ABSTRACT
Rowthorn and Brown (Chapter 2) and Laxminarayan and Weitzman (Chapter 3) characterize the socially optimal way to treat one kind of bacterial infection in a population when multiple antibiotics are available. The first of these formulations is static, whereas the second is dynamic. In both formulations, choosing the antibiotic best suited for each patient in isolation is not socially optimal because such a policy disregards the effects such a choice imposes on third parties. When the third-party effects of such choices are taken fully into account, it often turns out—for different reasons in the two formulations—to be optimal to treat the same type of infection in different individuals at the same time with different antibiotics.
