Comparing a Distributed Parameter Model-Based System Identification Technique with More Conventional Methods for...

Area of Science:

• Biomedical Engineering
• Analytical Chemistry
• Pharmacokinetics

Background:

• Accurate estimation of blood or breath alcohol concentration (BAC/BrAC) is crucial for clinical and research applications.
• Transdermal alcohol concentration (TAC) measured by biosensors offers a non-invasive alternative to traditional methods.
• Estimating BAC/BrAC from TAC involves solving an ill-posed inverse problem, requiring robust deconvolution techniques.

Purpose of the Study:

• To evaluate and compare three distinct methods for estimating BAC/BrAC from biosensor-measured TAC.
• To assess the performance of a novel system identification/quasi-blind deconvolution scheme against conventional techniques.
• To determine which method best captures key pharmacokinetic parameters of alcohol distribution.

Main Methods:

• A system identification approach using a distributed parameter model for ethanol transport in the skin.
• Comparison with two conventional methods: frequency domain deconvolution and time series analysis (ARMA model).
• Evaluation based on five statistical measures: peak BAC/BrAC, time of peak, ascending/descending slopes, and area under the curve.

Main Results:

• The novel system identification method demonstrated comparable or superior performance in estimating key BAC/BrAC parameters.
• All evaluated methods provided estimates for peak BAC/BrAC, time of peak, slopes, and area under the curve.
• The distributed parameter model offered a potentially more accurate representation of skin ethanol transport dynamics.

Conclusions:

• The system identification/quasi-blind deconvolution scheme is a viable and potentially advantageous method for estimating BAC/BrAC from TAC.
• This approach may improve the accuracy and reliability of non-invasive alcohol monitoring.
• Further validation is warranted to establish its clinical utility in diverse populations and conditions.