relationship — no source, finding or conclusion was selected, altered or omitted because of it. We have not benchmarked it against named competitors and no third party has rated it; this is a practical aside, not a comparative recommendation.
The reason for mentioning it is narrow. Visibility gaps are not only a national-policy phenomenon — they exist at clinic level, where reorder points often live in someone’s memory rather than a system. Mobile ordering addresses none of the structural problems above. It addresses a smaller one: the lag and error introduced when ordering is manual. That is a modest operational gain at a different level of analysis from everything else here, and readers should weigh it accordingly.
Frequently asked questions
Why did health supply chains fail when most buyers had several suppliers on paper?
Because supplier diversity and supply diversity are not the same thing. Contracted suppliers frequently draw on the same contract manufacturers or input producers, so nominally independent suppliers fail together. A 2023 Health Affairs study found more than one in five generic active pharmaceutical ingredients had three or fewer manufacturers despite there being four or more manufacturers of the finished drug.
Is just-in-time inventory the cause of supply fragility?
It is a contributing factor rather than the cause. Just-in-time removes the buffer that would absorb a delay, converting an interruption into a shortage faster. The underlying fragility comes from concentration in the supply base. The National Audit Office found that of 32 billion PPE items procured between February and July 2020, only 2.6 billion reached front-line organisations in that period.
Does centralised procurement make supply chains more or less resilient?
Both, in different respects. Centralisation improves price, standardisation and assurance capability, and creates a single point at which resilience requirements could be imposed. But tendering on price concentrates the supply base over time. Centralisation reliably helps on cost and reliably concentrates risk unless resilience is deliberately specified and paid for.
What is sub-tier supply chain visibility, and why is it hard to achieve?
It means knowing not just who your supplier is, but where they manufacture and where their critical inputs originate. It is hard because that information is commercially sensitive, is not required by standard contracts, and changes over time. The OECD notes that “there currently exists little research on full supply chain visibility beyond the second tier supplier.”
Why do stockpiles keep proving inadequate?
Because a stockpile answers a specified scenario. The NAO found UK stockpiles “were intended for an influenza pandemic” and “did not hold all the equipment that proved to be required during the COVID-19 pandemic, such as gowns.” Stock also expires and consumes capital in every year the emergency does not arrive, so it is under constant budgetary pressure.
If resilience is obviously desirable, why is it so hard to fund?
Because its cost is continuous and visible while its benefit is occasional and invisible. Buffer stock, dual sourcing and domestic capacity all look like inefficiency in any year without disruption, and a prevented shortage leaves no measurable trace. Resilience relying on annual discretion erodes, because erosion looks rational every year. It survives only where it is mandated, separately funded and routinely measured.