Hospitalizations are often predictable - if you look early enough
Most risk tools score a patient once. Passage Healthcare models how health evolves - vitals, utilization, meds, and social context - so rising acuity shows up before the bed is filled.
Even the smallest datasets update the posterior. Sparse claims, a missed fill, a quiet stretch of encounters - weak signals still reshape who we watch first.
Inference under sparse data
The Care Pathway. Diagnosed.
Move the scales to change how many clinical signals map onto the patient - denser overlays mean higher modeled risk.
Signal density
38 datapoints