Simone Marini
University of Florida, FL·Public Health & Epidemiology
Clinical data about a patient, lab values, notes a doctor typed, imaging, information about housing or transportation, arrive in many formats and accumulate over years. Most prediction tools reduce all of that to a single yes-or-no forecast and cannot explain their reasoning.
- Fields
Public Health & Epidemiology
Medicine & Clinical Research
Computer Science
- Topics
HIV care continuum
people living with HIV
antiretroviral therapy
- Grants
- 2 known grants, through Dec 2030
What they work on
- Build an agentic AI system that predicts multiple HIV outcomes at once, loss of care, viral suppression, CD4 count, new acute events, and new comorbidities, from longitudinal electronic health records, clinical notes, social-behavioral determinants, and medical imaging.
- Develop an antiretroviral regimen optimization system using causal AI, so recommendations reflect the likely effect of changing treatment across several HIV health endpoints rather than statistical association alone.
- Assess how such a system would be implemented in practice through qualitative evaluation with infectious disease specialists and community healthcare experts, using their input to design the clinical applications and human-interpretable outputs.
- Identify gene modules in the pain-endocannabinoid system whose over- or under-expression is associated with medical marijuana treatment and with change in chronic pain, using targeted gene expression in older adults with chronic musculoskeletal pain.
- Test whether changes in chronic pain severity track with changes in allostatic load, a composite index of cumulative physiological wear, over three months in 80 patients initiating medical marijuana and 40 who are not.
- Develop a causal, personalized prediction model combining genetic, sociodemographic, and psychological/behavioral data to identify which older adults are most likely to benefit from medical marijuana for musculoskeletal pain.
More about the research
Two projects here build computational systems that handle many outcomes at once, work from several kinds of data together, and produce explanations clinicians can inspect: one for the care of people living with HIV, one for older adults considering medical marijuana for chronic pain.
Why it matters
People living with HIV face several interacting health risks at once, and clinicians currently lack tools that forecast them jointly or explain their reasoning; an interpretable system for predicting loss of care, viral suppression, and new conditions, plus drug regimen optimization, could support better-targeted clinical decisions. For older adults, about two-thirds of whom use medical marijuana primarily for chronic pain, evidence on benefits and risks is thin; identifying genetic, sociodemographic, and behavioral predictors of response could help clinicians decide for whom the treatment is worth trying and how to monitor it.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Danielle Haley
Boston University, MA·Public Health & Epidemiology
Determine whether cannabis use disorder is associated with new falls, fractures, delirium and dementia, and with HIV viral suppression and antiretroviral therapy adherence, in about 180,000 Veterans Aging Cohort Study patients.
Selen Bozkurt
Emory University, GA·Medicine & Clinical Research
Translate clinical trial protocol eligibility text into computable phenotypes and query packages tied to standard vocabularies, and validate them against expert references and prescreening logs.
Bryan Lau
Johns Hopkins University, MD·Public Health & Epidemiology
Assess how HIV status relates to COVID-19 hospitalization and post-acute sequelae of COVID-19 (lasting after-effects), using Medicaid claims from 27 states and Washington, DC.
The school
University of Florida: 65 labs on PI Pages, $1.3B a year on research, 795 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.