New preprint by Jianghui Xiong, PhD, and the DeepoMe team introduces SteeraMed, a steerable biomedical world model for N-of-1 intervention reasoning across chronic diseases and aging
BEIJING, China — May 25, 2026 — DeepoMe today announced the publication of a new preprint introducing SteeraMed, which the company describes as the first biomedical world model designed for chronic disease management and aging intervention. The manuscript, titled “SteeraMed: A Biomedical World Model for N-of-1 Intervention Reasoning Across Chronic Diseases and Aging,” is now available on Preprints.org:
https://www.preprints.org/manuscript/202605.1578
Developed by Jianghui Xiong, PhD, and the DeepoMe team, SteeraMed represents a new category of biomedical AI: a world model for individualized health trajectories. Rather than only predicting disease risk or generating static health recommendations, SteeraMed is designed to reason over the dynamic relationship between an individual’s biological state, possible interventions, expected future trajectories, real-world feedback, and evolving evidence.
The framework is part of DeepoMe’s broader mission to develop next-generation AI systems for chronic disease management, precision health, and aging intervention.
“Chronic disease management and aging intervention require more than risk prediction,” said Jianghui Xiong, PhD, CEO of DeepoMe. “They require a system that can reason over state, intervention, trajectory, feedback, and evidence over time. SteeraMed was created to explore this next generation of biomedical AI — moving from prediction-centered models toward steerable world models for individualized health intervention.”
The First Biomedical World Model for Chronic Disease Management and Aging Intervention
Conventional medical AI systems are often designed for diagnosis, triage, imaging interpretation, medical record summarization, or population-level risk scoring. While these tools can be valuable, they are typically not built to model how an individual’s biological state may change over time under different intervention strategies.
SteeraMed is designed around a different question:
If an individual is in a specific biological state today, how might that state evolve under different chronic disease management or aging intervention strategies?
This makes SteeraMed fundamentally different from traditional clinical decision support tools. It is not simply a chatbot, risk calculator, or recommendation engine. Instead, it is a biomedical world model framework organized around five core components:
- State — the current biological, clinical, and phenotypic condition of an individual;
- Action — potential interventions, including lifestyle, nutrition, pharmacologic, behavioral, and other health strategies;
- Transition — how the state may evolve after an intervention;
- Evidence — mechanistic knowledge, clinical literature, biomarkers, and real-world observations;
- Feedback — follow-up data used to adjust reasoning and guide the next step.
By connecting these components, SteeraMed aims to support a more adaptive form of AI reasoning for complex chronic disease and aging-related intervention scenarios.
Moving Beyond Static Recommendations
Chronic diseases and aging are dynamic, multifactorial processes. A person’s biological state changes continuously, and the impact of an intervention may depend on baseline condition, timing, dosage, adherence, comorbidities, lifestyle, environment, and biomarker feedback.
Most healthcare AI systems are not designed to handle this level of longitudinal complexity. They may predict risk or summarize information, but they often do not model how an individual’s health trajectory may shift under different intervention pathways.
SteeraMed is designed to address this gap through N-of-1 intervention reasoning, where each individual becomes the core unit of analysis.
In this framework, the key questions become:
- What is the individual’s current biological state?
- What mechanisms may be driving their chronic disease or aging phenotype?
- Which interventions may plausibly shift the trajectory?
- What are the expected benefits, uncertainties, and risks?
- What feedback should be monitored?
- How should the plan be adjusted as new data emerges?
A Steerable Framework for Precision Health and Longevity Medicine
A defining feature of SteeraMed is steerability.
In biomedical AI, steerability refers to the ability of researchers, clinicians, and domain experts to inspect, question, correct, and guide an AI system’s reasoning within clearly defined boundaries. This is especially important in chronic disease management and aging intervention, where AI systems must remain transparent, auditable, and human-supervised.
SteeraMed is designed to support:
- Mechanism-aware intervention reasoning;
- Longitudinal biomarker interpretation;
- Counterfactual comparison of possible health trajectories;
- Human-in-the-loop expert review;
- Feedback-based plan adjustment;
- Uncertainty-aware decision support;
- Evidence-chain tracking across interventions and outcomes.
This steerable design is intended to make biomedical AI more useful for real-world precision health workflows, where the goal is not simply to provide a one-time recommendation, but to help model, monitor, and guide an adaptive health trajectory over time.
More information about the SteeraMed framework is available at:
Potential Applications Across Chronic Disease and Aging
SteeraMed is initially positioned for complex, longitudinal health scenarios where static recommendations are insufficient. Potential application areas include:
- Metabolic health management;
- Cardiometabolic risk reduction;
- Chronic inflammation and immune dysregulation;
- Biological aging assessment;
- Longevity intervention planning;
- Multi-omics-informed health optimization;
- Personalized lifestyle and nutrition intervention;
- Longitudinal biomarker monitoring;
- N-of-1 intervention tracking.
The framework may be especially relevant to aging intervention because aging is not a single-disease or single-biomarker problem, but a multi-system biological process involving interactions among metabolism, inflammation, immunity, epigenetics, stress response, lifestyle, environment, and time.
SteeraMed aims to provide a structured reasoning layer for understanding how interventions may influence an individual’s biological aging trajectory over time.
From Medical AI to Biomedical World Models
The publication of the SteeraMed preprint reflects a broader shift in healthcare AI.
The first wave of medical AI focused on classification and prediction.
The second wave focused on generative AI and clinical copilots.
The next wave may be defined by biomedical world models — systems capable of representing biological states, reasoning over possible interventions, modeling future transitions, incorporating feedback, and supporting individualized decision-making.
DeepoMe is positioning SteeraMed as an early framework for this emerging category.
“Healthcare does not only need larger models,” said Xiong. “It needs models that can reason about change. Chronic disease management and aging intervention are trajectory problems. The future of precision health will depend on AI systems that can help model, monitor, and steer those trajectories.”
Research Status and Limitations
The SteeraMed manuscript is currently available as a preprint and has not yet undergone peer review. SteeraMed is presented as a research framework and conceptual architecture for biomedical world modeling.
SteeraMed is not a clinically validated treatment system, not a diagnostic product, and not a medical device. Further validation will be required through real-world biomedical datasets, prospective studies, clinical expert review, longitudinal outcome evaluation, and appropriate regulatory pathways where applicable.
About DeepoMe
DeepoMe is developing next-generation biomedical AI systems for chronic disease management, precision health, and aging intervention. The company’s work focuses on biomedical world models, systems biology, biomarker interpretation, N-of-1 reasoning, and human-in-the-loop AI frameworks for individualized health trajectory modeling.
For more information, visit:
About SteeraMed
SteeraMed is a biomedical world model framework for N-of-1 intervention reasoning across chronic disease and aging. It is designed to organize individualized biological states, possible interventions, expected trajectories, evidence chains, feedback signals, and expert steering into a structured AI reasoning framework.
For more information, visit:
Media Contact
Jianghui Xiong, PhD
CEO, DeepoMe
Email: info@deepome.com
Website: https://deepome.com
Preprint
“SteeraMed: A Biomedical World Model for N-of-1 Intervention Reasoning Across Chronic Diseases and Aging”
https://www.preprints.org/manuscript/202605.1578
