Related Experiment Video
Updated: May 15, 2025
08:48
A Rat Model of EcoHIV Brain Infection
Published on: January 21, 2021
3.1K
Predicting Neurobehavioral Outcomes in People with HIV
Ronald J Ellis1, Bin Tang2, Robert K Heaton2
1Department of Neuroscience, University of California San Diego School of Medicine.
Research Square
|April 8, 2025
Summary
Researchers identified four distinct long-term health patterns in people with HIV (PWH) using machine learning. Certain health conditions at the start predicted poorer long-term outcomes, aiding personalized HIV care.
Area of Science:
- Biomedical research
- Public health
- Neuroscience
Background:
- HIV infection impacts multiple body systems.
- Longitudinal studies are crucial for understanding disease progression.
- Biopsychosocial factors significantly influence health outcomes in people with HIV (PWH).
Purpose of the Study:
- To identify complex, multidimensional, longitudinal biopsychosocial phenotypes (MLBPSPs) in PWH.
- To evaluate the association between baseline clinical characteristics and MLBPSPs.
- To inform personalized interventions for vulnerable PWH subpopulations.
Main Methods:
- Machine learning algorithms were employed to analyze data from 506 PWH.
- Biopsychosocial data including neurocognition, mood, cognitive symptoms, and daily living activities were collected across four visits.
- Longitudinal trajectories and cluster analysis were used to define MLBPSP clusters.
Main Results:
- Four distinct MLBPSP clusters were identified, reflecting varying trajectories of neurocognition, mood, and function.
- The largest cluster (C1, N=231) demonstrated stable, optimal functioning over 18 months.
- Baseline chronic pulmonary disease, neuropathic pain, polypharmacy, and creatinine levels predicted adverse MLBPSP trajectories.
Conclusions:
- Distinct longitudinal biopsychosocial phenotypes exist in PWH.
- Baseline clinical factors can predict future adverse health trajectories.
- Findings support the development of targeted interventions for at-risk PWH.
