Opioids are a broad group of effective pain-relieving medicines that can become highly addictive in some individuals. According to government sources, nearly 40 million people are addicted to illicit drugs worldwide. In 2017, the U.S. Department of Health and Human Services declared the opioid crisis a national public health emergency. To combat the opioid epidemic, researchers at University of California San Diego School of Medicine will develop an AI model that will more accurately predict opioid addiction in high-risk patients.
The project is funded through a three-year contract with Wellcome Leap as part of a $50 million groundbreaking initiative, called Untangling Addiction. The goal is to revolutionize how we understand opioid addiction and leverage innovative tools, such as artificial intelligence and predictive modeling, to intervene. UC San Diego School of Medicine was one of 14 locations worldwide to receive the funding.
Controlled opioids in the health care setting are still an important part of adequate pain control and used for standard care. However, it is critical to know who is receiving these drugs and the risk it carries with some patients."
Rodney Gabriel, MD, lead researcher for the project, chief of perioperative informatics in the Department of Anesthesiology at UC San Diego School of Medicine and clinical director of anesthesiology at UC San Diego Health
"The AI model will help to identify who is most at risk for an opioid addiction and implement useful resources to help manage their opioid regimen. This way, we can better manage pain in this patient population and also avoid the potentially dangerous downstream consequences of addiction."
The model will use generative artificial intelligence (GenAI), which can produce various types of content. It offers a more holistic approach, which can help with understanding and predicting multiple aspects of a patient's prior and future behaviors.
"GenAI provides more sophisticated ways to predict multiple outcomes based on patterns discovered from large patient datasets," Gabriel said. "We want to better predict risk of addiction the moment a patient is given an opioid prescription to the moment they would start to become addicted."
Researchers will develop electronic health record (EHR) foundation models in a secure platform, which will leverage large multi-institutional datasets to incorporate genomic, social determinants of health, clinical, procedural and demographic data to predict the development of opioid use disorder and related outcomes among any patient initially prescribed an opioid.
"Anesthesiologists have access to a variety of secure data, which we review to safely get a patient through surgery. Dr. Gabriel's research focus is how AI-assisted knowledge of a patient's risks can optimize their overall care, and in this particular instance, decrease the chances of addiction," said Ruth Waterman, MD, chair of the Department of Anesthesiology at UC San Diego School of Medicine and anesthesiologist at UC San Diego Health.
"What will be gained from this project will be translatable to many other areas of a patient's health care journey, resulting in better outcomes and care."
When the predictive tool is ready to be tested in clinical settings, Gabriel and his team will partner with the Joan & Irwin Jacobs Center for Health Innovation at UC San Diego Health (JCHI), which provides a unique environment for integrating AI approaches into clinical care.
"JCHI will be an indispensable resource for computational, technical and collaborative needs," said Gabriel. "Our project aligns with the vision of JCHI, in which we leverage AI to make impactful and significant changes for the future of medicine."
For Karandeep Singh, MD, who was named the inaugural chief health AI officer at UC San Diego Health, real-world evaluations of GenAI's potential are critical.
"Generative AI has the potential to help us better understand people's risk, but this idea hasn't really been put to the test in most areas of medicine," said Singh. "This project will be key towards helping us understand the potential of generative AI in identifying opioid risk."
The ultimate goal of the project is to develop a commercially available genomic and microbiome panel that clinicians can use to easily assess opioid addiction, as well as to develop automated approaches using AI to integrate into EHR systems.
"This will allow us to make real-time predictions of risk throughout a patient's entirety of care and lead the way in the prevention of opioid addiction," said Gabriel.
Impact of glucagon-like peptide-1 receptor agonists on alcohol consumption