Event-based control of postoperative analgesia using physiological nociception signals

This paper presents an event-based dual-loop analgesia control strategy for communicative patients in the Post-Anesthesia Care Unit (PACU). The work is a simulation-based proof-of-concept and preliminary feasibility study, evaluated using three reconstructed patient cases and Monte Carlo-generated synthetic patients. The approach combines an impedance-derived objective nociception index, (ΔT), obtained from the Anspec-Pro device, with sparsely sampled subjective Numerical Rating Scale (NRS) pain scores. The dynamic relationships from drug administration to (ΔT), and from (ΔT) to NRS, are modeled using a first-order plus time-delay (FOPTD) model with integral action and only a FOPTD model, respectively. Model parameters and delays are identified from postoperative patient data using least-squares estimation and correlation-based delay identification. The identified models reproduce the measured equivalents. The identified parameter distributions across patients follow log-normal distributions, enabling population-level Monte Carlo simulations. Population-average models are used to tune an inner proportional–integral–derivative controller with a robustness margin of (0.7) regulating (ΔT), while an outer event-based proportional-integral controller adjusts the setpoint when NRS exceeds a pain threshold. Robustness analyses using 1000 virtual patients show that the proposed closed-loop methodology maintains reported pain levels within the desired clinical range for 98.97% of patients without disturbance and for 89.78% of patients under a hypothetical disturbance scenario. These results indicate that the proposed dual-loop scheme can robustly anticipate and limit pain excursions, detect pain onset prior to patient self-reporting, and respond more rapidly to provide comfort, thereby suggesting its feasibility as a decision-support tool in the PACU.