Fuzzy-logic based automatic control of hemodynamics

We propose an automatic closed-loop-system for hemodynamic control on the intensive care unit in patients after cardiac surgery. All possible combinations of four standard hemodynamic parameters (mean arterial pressure, on-line cardiac output, central venous pressure and systemic vascular resistance) and four treatment regimens (fluid administration, noradrenaline, dopamine and nitrates) were linked by knowledge-based rules in a fuzzy-logic system. After defuzzification, the loop was closed by integration of automatic control of infusion and syringe pumps. The system proved to be able to observe and handle any cardiocirculatory disturbances in circulatory models and animal experiments. The system remained stable throughout any observation period. Hemodynamic changes and therapeutic interventions were documented in an electronic experimental record. In several attempts for automatic therapy control in the past, failures were mainly due to insufficient computer capacity, inadequate sensors and the lack of appropriate mathematical models. We anticipate an improvement of postoperative patient care by induction of therapies and paperless on-line documentation permitting to spend more time to treat the critically ill.