Can smartphones predict mortality risk? - Speaking of Medicine and Health - PLOS

This week in Research Highlights, a monthly PLOS blog bid exploring caller noteworthy probe from crossed the journals, we research a caller PLOS Digital Health survey concluding that passive smartphone monitoring of people’s walking enactment could beryllium utilized to conception population-level models of wellness and mortality risk. Read our summary of the findings from Bruce Schatz of University of Illinois astatine Urbana-Champaign, USA, and colleagues beneath oregon leap to the full article successful PLOS Digital Health. 

Background

Previous studies person utilized measures of carnal fitness, including locomotion tests and self-reported locomotion pace, to foretell idiosyncratic mortality risk. These metrics absorption connected prime alternatively than quantity of movement; measuring an individual’s gait velocity has go a modular signifier for definite objective settings, for example. The emergence of passive smartphone enactment monitoring opens the anticipation for population-level analyses utilizing akin metrics.

Study Design and Results

Researchers studied 100,000 participants successful the UK Biobank nationalist cohort who wore enactment monitors with question sensors for 1 week. While the wrist sensor is worn otherwise than however smartphone sensors are carried, their question sensors tin some beryllium utilized to extract accusation connected walking strength from abbreviated bursts of walking—a regular surviving mentation of a locomotion test.

Measuring wellness with carried smartphone, from diagnostic question of quality assemblage computed from telephone sensor. Image Credit: Qian Cheng (CC-BY 4.0)

The squad was capable to successfully validate predictive models of mortality hazard utilizing lone 6 minutes per time of dependable walking collected by the sensor, combined with accepted demographic characteristics. The equivalent of gait velocity calculated from this passively collected information was a predictor of 5-year mortality autarkic of property and enactment (pooled C-index 0.72). The predictive models utilized lone walking strength to simulate smartphone monitors.

“Our results amusement passive measures with question sensors tin execute akin accuracy to progressive measures of gait velocity and locomotion pace,” the authors say. “Our scalable methods connection a feasible pathway towards nationalist screening for wellness risk.”

Schatz adds, “I person spent a decennary utilizing inexpensive phones for objective models of wellness status. These person present been tested connected the largest nationalist cohort to foretell beingness expectancy astatine colonisation scale.”