There is little consensus about treatment for the diseases caused by the common non-tuberculous mycobacteria (NTM). Guidelines are based largely on retrospective, non-controlled studies. Where susceptibility testing data is available, in vitro testing often correlates poorly or not at all with the clinical response to treatment. Performing non validated susceptibility testing is likely to confuse treatment rather than aid it. In the clinical situation, the only valid indication for performance of susceptibility testing is if the test will produce a result that will help predict treatment outcome. A ?susceptible? result should predict treatment success with that drug. Conversely, the detection of resistance in a susceptibility test should predict treatment failure.