AI spots risk years ahead

A broken hip can change an older person’s life in seconds, taking away mobility, independence and, in some cases, survival. Now researchers in Sweden say information already sitting in health records could help spot danger before the fall happens.

Scientists analysed records from more than 3.5 million people aged 50 and above, following them for up to a decade. During that period, 142,327 people suffered a hip fracture. Using this enormous pool of information, the team developed an artificial intelligence tool called FRACTURE-ML.

Tap here to add The Daily Star as a trusted source

The tool searches routinely collected details, including illnesses, medicines, previous treatments and basic social information, to identify people most likely to break a hip. Unlike many existing checks, it does not require a face-to-face appointment, weight measurement or questions about lifestyle.

In testing, FRACTURE-ML identified nearly seven times more people at risk of suffering a hip fracture within two years than screening currently used in Sweden. Remarkably, a simpler version using only 35 pieces of information performed almost as well.

The researchers say this could allow preventive care to reach vulnerable people earlier. However, the tool does not include smoking or alcohol habits, and researchers stress that it still needs testing in other countries before widespread use.

Source: PLOS Medicine