Event-native transformer
Model medicine as a sequence
Represent chronological diagnoses, procedures, medications, demographics, costs, and timing as compact medical-event tokens instead of generic language.
06 — Predictions
Large Medical ModelOne history. Many possible futures.
Model longitudinal medical-event sequences and simulate multiple possible futures to forecast risk, care pathways, timing, and cost with GenHealth's Large Medical Model.
Use patient histories and cohort data to understand what may happen next, why it may happen, and where earlier action can improve care and cost.
Model codable medical events, not generic text
Simulate many possible patient futures
Forecast conditions, procedures, medications, and cost
Loading available demo times
What it enables
Built for Large Medical Model in real operations.
Transform chronological medical events into distributional forecasts that show not just a score, but the events and timing behind it.
Read the Large Medical Model whitepaperLongitudinal scale
Learn from broad patient histories
The published evaluation describes training on event sequences from more than 140 million patients across commercial, Medicare, and Medicaid populations.
Simulated futures
See a distribution, not one guess
Generate multiple possible future event sequences, then aggregate them with Monte Carlo simulation to estimate the probability of what may happen next.
Actionable forecasts
Understand what may happen and when
Forecast conditions, procedures, medications, care paths, and cost over time so teams can investigate risk and act with more specific context.