Antimicrobial resistance
How antibiotic use turns a sensitive bacterium into a resistant one across a population, and what stewardship can do. Resistance wins when the pressure from treatment outweighs its fitness cost.
How much less the resistant strain spreads without treatment.
More options
Resistant share at the end
3%
After 10 years, of everyone carrying the bacterium.
Tipping point
20%
Treat a larger share of episodes than this and resistance takes over in the long run.
Resistance passes 50%
after 2.1 years
Long-run winner
Sensitive strain
Without stewardship
100%
Resistant share of carriage over time
For your protocol
A two-strain carriage model (sensitive and resistant; basic reproduction number 2, mean carriage 30 days, resistance fitness cost 20%) with 30% of sensitive carriage episodes ended by antibiotic treatment, reduced by 50% from year 3 through stewardship, projects the resistant share of carriage to move from 2% to 3% over 10 years, passing 50% after about 2.1 years. In this model resistance takes over in the long run when more than 20% of episodes are treated (the fitness cost). It is a teaching model with homogeneous mixing and no co-colonisation, within-host dynamics or compensatory evolution.
How this simulation works
A two-strain model of carriage in a well-mixed population (after Bonhoeffer, Lipsitch and Levin, 1997). Carriers of the sensitive strain transmit at rate β and clear naturally at rate γ (one over the carriage duration); treatment clears them at an extra rate τ, set from the share of episodes treated. The resistant strain is untouched by treatment but transmits less (by the fitness cost). A small chance of resistance emerging with each course can be added. In this model the resistant strain takes over in the long run exactly when the treated share exceeds the fitness cost. Solved with fourth-order Runge–Kutta. A teaching model: real systems have co-colonisation, several drugs, compensatory mutations and structured contact.