We designed a randomized clinical trial to determine the efficacy of plasma exchange in Guillain-Barre syndrome, an acute neurological disease leading to paralysis in previously healthy subjects. Some patients require temporary mechanical ventilation during the course of the disease. Ventilatory status was analysed using a three-state model with a transient state defined by use of temporary mechanical ventilation, and the
probability of being mechanically ventilated modelled as a function of time since randomization. The plot of this function reflects the effect of plasma exchange on ventilatory status showing that fewer plasma-exchanged patients than controls required ventilatory assistance and that the duration of assistance was shorter. Similar functions can be used in a more general framework when the response criterion includes a transient state.