Accepted for/Published in: JMIR Formative Research
Date Submitted: Jan 3, 2023
Open Peer Review Period: Dec 26, 2022 - Feb 20, 2023
Date Accepted: Sep 1, 2023
(closed for review but you can still tweet)
Community Substance Use: A Pilot Correlative and Machine Learning Study Using Wastewater
ABSTRACT
Background:
Substance use disorder and associated deaths have increased in the United States, but methods for detecting and monitoring substance use utilizing rapid and unbiased techniques are lacking. Wastewater-based surveillance is a cost-effective method for monitoring community drug use. However, the examination of the results often focuses on descriptive analysis.
Objective:
The current study aimed to utilize machine learning to better understand geographic differences and explore commonalities of substance use in the United States. Further, it looked to validate trends in wastewater levels of drugs and metabolites with other forms of substance use surveillance.
Methods:
Wastewater was sampled across the United States (n=12). Selected drugs with misuse potential, prescriptions, and over-the-counter drugs and their metabolites were sampled across geographic locations for 7 days. Machine learning was utilized to assess geographical patterns of drug use.
Results:
Geographic variations in the wastewater drug or metabolite levels were found. Specifically, results revealed a higher use of methamphetamine (z=-2.27, p=0.02) and opioids-to-methadone ratios (oxycodone-to-methadone: z=-1.95, p=0.05; hydrocodone-to-methadone: z=-1.95, p=0.05) in states west of the Mississippi River compared to the east. Machine learning suggested temazepam and methadone were significant predictors of geographical locations. Precision, sensitivity, specificity, and F1 scores were 0.88, 1, 0.80, and 0.93, respectively. Finally, cluster analysis revealed similarities in substance use among communities.
Conclusions:
These findings suggest that wastewater-based surveillance is an effective form of surveillance for substance use. Further, machine learning may help uncover geographical patterns and detect communities with similar needs for resources to address substance use disorders. Utilizing automated machine learning, these advanced surveillance techniques may help communities develop tailored treatment and prevention efforts.
Citation
Request queued. Please wait while the file is being generated. It may take some time.
Copyright
© The authors. All rights reserved. This is a privileged document currently under peer-review/community review (or an accepted/rejected manuscript). Authors have provided JMIR Publications with an exclusive license to publish this preprint on it's website for review and ahead-of-print citation purposes only. While the final peer-reviewed paper may be licensed under a cc-by license on publication, at this stage authors and publisher expressively prohibit redistribution of this draft paper other than for review purposes.