Paroxysmal atrial tachycardia may be useful as a model for designing clinical trials of paroxysmal tachycardia. Individual occurrences of paroxysmal atrial tachycardia are clinically independent events, and the time periods between events follow an exponential probability distribution; i.e. paroxysmal atrial tachycardia behaves like a Poisson process. We propose that treatment of paroxysmal atrial tachycardia can be studied efficiently using a clinical trial with a two period cross-over design in which placebo and investigational therapy are delivered in random order. A screening phase, dose ranging phase, and both limbs of a cross-over plane can be completed in 180 days or less in each patient.