In clinical microbiology this
problem of resolution at boundaries was recognized early
on and was addressed by introducing the concept of the
‘resistance breakpoint’. There are two different breakpoints
commonly used in clinical microbiology: the
‘clinical breakpoint’ that predicts therapeutic outcome,
and the ‘epidemiological breakpoint’ that acts as an early
warning system of changes in the frequency of isolates
with increased resistance in the wild-type population
[40,41]. These breakpoint systems, being based on
MIC, implicitly recognize that there are problems in
measuring resistance. One problem is that the measurements
are inherently crude, and measuring MIC repeatedly
on the same sample gives a distribution of values that
can only locate MIC (which is defined as a step function)