A digital twin framework for forensic reconstruction of alcohol intake via fast and slow metabolite kinetics

Accurately determining the timing and/or amount of consumed alcohol is critical in healthcare and forensic contexts, yet self-reported data are often unreliable due to stigma and legal implications. Objective biomarkers are therefore essential. Commonly used markers such as blood- and breath alcohol concentrations (BAC, BrAC) decline rapidly, limiting their utility for reconstructing drinking scenarios. Such reconstructions are essential for legal assessments, including refuting the “hipflask” defence. Markers with slower kinetics, such as ethyl glucuronide (EtG), ethyl sulphate (EtS), and urine alcohol concentration (UAC) provide complementary temporal information; however, current approaches fail to fully exploit available data. Here, we introduce a unified physiological digital twin that mechanistically integrates BAC, BrAC, EtG, EtS, and UAC within a single framework. This model captures the joint dynamics of rapid- and slower markers, enabling personalized simulations of alcohol intake and metabolism. We show that this integrated approach substantially improves reconstruction of past drinking events and supports complex forensic assessments requiring high temporal precision. To ease practical application, we provide an interactive web tool that allows users to evaluate hypothetical drinking scenarios and visualise individualized biomarker trajectories. Our work establishes a foundation for precision modelling of alcohol kinetics, bridging gaps between clinical, forensic, and computational domains.