In many cases, physicians prescribe antibiotics without knowing whether an individual patient is infected with a susceptible or resistant pathogen. As the proportion of resistant organisms in a community increases, physicians substitute away from older-inexpensive drugs to newer, more expensive agents as first line therapy. This paper explores the implications of resistance-induced antibiotic substitution for epidemiological models to predict future resistance levels, efforts to measure the health care costs associated with resistance, and policies to improve physicians' antibiotic prescribing decisions. The extent of resistance-induced substitution in outpatient settings is documented using a data set consisting of observations on initial physician office visits for otitis media in the US controlling for new product introductions and price increases, per prescription antibiotic spending increased by 22% between 1980 and 1996, corresponding to a steep increase in resistance levels over the same period.