A rule-based postoperative pain controller: simulation results

We describe a system for controlling postoperative pain, a phenomenon that is difficult to treat by conventional control methods due to interpatient variability, interferences, non-linearity and the lack of a plausible, well-defined mathematical model. The system consists of two phases. In Phase 1 a closed-loop fuzzy controller implementing a suitable control strategy brings the patient to a zero-pain state. In Phase 2, an open-loop computer-assisted continuous infusion controller maintains a constant concentration of the analgesic (alfentanil) in plasma, subject to an upper safety limit on infusion rate; the set-point of this controller is periodically revised (either maintained or reduced) on the basis of feedback on the duration of zero pain (set-point reduction is necessary because the open-loop system has no means of knowing whether analgesic is accumulating in the patient). Pain is quantified by the patient on a numerical scale of 1 to 10 at 1.5-min intervals during Phase 1 and 9-min intervals during Phase 2. In simulation trials in which a fixed approximate model was used for the effect of sedation on pain while the pharmacokinetics of alfentanil were varied from one simulated patient to another, zero pain was achieved in under 15 min with minimal overshoot in plasma drug concentration and was maintained, with only minor deviation, by means of low drug concentrations.