Interferon therapy is used widely for chronic hepatitis C but only a minority of treated patients achieve a long-lasting sustained response. We have developed, by logistic regression, a mathematical model to estimate the probability of sustained response in an individual patient with chronic hepatitis C when treated with interferon-alpha (IFN-alpha). The model, which includes age, sex, disease duration, pretreatment serum gamma-glutamyl-transpeptidase, alanine aminotransferase and virus genotype, was developed from a database of 307 patients and validated in a new set of 200 patients. It performed well as goodness-of-fit (P = 0.71 and P = 0.15 in the development and test sample, respectively) and discrimination (area under receiver operating curve = 0.79 in the development and 0.78 in the test sample, respectively). This model may provide decision support in the treatment of chronic hepatitis C with IFN-alpha.