Examining new-generation transdermal alcohol biosensor performance across laboratory and field contexts

Area of Science:

• Biomedical Engineering
• Analytical Chemistry
• Public Health

Background:

• Wrist-worn transdermal alcohol sensors offer a novel method for monitoring alcohol intake.
• Previous hardware and data analysis challenges have limited the widespread adoption of these devices.
• Recent technological advancements prompt an updated evaluation of new-generation transdermal sensor performance.

Purpose of the Study:

• To assess the performance of a new-generation wrist-worn transdermal alcohol sensor.
• To evaluate the accuracy of transdermal alcohol concentration data converted to Breath Alcohol Concentration (BrAC) estimates.
• To compare sensor performance in both controlled laboratory and real-world field settings.

Main Methods:

• Utilized machine learning models to convert transdermal alcohol concentration data to BrAC estimates.
• Evaluated accuracy by comparing sensor-derived BrAC estimates with real-time breathalyzer measurements.
• Conducted studies in a large laboratory sample (N=256) and a pilot field sample (N=27).

Main Results:

• The newest sensor version demonstrated significantly lower error rates (0-7%) compared to older prototypes (29-53%).
• Average BrAC estimates were within 0.007 of true BAC in the lab and 0.019 in the field.
• Power-law projections suggest >25% accuracy improvement in real-world contexts with sufficient training data.

Conclusions:

• The latest transdermal wrist sensor shows potential for accurate alcohol consumption assessment in field settings.
• Further research with larger field samples is necessary to fully understand device utility.
• Continued development may enhance the accuracy and applicability of transdermal alcohol monitoring.