Surveillance work has generally let us know whether a particular compound is still good for curing disease. The answer is yes as long as a very high percentage of the isolates are considered susceptible. The data can also be used to warn us of upcoming resistance, but the system is not very good for that. There are two reasons. First, when one uses susceptibility breakpoints to determine whether a strain is resistant or susceptible, a strain can have many resistance mutations and still be considered susceptible. If those mutations accelerate the development of resistance, the surveillance work will give you a false sense of security. Second, the primary data for surveillance is generally derived only from the dominant members of the infecting population. This is because clinical labs generally only look at well-isolated colonies. Consequently, the growth of mutant subpopulations will not be observed. The slide shows what is called creeping resistance. With time the curve shifts to higher concentrations. The result is that resistance is thought to appear suddenly when it actually develops gradually.