According to the World Health Organization, a stroke has been the second most common cause of death in the world in the last 15 years. An ischemic stroke accounts for almost 80 % of all cases. The University Hospital Ostrava in the Czech Republic collects various information about patients who were transported there after suffering from an acute ischemic stroke, such as the affected brain hemisphere, duration of medical procedure or presence of hypertension. The objective of this paper was finding a model which would be able to predict patient's clinical outcome three months after an ischemic stroke based on the collected data. It was also desirable to analyse importance of the considered variables. For this purpose, the random forests algorithm was used. To avoid biased variable importance, we used an alternative approach to the random forests which uses the conditional inference trees. Firstly, the commonly used modified Rankin Scale was used for describing the patient's outcome three months after a stroke. Secondly, only two values for the clinical status were considered, by meaning they correspond with the values 0–3 and 4–6 of modified Rankin Scale. The best performance was achieved with the second approach to description of the clinical outcome with the calculated classification accuracy 86 %.