Can smartphones predict the risk of death?

A new study by Bruce Schatz of the University of Illinois at Urbana-Champaign and colleagues, published Oct.
20 in the open-access journal PLoS Digital Wellness, suggests that passive monitoring of people's walking activity by smartphones can be used to build population-level models
of health and death risk.

Previous studies have used measures of physical health, including walking tests and self-reported walking speed, to predict an individual's risk
of death.
These metrics focus on the quality rather than quantity of movement; For example, in some clinical situations, measuring an individual's gait speed has become a standard practice
.
The rise of passive smartphone activity monitoring has made it possible
to use similar metrics for population-level analysis.

In the new study, researchers studied 100,000 participants in the UK Biobank National Cohort who wore activity monitors with motion sensors for a week
.
While wrist sensors are worn differently than smartphone sensors, their motion sensors can both be used to extract information about walking intensity from short periods of walking – a walking test
in everyday life.

Using only 6 minutes of steady walking per day collected by sensors, combined with traditional demographic characteristics, the team successfully validated a predictive model
for death risk.
The equivalent of gait speed being a predictor of 5-year mortality independent of age and sex, calculated from passively collected data (composite c-index 0.
72).

The predictive model only uses walking intensity to simulate smartphone displays
.

"Our findings suggest that passive measurements using motion sensors can achieve an accuracy similar to active measurements of gait speed and walking speed," the authors said
.
"Our scalable approach provides a viable pathway
to national health risk screening.
"

Schatz added: "I spent 10 years using cheap phones for clinical modeling
of health conditions.
The data has been tested in the nation's largest cohort to predict life expectancy
at population size.
”

Journal Reference:

1. Haowen Zhou, Ruoqing Zhu, Anita Ung, Bruce Schatz.
   Population analysis of mortality risk: Predictive models from passive monitors using motion sensors for 100,000 UK Biobank participants.
   PLOS Digital Health, 2022; 1 (10): e0000045 DOI: 10.
   1371/journal.
   pdig.
   0000045