Determining blood and/or breath alcohol concentration from transdermal alcohol data

Abstract

We develop a scheme for the blind deconvolution of blood or breath alcohol concentration from biosensor measured transdermal alcohol concentration (TAC). The scheme is based on a distributed parameter model with unbounded input and output for the transdermal transport of ethanol from the blood through the skin to the sensor. The estimation of the convolution filter that serves to calibrate the underlying model to a particular subject and device is formulated as a nonlinear least squares fit of unknown parameters appearing in the model to a subject's laboratory alcohol administration session data. The deconvolution is formulated as a regularized quadratic programming problem using the calibrated model. A scheme to automatically locate distinct drinking episodes in the TAC data based on a Hodrick Prescott filter is discussed. Numerical results involving actual patient data is presented.