Linking Genetic and Clinical Data to Optimize Surgical Opioid Analgesic Prescribing and Predict Risks of Opioid-Related Adverse Drug Events

PROJECT SUMMARY While opioid drugs are very effective for controlling pain, they are highly addictive. The current clinical decision making for post-surgical opioid prescribing is still based on oversimplified information. Both opioid use disorder and opioid-related pain response have strong genetic underpinnings. A better system is needed to guide the clinical use of genotype data and facilitate better-informed opioid prescribing decision-making. Electronic health record (EHR) offer a largely untapped source of information to conduct clinically oriented studies. By integrating genomic data and EHR dataset in large-scale clinical Biobank, we can now perform "clinical-driven genomic research". This approach is particularly robust for complex diseases with a genetic susceptibility. I propose to develop a model to recommend optimal opioid dosage (oral morphine milligram equivalents) for post-surgical pain relief and predict the risk of post-discharge harmful outcomes. The large-scale clinical and genomic databases from the Mass General Brigham (MGB) healthcare system, including eight hospitals that share a centralized database, will be utilized for model development. The proposed research will not only have strong immediate potential to improve clinical practice in this urgent area but will also provide strong preliminary work to support my first NIH R01 application. From the previous projects, I have developed opioid use disorder clinical phenotypes and identified associated genetic markers. I have also developed machine learning models to predict risks of complex diseases and explored clinical application potentials. The proposed project will combine my existing analytical skills with the newer expertise that I seek to develop through this award program, including a nuanced understanding of analgesic strategies, addiction, and implementation of clinical decision support systems, as well as artificial intelligence applications in healthcare system. I have invited several established investigators to form a strong, multi-disciplinary mentorship team. They include experts in the fields of addiction, pain management, genetic analysis, implementation methods and artificial intelligence techniques. With their support, I intend to dedicate my longer-term research career to study functional genomics in the fields of opioid addiction and pain by using integrated analysis of harmonized big data. The K01 program outlined in this application will provide timely and strategic support for my continued advancement towards those goals.