Estimation of optimal modeling weights for a Bayesian-based closed-loop system for propofol administration using the bispectral index as a controlled variable: a simulation study

Abstract

Background: Implementing Bayesian methods in a model-based closed-loop system requires the integration of a standard response model with a patient-specific response model. This process makes use of specific modeling weights, called Bayesian variances, which determine how the specific model can deviate from the standard model. In this study we applied simulations to select the Bayesian variances yielding the optimal controller for a Bayesian-based closed-loop system for propofol administration using the Bispectral Index (BIS) as a controlled variable.

Methods: The relevant Bayesian variances determining the modeling process were identified. Each set of such Bayesian variances represents a potential controller. The set, which will result in optimal control, was estimated using calculations on a simulated population. We selected 625 candidate sets. Similar to our previous closed-loop performance study, we applied a simulation protocol to evaluate controller performance. Our population consisted of 416 virtual patients, generated using population characteristics from previous work. A BIS offset trajectory similar to a surgical case was used.

Results: We were able to develop, describe, and optimize the parameter setting for a patient-individualized model-based closed-loop controller using Bayesian optimization. Selection of the optimal set yields a controller performing with the following median absolute prediction errors at BIS targets 30, 50, and 70: 12.9 +/- 2.87, 7.59 +/- 0.74, and 5.76 +/- 1.03 respectively.

Conclusions: We believe this system can be introduced safely into clinical testing for both induction and maintenance of anesthesia under direct observation of an anesthesiologist.