Over 800,000 Americans are living with end-stage kidney disease, and most of them are on hemodialysis three times a week. Nearly all of them will be hospitalized within their first year. Most admissions trace back to two culprits: fluid overload and infections — both theoretically preventable with early enough intervention.
The problem is that dialysis clinics only see patients three times a week. Between sessions, no one is watching. By the time a nurse or nephrologist notices something is wrong, the window for outpatient intervention has often already closed. Researchers at Fresenius Medical Care and Interwell Health asked whether AI could change that — by identifying high-risk patients before they deteriorate and routing them to nurses who can actually do something about it.
The data is more nuanced than the headlines suggest…
