Abstract Cough provocation testing was originally developed for testing new anti-tussive agents and still remains a mandatory tool in their development. The methodology has been developed into single breath and multiple breath techniques. Neither has been fully validated; however, single breath techniques have been used in a number of patient series of more than 100 patients. These studies show that the cough reflex is more sensitive in patients with cough and becomes normal on recovery. Studies in the common cold and in angiotensin converting enzyme inhibitor-associated cough suggest that this may be cause and effect. These results point to a utility for cough provocation testing in epidemiology but not in routine clinical practice. The one exception in clinical practice is in the assessment of patient post stroke where it is a better method, than voluntary cough testing, for assessing damage to the cough reflex. Other than use in epidemiology it remains an invaluable tool in hypothesis testing.