Glucose Digital Twin
A mechanistic glucose–insulin simulator (Hovorka core + gut absorption, endogenous secretion, medication PK/PD, and physiologic modifiers). Build a regimen, run it, and predict the 24-hour glucose trajectory, time-in-range, and hypo/hyper risk.
Calculators — A1c, ISF/ICR & starting regimen
An A1c of 7.0% ≈ an average glucose of 154 mg/dL (eAG).
Reverse: 154 mg/dL ≈ A1c 7%
Meal boluses (by ICR): B 3U · L 4U · D 5.5U — for 45/60/75 g.
Steroid hyperglycemia: increases hepatic glucose output and impairs peripheral insulin sensitivity.
Regimen library — load an in/outpatient scenario
Ten ready-made regimens across both care settings — load one to see exactly what it does, then tweak it or pin it and compare against another.
Import from Epic / EMR / CGM
Paste a glucose flowsheet, CGM export, or lab report. The app extracts every reading, marks when Lantus/insulin was given, and flags DKA/HHS — then overlays it all on the chart.
Predicted glucose trajectory
Regimen builder
Meals
Insulin boluses
Basal insulin
No basal segments (use Auto-basal toggle for a tuned rate).
Oral / injectable agents
No agents added.
Exercise
Times are minutes from simulation start (07:00).
Dose advisor — predict & recommend
Set the current glucose (drag the chart handle or the slider). The advisor simulates forward and recommends a dose, showing exactly how far it drops you.
Patient physiology & dosing
Type 1 — autoimmune β-cell destruction; C-peptide typically undetectable.
Physiologic & environmental modifiers
Predictions & guidance
- Run a scenario to see predictions.
Model reference & data layer
Every variable, parameter, formula, and clinical threshold the engine uses — the same canonical data other condition pathways can import from @/lib/glucose-twin.
Research & education tool. Equations adapted from the Hovorka/Cambridge, Bergman, Dalla Man/Cobelli and Sorensen/Panunzi model families plus standard PK/PD templates. Not a medical device and not for individual treatment decisions.