Surviving is not the same as staying independent — AI now flags the heart patients who need help keeping both

For older people who leave hospital after heart failure, the danger is not always dying — it is quietly losing the ability to walk, dress and live at home unaided, even when the survival odds look good. Researchers at Juntendo University in Japan studied 6,382 patients aged 65 and over who had been in hospital for heart failure, had physiotherapy, and were fully independent before they went in. They trained a computer model on 77 clues collected during care — walking speed, handgrip strength and balance tests alongside the usual medical measures — to predict who would still be independent a year later, who would decline, and who would die. A simpler version using just ten clues, led by how well people could physically move, did better than standard frailty questionnaires — and walking speed often told doctors more than scans of the heart itself. Most strikingly, the model found 1,732 patients who looked safe on survival alone (an estimated 89% chance of living a year) but were at high risk of losing independence — and 57% of them did decline. Spotting that risk before discharge, the team says, lets hospitals plan rehabilitation and support while it can still make a difference. The study was published in the Journal of Cardiac Failure, and the team has built a prototype web tool; it still needs testing in routine care.

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