Smartphones could tell how soon people will die

A study of users’ walking habits could be used to predict mortality in the next five years, researchers suggest

(Pixabay)

Smartphones might be able to detect how soon someone might die by analysing how they walk, researchers have suggested.

In a new study, over 100,000 participants wore activity monitors for one week, but it only took six minutes of walking data per day for researchers to predict someone’s risk of dying within the next five years.

Combined with the wearer’s demographic information, their gait speed could be gathered by the phone and used in an algorithm to predict mortality rates; of the middle-aged and senior adults in the study, approximately two per cent died in the following years.

“Our results show passive measures with motion sensors can achieve similar accuracy to active measures of gait speed and walk pace,” the authors say. “Our scalable methods offer a feasible pathway towards national screening for health risk.”

While the participants in the study wore motion sensors on their wrists, the researchers say that the same data could be gathered using inexpensive phones as their accelerometers are good enough to gather the data and “even the cheapest flip phones incorporating motion sensors”

The researchers also suggest that this could be useful in poorer countries, as even in Sub-Saharan Africa approximately 48 per cent of people have smartphones. In the UK, smartphone ownership is at 92 per cent.

Generally, there are two main ways for measuring physical activity: self-reporting via a questionnaire or walking a fixed distance, and passively collecting data through sensors. The issue with the former, the researchers say, are “problematic to scale for population level assessment, due to logistic difficulty of getting large numbers of people to perform the required tasks on a routine basis.”

However, sensor-based research is has a major advantage as it does not require people to change their day-to-day activities.

The study has been published in PLOS Digital Health.