Sensors that measure alcohol in the perspiration on a person's skin are currently available and are used by courts to monitor DUI offenders and by researchers in NIAAA-funded research and treatment efficacy studies. The potential of the two available devices (SCRAM and WrisTAS) for research and applied use is considerable, but there are serious challenges for users. There are no shared procedures within or across sensors for identifying alcohol use episodes, transdermal alcohol concentration (TAC) values are not quantitatively equivalent to blood or breath alcohol content (BAC/BrAC), and TAC curves are delayed relative to BAC/BrAC curves. Furthermore, sensor data files are large and, due to the nature of the data, calculating summary variables can be very time consuming. These challenges compromise the validity and reliability of research with sensor data, and limit the adoption of the alcohol biosensor technology in research and treatment. The primary objective of this proposal is to use existing data and known experts in addiction research, epidemiology, mathematics, and biostatistics to advance research with alcohol transdermal sensors and to address the key challenges for users of the sensors.
The specific aims are to: (1) establish the best methods for detecting alcohol use episodes with both available transdermal sensors, and (2) develop methods to deconvolve BAC estimates from TAC for both available sensors, and in doing so eliminate the observed time lag in TAC. Simultaneous with addressing each aim, we will produce specialized analytic tools that will incorporate the research findings.
For Aim 1, we will modify a current prototype, the Transdermal Alcohol Sensor Macro (TASMAC) that contains Visual Basic code to read each line of a sensor data input file;according to user-defined parameters the tool will identify alcohol consumption episodes, create summary variables, and format a data output file suitable for use with statistical software.
For Aim 2, the BAC Estimator software will be developed by Dr. Gary Rosen using methods that do not require an alcohol challenge to calibrate the estimation model. Both tools will have graphical user interfaces for ease of use, and both tools will work within Microsoft EXCEL. We will use a collaborative, interdisciplinary approach with eight researchers who have used one of the sensors;collaborators will provide data from controlled laboratory administration studies, field research, and intervention trials to evaluate research aims and develop the tool functions, and will test the tools with their data as part of the evaluation. Research findings and the resulting tools will simplify complex data processing and calculations, thereby removing a barrier to conducting basic, epidemiological, and applied clinical research with transdermal alcohol sensors.
Transdermal alcohol sensors have primarily been used to monitor almost 200,000 individuals who have been court ordered not to use alcohol. Research and treatments incorporating alcohol sensors will develop more quickly if we can validate alcohol use detection criteria and derive accurate estimates of alcohol concentration and alcohol impairment. The tools developed in this application will help translate our research findings into practice, will recommend best practices, and will accelerate the adoption of alcohol sensors for use in research and treatment to help reduce the problem of high-risk drinking.
|Barnett, Nancy P; Meade, E B; Glynn, Tiffany R (2014) Predictors of detection of alcohol use episodes using a transdermal alcohol sensor. Exp Clin Psychopharmacol 22:86-96|