PID control of depth of hypnosis in anesthesia for propofol and remifentanil coadministration

Tese de mestrado, Engenharia Biomédica e Biofísica, 2022, Universidade de Lisboa, Faculdade de CiênciasThe purpose of general anesthesia is to deeply sedate a person so that they lose consciousness, sensitivity, and body reflexes, and so that surgeries can be safely performed without the patient feeling pain or discomfort during the procedure. General anesthesia is a combination of the effect of three components, namely hypnosis, analgesia, and neuromuscular blockade. Each component is regulated through the action of a specific drug, or through the combined effect of two or more drugs. In recent years there have been many advances in the field of automatic control systems for drug delivery during anesthesia, which can be implemented using a wide variety of controllers and process variables. The reason behind these advances is that an automatic control system can provide several benefits, such as a reduction in the anesthesiologist's workload, a reduction in the amount of medication used (which implies a faster and better recovery time for the patient in the postoperative phase), and, in fact, a more robust performance with fewer episodes of over- or under-dosing of the drug. A proportional-integral-derivative controller (PID) continuously calculates the error value that is the difference between the desired value and the measured process variable and applies a correction that is based on proportional, integral and derivative terms. In this dissertation, a specific PID control system for propofol and remifentanil is proposed to regulate the hypnosis component during anesthesia using the bispectral index (BIS) as the process variable. Infusion rates of both drugs are also controlled. The adjustment of the PID parameters, so that the BIS was closer to what was expected, was done using a genetic algorithm. The implementation of the control system was done in Simulink in order to simulate a surgery. The simulation scheme includes the patient models for both drugs, a disturbance profile, and two different PID controllers for the two phases of anesthesia - induction and maintenance. Aspects such as noise in the BIS signal and artifacts were taken into account in the system and a suitable noise filter was applied in the control algorithm. In addition, a ratio between the infusion rates of propofol and remifentanil has been introduced to allow the anesthesiologist to choose the appropriate opioid-hypnotic balance In the end, a performance analysis of the control system was made based on seven performance indices (namely the integrated absolute error, the settling time, the median performance error, the median absolute performance error, the wobble, and the above and below recommended BIS values). Although there are many types of control systems for the automatic control of hypnosis depth described in the literature, these are not usually used in clinical practice. Therefore, it is important to continue research to produce robust and user-friendly systems that integrate clinicians' clinical knowledge and meet their actual needs