Surviving isn't the same as living: a new tool spots which heart patients risk losing their independence

For an older person leaving hospital after heart failure, the scariest outcome is not always dying — it is surviving but no longer being able to live independently. Yet most risk scores only predict death or readmission. At Juntendo University in Japan, Kanji Yamada, Tetsuya Takahashi and colleagues studied 6,382 patients aged 65 and over and taught a machine-learning model to predict something more human: who would keep their independence a year later. Published in the Journal of Cardiac Failure, the model — built on simple measures like walking speed and grip strength taken at discharge — found a hidden group of 1,732 people with an 89% chance of surviving the year but a 57% rate of losing independence. Naming that group early means rehabilitation and home support can reach them while they can still change the ending.

SDG 3 · Good health and well-being — Mechanism: tech innovation