Prototype of a breath-based analysis system for medication compliance monitoring

Abstract

Poor compliance to prescribed medication regimens in clinical trials continues to be a major obstacle to drug development success and cost reduction, and there is no gold standard technology that is cheap and effective to reliably measure drug compliance. This paper presents a new drug compliance scheme based on the detection of volatile markers in exhaled breath using a commercial off-the-shelf electronic nose (e-nose) system. A mouth air collection system was developed and interfaced with an e-nose for automated sampling and analysis of breath samples. Advanced signal processing techniques were used to develop classifiers for discriminating between blank breath and breaths masked with volatiles. The compliance monitor was able to achieve 100% detection rate in separating a volatile tagged breath from a blank breath, and 98% discrimination rate between volatile types. The proposed compliance monitor demonstrates a cheap and reliable method of measuring drug compliance in clinical trials.