Patients suffering from a certain disease generally differ greatly regarding the course of their disease. The identification of factors from which much of this variability between patients can be explained may be of great importance. The study of such prognostic factors in clinical medicine is usually done for a variety of reasons which to some extent may overlap each other. One of the reasons may be to better understand how the disease is likely to behave. It is hardly ever possible nowadays to study the pure natural history of a disease because usually some form of treatment will have been undertaken. The treatment often will be curative in intent, but may also be directed towards the relief of symptoms or avoiding or delaying these in case no curative treatment is available. It is known from their natural history that certain illnesses do not require treatment at all because they are self-limited processes. Knowledge about prognostic factors in such cases will make it possible to provide information to the patient about the probable duration and course of the disease. Knowledge of factors which are related to the outcome of disease may be helpful to physicians in understanding the mechanism of the disease. Such better understanding may assist in modifying treatment protocols, searching for new treatments or fonnulating strategies for the optimal use of expensive medical tests. Sometimes knowledge of prognostic factors may be of use to alter the course of disease by remedial action. This will particularly be so if these factors are elements of lifestyle such as diet, exercise or habits. For instance, patients with alcoholic liver cirrhosis who stop taking alcohol appear to have a better prognosis regarding survival as compared to those who continue drinking alcohol [Borowsky et ai, 1981]. Such scientifically based facts may be helpful in encouraging patients to change their drinking habits. Knowledge of prognostic factors often will be of importance in planning therapeutic trials. It may be desirable to evaluate certain therapies only in patients who belong to certain prognostic categories. Also the number of patients required in a randomized clinical trial depends on the distribution of prognostic factors among patients to be included in the study.