Closed-Loop Anesthesia

Automating Anesthetic Processes

My most recent project was for my control systems class.  I wanted do something original, and I was reminded of my days working as an orderly in an operating room at the nearby hospital.  My observations of the anesthesiologists were that much of what their work entailed processes that could easily be controlled by a computer.  In the words of my professor, “anesthesiology is so open loop rn.”  So I decided to build a program that would automate anesthesia, designed for a very simple system.

The model I used is primarily based on a two stage system that predicts the effects of Propofol in a patient.  The first stage is concerned with the pharmacokinetics (PK), which models how the concentration of Propofol varies in the body as the drug is metabolized.  The second stage is concerned with the pharmacodynamics (PD), which models the plasma in drug concentration to clinical-site effect.  This PK/PD model is represented in its state-space form below in Equation 1.

The Hill Sigmoid equation, as as seen in Equation 2 below, is a widely used technique for relating the wakefulness of a patient to the effect-site concentration.  The wakefulness of a patient is measured on a scale from 0 to 1, where 0 corresponds to a complete state of coherency, and 1 represents the patient being in a coma.  The ideal depth of hypnosis occurs at a wakefulness of 0.5.

With these governing equations, I built a PID controller that would successfully bring a patient to a wakefulness of 0.5, while also taking into account potential disturbances.

In order to obtain the results above, I had to make various simplifying assumptions.  This current model breaks for adult patients, and would only superficially work for children.  This project was meant to be a fun demonstration of the power of control systems.

Feel free to contact me with any questions