The Global Antibiotic Resistance Crisis As A Forced Cooperation Problem

Let's get the biology straight before we get to the politics.

Bacteria reproduce fast. E. coli can divide every twenty minutes under ideal conditions. In eight hours, a single bacterium can become sixteen million. Every division is an opportunity for mutation, and mutations happen constantly.

Most mutations are neutral or harmful. But when you bathe a bacterial population in an antibiotic, you create selection pressure. The bacteria with mutations that happen to resist the drug survive. They reproduce. Their resistant offspring become the dominant strain. This is not a theory. It is observed, repeatable, Tuesday-afternoon-in-a-petri-dish science.

Worse, bacteria can share resistance genes horizontally through plasmids — small loops of DNA that transfer between different species. A resistance gene that evolved in Pseudomonas can end up in Klebsiella inside your gut within a week. This is why resistance to one drug in one bug can cascade into resistance across dozens of bacteria types. The ecosystem rewires itself.

The specific strains that keep WHO microbiologists up at night:

- MRSA (methicillin-resistant Staphylococcus aureus): skin and soft-tissue infections, hospital-acquired pneumonia - CRE (carbapenem-resistant Enterobacteriaceae): bloodstream infections with mortality rates up to 50 percent - MDR-TB and XDR-TB: multidrug- and extensively drug-resistant tuberculosis - Drug-resistant gonorrhea: now resistant to nearly every antibiotic historically used to treat it - Candida auris: a fungus, technically, but following the same selection logic

According to the Lancet's 2022 Global Research on Antimicrobial Resistance (GRAM) study, AMR was directly responsible for 1.27 million deaths in 2019 and associated with 4.95 million deaths. That's already more than HIV/AIDS and malaria combined. And that's the baseline before the projected curve bends up.

The discovery of penicillin by Alexander Fleming in 1928 kicked off what we now call the golden age of antibiotics. Streptomycin (1943), tetracycline (1950s), vancomycin (1958), quinolones (1960s), daptomycin (1986). Then — silence.

The last novel class of antibiotic to reach clinical use was the oxazolidinones, approved in 2000 but discovered in the late 1980s. Everything since has been a modification of existing classes. That's like if the only new vehicles for forty years were variants of the 1980 Toyota Corolla.

Why did this happen? Three reasons, and it's worth sitting with each:

Economics. A cancer drug taken daily for years generates billions in revenue. An antibiotic taken for ten days generates thousands per patient. Worse, the best antibiotics get held in reserve (rightly, to slow resistance), which means low sales volume. Nobody wants to develop a product you're actively told not to use.

Scientific difficulty. The easy soil microbes — the ones producing natural antibiotics we could isolate — got mined out by the 1970s. What's left requires deep chemistry, often in molecules that don't penetrate bacterial membranes or get pumped out by efflux pumps. The frontier got harder exactly as the incentives got worse.

Regulatory friction. Clinical trials for antibiotics are weird. You can't ethically test a new drug against a placebo for a life-threatening bacterial infection. You need non-inferiority trials against existing drugs, which require huge patient populations with specific resistance profiles, which are hard to recruit.

The result: in 2019, the Pew Trusts counted only 42 antibiotics in clinical development globally. Compare that to roughly 1,800 oncology drugs in development. An order of magnitude gap in a field where we're losing ground.

Small biotechs like Achaogen (bankrupt 2019, four months after FDA approval for plazomicin) have shown that even getting a drug approved doesn't generate enough revenue to survive. The market has spoken, and it said: let people die.

Now go to the farm.

The United States uses roughly 11 million kilograms of antibiotics in food animals each year. China uses more than 80 million kg — nearly half the global livestock antibiotic supply. Humans, globally, use maybe 40 million kg.

Most of this agricultural use is not for treating sick animals. It's for growth promotion and prophylaxis — continuous low doses in feed and water to make animals grow faster on less food and to prevent disease in crowded industrial operations.

This is a laboratory for breeding resistance. You keep the dose below the killing threshold, which means you select for mutants that tolerate the drug without eliminating susceptible bacteria. You run this experiment 24 hours a day across hundreds of thousands of animals for decades.

What comes out:

- Resistant bacteria in meat (studies find resistant E. coli in roughly 80 percent of supermarket chicken in multiple countries) - Resistant bacteria in manure, which fertilizes crops - Antibiotic residues in groundwater - Transfer to farm workers, who carry resistant strains into their families and communities

When colistin — a last-resort antibiotic for humans — started failing in the 2010s, researchers traced the resistance gene (mcr-1) back to Chinese pig farms, where colistin had been used as a growth promoter for years. The gene has since been found in 40+ countries.

The EU banned antibiotic growth promoters in 2006. The US FDA issued voluntary guidance in 2017 (Guidance 213) that effectively ended most growth-promotion uses. But India, China, Brazil, and most of Southeast Asia and sub-Saharan Africa still use antibiotics in livestock with minimal oversight.

