NIDA Clinical Trials Network, CTN-0073-Ot
6/1/16 - 5/31/20
Lisa A. Marsch, PhD (Geisel School of Medicine at Dartmouth); Santosh Kumar, PhD (University of Memphis)
Other Project Staff
Dr. Emre Ertin (sensor expert who designed AutoSense and EasySense sensors at Ohio State); Dr. Kenzie Preston (Chief of the Clinical Pharmacology and Therapeutics Research Branch at the NIDA Intramural Research Center); Dr. August Holtyn (Research Associate at the Center for Learning and Health at the Johns Hopkins University School of Medicine); Dr. Udi Ghitza (CCTN); Dr. Dee Blumberg (CCC); Jennifer McCormack (DSC); Shahin Samiei (University of Memphis); Andrea Meier (Dartmouth); Bethany McLeman (Dartmouth); Samantha Auty (Dartmouth);Carmen Rosa (CCTN); Syed Hossain (University of Memphis); Massoud Vahabzadeh (Chief of the Biomedical Informatics Section at the NIDA Intramural Research Center); Jia-Ling Lin (Biomedical Informatics Section at the NIDA Intramural Research Center); Mustafa Mezghanni (Biomedical Informatics Section at the NIDA Intramural Research Center)
The goal of this project is to investigate methods to detect cocaine use from heart rate data captured by smartwatches, so this approach can be deployed widely in the NIDA Clinical Trials Network (CTN). This method will enable us to automatically detect cocaine use and the precise timing of such use. And, this approach can nicely complement self-report methods that suffer from temporal inaccuracy in reporting cocaine use in the field setting. Detection of cocaine use via smartwatches will build upon, and extend, our recently developed methods to identify cocaine use from interbeat interval heart rate data obtained from electrocardiogram (ECG) sensors and physical activity from accelerometer data.
The specific aims of the project are:
1.Develop a smartwatch device that can reliably detect interbeat interval and can last the entire day on a single charge of battery with continuous sensor data collection.
2.Conduct a user study to determine the feasibility of using smartwatches to collect reliable interbeat interval and physical activity data in the natural field setting. This study will provide the data necessary to determine under what conditions high quality data can be obtained from smartwatches, identify common failure scenarios, and understand wearability/usage patterns.
3.Adapt the computational model for detecting cocaine use from interbeat interval, so it can be applied to the interbeat and physical activity data obtained from smartwatches. We will also assess the degree of specificity of the model relative to other stimulant use.
Public Health Relevance
Illicit drug use results in significant consequences including morbidity, mortality and health care costs. Detection of cocaine use relies heavily on inaccurate self-report or intrusive urine screens. Detection by smart watches offer many benefits to researchers, providers and consumers.