The data collected by mobile phone motion sensors from just six minutes of walking could be enough to predict their likelihood of dying within the next five years, according to US researchers.

A study by Bruce Satz of the University of Illinois and published in the digital health journal PloS Digital Health according to New Scientist found that the accuracy of the above method is roughly comparable to that of other, more traditional prediction methods.

Previous studies have assessed the risk of premature death based on overall physical activity using sensors in other devices such as wristbands and smartwatches. Almost all people now have “smart” mobile phones (smartphones) with similar motion sensors, but they do not carry the phone with them all the time, so it is impossible to reliably record their overall level of physical activity.

To find an alternative predictor of risk, the researchers analyzed the health data of 100,655 people who wore an electronic motion-sensing bracelet for a week. Over the next five years, about 2% of the participants died from various causes.

An analysis of AI—using an algorithm that estimated risk of premature death based only on speeding up during a six-minute walk—showed that in the presence of diseases such as heart and lung disease, people often slow down their pace and then they speed it up again in small doses. Based on acceleration and deceleration analysis, the new model can predict the future risk of death with an accuracy comparable to other more traditional indicators (for example, using health questionnaires or total daily activity).

While the new study used motion sensors worn on the wrist, Szaz says smartphones can also measure acceleration during a short walk, predicting a similar risk. The researchers even suggest that it is possible to use data from mobile phones at the level of the entire population of the country to assess the health status of residents.

RES-EMI