M Van Horn
University of New Mexico, NM·Psychology
Many treatments for alcohol use disorder (AUD) work modestly, relapse is common, and no single therapy clearly beats the others. One reason may be that different people respond best to different treatments, but past studies rarely found predictors strong enough to guide clinical choices.
- Fields
Psychology
Mathematics & Statistics
Public Health & Epidemiology
- Topics
alcohol use disorder
alcoholism
heavy drinking
- Grants
- 1 known grant, through Mar 2028
What they work on
- Develop and cross-validate machine-learning predictions of each person's response to five medications that reduce heavy drinking and six behavioral relapse-prevention programs, using pooled data from 11 randomized trials.
- Test whether individual differences in treatment effects are significant for each intervention, and provide predictive intervals that describe each person's expected response.
- Develop and test new methods to combine data across multiple trials and improve the precision of predicted individual treatment effects (PITEs), using simulations to check cross-validation.
More about the research
This project tests whether machine learning, which combines many patient characteristics at once, can predict which treatment will work best for a given person.
Why it matters
People seeking help for alcohol use disorder often try treatments by trial and error. Validated predictions of which medication is most likely to reduce a person's heavy drinking, and which behavioral program best prevents relapse, could help clinicians match patients to treatments and improve outcomes.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Renee Martin-Willett
University of Colorado Boulder, CO·Psychology·Early-career
Test whether cannabidiol (CBD), a non-intoxicating cannabis compound, improves sleep in adults who report heavy drinking, using a randomized placebo-controlled trial.
Andre Miguel
Washington State University, WA·Medicine & Clinical Research
Test whether rural smokers with alcohol use disorder who receive varenicline plus app-based contingency management submit more breathalyzer samples showing alcohol abstinence and non-heavy drinking days than those using a control app.
Brandon Oberlin
Indiana University-Purdue University-Indianapolis, IN·Psychology
Test whether 30 daily virtual reality encounters with digitally age-progressed Future Selves improve abstinence and alcohol use outcomes at 1 and 6 months in people in early recovery from alcohol use disorder, compared with a virtual reality control condition.
The school
University of New Mexico: 27 labs on PI Pages, $355M a year on research, 182 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.