Parametric modeling of stroke recurrence.

Stroke recurrence has been investigated primarily with respect to prognostic factors predictive of recurrence. Several parametric functions are considered in modeling the distribution of ischemic stroke recurrences recorded within the Stroke Data Bank. A linear hazard function is shown to be the best-fitting function among those considered. This method of parametric modeling may lead to a more informed approach to treatment of ischemic stroke and secondary prevention and may enhance future investigations of prognostic factors as well.