Smartphones may be used as motion sensors to predict population-level mortality risk

Using smartphones as motion sensors may be the key to population-level analysis for mortality risk, researchers wrote.

Previous research has demonstrated that walking more than 10,000 steps per day reduces mortality risk for prediabetes and diabetes, and that combinations of aerobic exercise and muscle strengthening may reduce mortality risk. Now, many smartphones have embedded accelerometers, which can be used as motion sensors. These can be effective in measuring quality of physical activity through walking intensity by determining gait speed.

“Using smartphones as passive monitors for population measurement is critically important for health equity, since they are already ubiquitous in high-income countries and increasingly common in low-income countries,” Haowen Zhou, of the University of Illinois department of statistics, and colleagues wrote in PLOS Digital Health.

In a study of 100,000 participants , the researchers used U.K. Biobank data to extract walking window inputs from wrist sensors, simulating smartphone data. The participants wore activity wrist monitors for 1 week, and the researchers used the data from daily 6-minute walking sessions to predict mortality risk, measuring the intensity rather than the duration of the walk test.

The researchers were able to validate the predictive model using only sensors and demographics, they wrote, resulting in a C-index of 0.76 for 1-year risk, which then decreased to 0.73 for 5-year. In calculating the equivalent of gait speed, they found a predictor of 5-year mortality, independent of sex and age .

“Our results show passive measures with motion sensors can achieve similar accuracy to active measures of gait speed and walk pace,” Zhou and colleagues said in a press release accompanying the study. “Our scalable methods offer a feasible pathway towards national screening for health risk.”

The most significant limitation to the study, the researchers wrote, is that although the wrist-worn sensors were equivalent to smartphone sensors, the wearing pattern is different, so the results from smartphones may be different.

“Our clinical studies have shown predictive models using only walking intensity can accurately compute pulmonary function for cardiopulmonary patients. Thus, our analysis with wearable sensors for predicting mortality is directly applicable for clinical practice with personal smartphones, already ubiquitous in the U.K. and the U.S. populations, and widespread in global populations.”

References

• Can smartphones predict mortality risk? https://www.eurekalert.org/news-releases/967527#:~:text=Can%20smartphones%20predict%20mortality%20risk%3F%20At%20population%20level%2C,and%20mortality%20risk%2C%20researchers%20conclude%20Peer-Reviewed%20Publication%20PLOS. Published Oct. 20, 2022. Accessed Nov. 3, 2022.
• Zhou H, et al. PLOS Digital Health. 2022;doi:10.1371/journal.pdig.0000045.