Clinical outcomes prediction in rudimentary form began as clinical observations of associations between single characteristics and pertinent outcomes. Multivariate modeling in intracerebral hemorrhage (ICH) has focused on determining outcomes and examining the independent effects of specific characteristics (e.g., intraventricular hemorrhage (IVH)) that could help explicate pathophysiological mechanisms and identify potential targets for intervention. Prognostic models have fostered the development of prognosis-based clinical trial methodology in which prognostic models are used to stratify patients. Models can be used to provide a sophisticated historical comparison for data collection in observational studies. Models are also used to define patient groups suitable for specific clinical trials and help to define relevant endpoints that can be prespecified for a particular group according to their expected outcome. Finally, mathematical outcome models have been used to identify specific findings or other characteristics that may affect outcome and be targets for intervention.