A prototype system is being designed to provide advance warning of adverse respiratory and circulatory conditions in postanesthesia patients. The system is designed to minimize artifacts in the database, validate data for epidemiologic studies, and identify variables that predict adverse events through the application of artificial intelligence techniques and statistical models. The system is of modular design. The modules include the patient database, statistical model, interface to monitoring devices, rules for evaluating the monitor data, clinical knowledge, and advice. >