We present a new approach to the design of a decision support system (DSS) in anesthesia which converts the available data to relevant information. Instead of a patient-driven design (patient modelling), we use a decision-driven design (anesthetist modelling). This approach results in a system consisting of three stages. First, the incoming data is validated to ensure reliable further processing for both DSS and anesthetist. Second, the validated data is analyzed to detect patterns that could trigger the anesthetist's response. Finally, from these patterns a strategic selection is made and offered to the anesthetist as information relevant for the decision that has to be made in the current context. Results show that this approach is feasible, but further evaluation is necessary before practical application is possible.