Sean X. Luo MD, PhD
Associate Professor of Clinical Psychiatry
Director, Addiction Data Science Workgroup
Center for Healing of Opioid and Other Substance Use Disorders (CHOSEN)
Division of Pathophysiology and Treatment Research
Department of Psychiatry
Columbia University and New York State Psychiatric Institute
Our group's overall goal is to leverage advanced computational methods to match patients with substance use disorders to optimal treatment. We combine bottom-up approaches — phase I/II trial data and single-site human subjects research — with top-down analytics of phase III clinical trial data, electronic medical records, claims databases, and mobile devices.
Data Science
Real-time prediction models in opioid use disorder
Developing models that predict individual patient risk for relapse or treatment dropout under methadone, buprenorphine, or extended-release depot naltrexone. We harmonized clinical trial datasets from three multi-site, NIDA-sponsored OUD trials and build a portfolio of projects around analytics, harmonization, and Common Data Models.
More information at the CTN Dissemination Library.
Predictive and Causal Inference in OUD/SUD Claims Data
Collaborating with Kara Rudolph's group, we are developing individual level modeling work that assesses whether informaton contained in administrative claims data can be used to make predictions and assess whether certain kinds of interventions are more effective.
Large Scale Informatics in SUD/OUD
We are developing several projects that capitalize on access to unique datasets including the TriNetX dataset, the Epic COSMOS dataset, and datasets from implementation science and other domains in addiction science. We also are involved in developing new techniques that applies contemporary machine learning (deep reinforcement learning, transfer learning) to OUD/SUD datasets.
Contribute to a study
Contact us to participate in a data science challenge to connect to learn more about whether your dataset might generate novel scientific questions. Links to future Kaggle and other open source projects will be posted here.
Open Source CompetitionsOpen-source code
Current results, working code, and the open-source teaching materials for our short course are released on our GitHub repository.
GitHubJoin the group
We are looking for individuals with a strong background in data science and technology development at various training levels — part-time or full-time — who want to apply their skills to addiction research.
Get in touchPractice, teaching,
& community.
Beyond the research program, our group contributes to data science quality improvement in mental health delivery, runs an informal forum for emerging addiction science, and maintains an open teaching curriculum for psychiatry trainees.
Data science quality improvement
We collaborate with organizations to provide data-driven services — predictive model construction, evaluation, and dissemination — for quality improvement and care delivery. Expertise in behavioral health data analytics applied to treatment prediction, claims analysis, quality-of-care measures, cost-benefit analysis, and technology development.
Ongoing collaborations with NewYork-Presbyterian Hospital (Evaluation Core) and the New York State Psychiatric Institute.
Addiction Data Science Workgroup
An informal work group offering a thoughtful, free-flowing forum for fellows and faculty within and outsdie of the department to discuss contemporary issues in addiction data science science, plan projects, and collaborate on grant writing. Open door virtual office hour.
Every last Monday of the month at 3:30PM except holidays.
Becoming a data-savvy psychiatrist
A short course developed for training psychiatry residents in the basics of data science. The course materials are released in the public domain on our GitHub and can be adapted for executive training programs aiming to provide basic analytics proficiency for mental health professionals.
Brief Biography
I am a practicing addiction psychiatrist and computational addiction science researcher at Columbia University. I trained in physics and mathematics at the University of Chicago and completed an MD/PhD at Columbia in theoretical neuroscience and machine learning under Richard Axel and Larry Abbott. I have focused my research interest in application of advanced statistical and machine learning techniques in substance use disorder treatment, dosing strategies, and overdose risk.
Our work is funded by NIDA, NIDA CTN, and the HEAL Inititive.
Recent work.
Full publication list in the curriculum vitae.