This project currently includes three studies each of which is a component of the Stroke Data Bank or its precursor, the Pilot Stroke Data Bank. The studies are: (1) Evolving Stroke. Using demographic, history, clinical and laboratory data, this study describes the temporal course of stroke-in-evolution and attempts to identify factors that cause or contribute to evolution. (2) Prognostic factors for 30-day mortality. Multiple logistic regression models, one for ischemic stroke and another for intracerebral hemorrhage were used to determine prognostic factors for 30-day mortality. Logistic models were derived using data from the pilot project: 620 ischemic strokes with 52 deaths and 94 intracerebral hemorrhages with 32 deaths. Potential risk factors (112 in all) were initially screened by univariate statistical methods and those screened positive were examined multivariately in the logistic model. These derived models will be cross-validated by examination of their predictive ability on data from the current Stroke Data Bank. (3) Discrimination between intracerebral hemorrhage and ischemic stroke. This study identifies factors, excluding CT information, available shortly after stroke onset that provide optimal classification into the two diagnostic categories.
