What antibiotic resistance actually is, and what drives it

The most common misconception about antibiotic resistance is that a person becomes resistant to antibiotics. They do not. It is the bacteria that become resistant, and the property belongs to the microbe, not the patient. That distinction is the key to understanding both why resistance spreads and why an individual's choices matter less than the collective ones — a resistant strain that evolves in one person, one farm, or one hospital can travel anywhere.

Resistance is, at bottom, ordinary evolution running fast. Bacteria divide in minutes, and any population large enough will contain rare variants that survive a given drug — through a random mutation that alters the drug's target, a pump that expels the antibiotic, or an enzyme that dismantles it. Expose the population to the antibiotic and the survivors are the ones that go on to reproduce. What makes bacterial resistance move faster than textbook evolution is that the genes for it are often not confined to a lineage. They ride on plasmids and other mobile genetic elements that one bacterium can pass to another, including across species, so a resistance trait can jump between unrelated bugs rather than only being inherited. The World Health Organization frames the whole phenomenon this way: antimicrobial resistance is a natural process accelerated by how the drugs are used [s2].

The scale, measured directly

For a long time the burden of resistance was asserted rather than counted. The first comprehensive attempt to measure it, the Global Research on Antimicrobial Resistance (GRAM) study published in The Lancet in 2022, estimated that in 2019 bacterial antimicrobial resistance was associated with 4.95 million deaths worldwide, of which 1.27 million were directly attributable to it [s1]. The two numbers answer different questions: "associated" counts deaths from drug-resistant infections, while "attributable" estimates how many of those deaths would have been avoided if the infections had been drug-susceptible instead [s1]. On the attributable figure, resistance already ranked among the leading causes of death globally in 2019, comparable to major established killers [s1].

The study also showed where the problem concentrates. A small number of pathogen-drug combinations drove most of the deaths — resistant E. coli, Staphylococcus aureus, Klebsiella pneumoniae, Streptococcus pneumoniae and a handful of others — and the burden fell hardest on regions with the least access to newer antibiotics and diagnostics [s1]. Resistance is often imagined as a rich-world hospital problem; the data put its heaviest toll in low-resource settings.

What actually drives it

The drivers are less exotic than the biology suggests. The WHO points to the misuse and overuse of antimicrobials in people, animals and plants as the main accelerants, alongside poor infection prevention, inadequate sanitation, and limited access to clean water and quality medicines [s2]. Two points inside that list are worth drawing out because they are widely misunderstood.

The first is agriculture. A large share of the antibiotics produced worldwide is used in food-producing animals, often not to treat sick animals but to promote growth or prevent disease in crowded conditions — and resistant bacteria and resistance genes selected there can reach people through food, water and the environment [s2]. Individual prescribing habits are only one tap feeding the same basin.

The second is the folk rule to "always finish the course." An analysis in The BMJ in 2017 argued that the evidence behind that instruction is weaker than assumed, and that for many infections there is no good evidence that stopping when you feel better drives resistance — while longer-than-needed courses expose more bacteria to the drug and may do the opposite of what the rule intends [s3]. The authors were careful: the answer varies by infection, some conditions genuinely require a fixed duration, and no one should self-adjust a prescription [s3]. The wider point they make is that the "complete the course" message was a public-health simplification, not a settled finding, and that the right course length is a question for evidence and for the prescriber, not a moral instruction to the patient.

What it means for a patient

For an individual, resistance is not abstract. It is the reason a routine urinary or wound infection occasionally does not respond to the first antibiotic tried, forcing a switch to a second or third drug that may be more toxic, need to be given by drip, or work more slowly — and, in the worst cases, the reason an infection that would once have been readily treatable becomes hard to treat at all [s1][s2]. The most consequential thing any one person can do runs against the instinct to demand a prescription: because the resource being protected is the drugs' future effectiveness for everyone, antibiotics taken when they cannot help — for a viral cold, for instance — carry the downside of resistance and adverse effects with none of the benefit [s2]. What any specific infection needs is a clinical judgement, and this explainer describes the forces at work rather than what to take.