Rapid detection and quantification of falsified Viagra using cloud-based portable NIR technology and machine learning

Abstract

The prevalence of falsified medications remains a global health challenge, intensified by globalization, internet accessibility, and the high profitability associated with low risks for this type of trafficking. This study demonstrates the innovative integration of portable Near-Infrared (NIR) spectroscopy with a cloud-based advanced data processing and management architecture, offering a rapid, non-destructive, and reliable solution for on-site detection and quantification of falsified Viagra tablets. Leveraging the advantages of portable NIR technology—such as its speed, ease of use, and ability to deliver nearly instantaneous results—this approach not only differentiates authentic from falsified tablets but also accurately determines their absolute sildenafil content. Utilizing data from authentic and seized falsified samples, Principal Component Analysis (PCA), Euclidean distance measurements and Support Vector Machine (SVM) highlight the capability of portable NIR devices to effectively distinguish between these groups. Such models can be seamlessly integrated into an online system paired with a mobile application, enhancing accessibility and efficiency in field settings. Furthermore, machine learning models were developed to quantify sildenafil content in falsified tablets, achieving excellent accuracy compared to a reference chromatographic method. These findings underscore the potential of portable NIR spectroscopy, combined with advanced data treatment, as a transformative tool for field deployment, empowering regulatory bodies and healthcare providers to ensure medication quality and safety with greater speed and precision.