One country's restraint doesn't matter much if another country's farms are downstream.

The One Health framework — formalized by a tripartite agreement between the WHO, FAO, and OIE (now WOAH) in 2010 — is the intellectual bedrock of any real AMR response. It states, simply: human health, animal health, and environmental health are inseparable.

In practice, that means:

- Veterinary antibiotic use must be included in surveillance - Wastewater treatment becomes a public health intervention (we already have data showing hospital and pharmaceutical plant effluent is a major reservoir of resistance genes) - Agricultural ministries and health ministries must actually talk to each other - Trade agreements must include AMR provisions

The WHO's Global Action Plan on AMR, adopted in 2015, required every member state to submit a National Action Plan. By 2023, roughly 170 countries had submitted plans. The problem: most plans are unfunded. A paper in the Lancet in 2022 found that only about 20 percent of National Action Plans had dedicated budgets.

Having a plan and funding a plan are different species.

The COVID-19 pandemic made one thing clear: pathogens don't wait for you to negotiate. The WHO Pandemic Agreement (negotiated 2021–2025, adopted in some form by the World Health Assembly) tried to codify what sovereign states owe each other during a biological emergency.

AMR is a slower pandemic. Same logic applies. A country that doesn't fund its surveillance is a country that exports blind spots. A country that doesn't regulate its pharmacy market is a country that manufactures future resistance. A country that subsidizes factory farms without antibiotic rules is a country that breeds the next untreatable infection.

The proposal — championed by the UN Interagency Coordination Group and various AMR scholars — is for a binding international AMR treaty. Something with teeth. Something that obligates wealthy countries to fund R&D pull incentives, obligates middle-income countries to regulate veterinary use, obligates low-income countries to accept surveillance support, and provides mechanisms for equitable access to new drugs when they arrive.

Whether this treaty happens in the next decade or not is one of the real tests of whether the species can coordinate on slow-moving existential threats.

A concrete proposal, since we should be specific.

The UK's NHS ran a pilot called the Subscription Model starting in 2020. Instead of paying per prescription, the NHS pays a flat annual fee (up to £10 million per drug) for guaranteed access to two new antibiotics. This delinks revenue from volume. A pharmaceutical company making a new antibiotic knows it will earn a steady income whether the drug is used once or ten thousand times.

The PASTEUR Act (Pioneering Antimicrobial Subscriptions To End Upsurging Resistance), introduced in the US Congress, would do something similar at roughly $6 billion over ten years. It's been introduced multiple times, never passed.

Calculations by the Duke-Margolis Center for Health Policy suggest global pull incentives in the range of $3–4 billion per drug would restart the pipeline. That's real money, but it's pocket change against $100 trillion in projected GDP losses.

This is the kind of coordinated bet that only makes sense if multiple countries agree to fund it together. One country doing it subsidizes free riders. Everyone doing it shares the cost and shares the benefit.

Classic collective action problem. Classic Law 1 moment.

- O'Neill, J. (2016). Tackling Drug-Resistant Infections Globally: Final Report and Recommendations. The Review on Antimicrobial Resistance. - Murray, C. J. L., et al. (2022). Global burden of bacterial antimicrobial resistance in 2019: a systematic analysis. The Lancet, 399(10325), 629–655. - WHO (2015). Global Action Plan on Antimicrobial Resistance. - Van Boeckel, T. P., et al. (2015). Global trends in antimicrobial use in food animals. PNAS, 112(18), 5649–5654. - Laxminarayan, R., et al. (2013). Antibiotic resistance — the need for global solutions. The Lancet Infectious Diseases, 13(12), 1057–1098. - Pew Trusts (2019). Antibiotics Currently in Global Clinical Development. - UK Department of Health and Social Care (2022). UK 20-year vision for antimicrobial resistance: 2024–2029 action plan.

1. Audit your own medicine cabinet. How many antibiotics have you taken in the last five years? How many were prescribed for viral infections (colds, flu, most sore throats)? Antibiotics do nothing to viruses, yet roughly 30 percent of prescriptions in the US are estimated to be unnecessary.

2. Trace one meal. Pick one meat product from your last grocery run. Look up the country of origin and its antibiotic regulations for livestock. Notice how little information is available.

3. Write the letter. Draft a letter to your national health ministry asking what percentage of the country's AMR action plan is funded. Send it. Post the response publicly.

4. The cooperation question. List five things your country could do on AMR that only make sense if three other countries also do them. Then ask: what mechanism would force that coordination?

5. The microbe exercise. For one day, notice every moment you trust that a small cut won't kill you. Every moment of that trust is built on antibiotics still working. What are you personally doing to keep that trust earned?

AMR is not a technical problem. The technical answers exist. AMR is a coordination problem — the kind that only gets solved when enough people understand they are part of a single system.

Bacteria are teaching us that borders are a fiction. The question is whether we learn the lesson before the body count confirms it.