Longitudinal data from sensors help assess health risks

Trillions of data points collected by wearable sensors now can be translated into empirically verified measures of health risk and benefit for patients, which can be used to quantify and enhance their length and quality of life.

That’s the contention of researchers who say they are the “first scientists to successfully translate a common piece of information from wearable sensors—step count—into a verifiable measure of health risk,” while estimating its impact on health and longevity.

The study, funded by Lapetus Solutions—a Wilmington, N.C.-based company that leverages analytics and cloud computing to more accurately predict mortality, morbidity and healthy lifespan—was published this month in the journal Computer.

“There are hundreds, if not thousands, of companies that are generating data from wearable sensors. However, most of that data isn’t used for anything,” says Jay Olshansky, chief scientist at Lapetus Solutions and professor at the University of Illinois at Chicago School of Public Health.

“We took it upon ourselves to develop the algorithms that are required to translate data from wearable sensors into something useful—in this case, we translated data from step count into actual risk of death,” adds Olshansky, who contends that the metrics have significant value for healthcare providers and patients.

To calculate the risk of death, expected gain in life expectancy and healthy-life expectancy, researchers used a plethora of data, including step count, age, sex, height, weight, walking speed, stride length, steps per mile as well as calories burned.

“We know that people who are sedentary have a significantly higher risk of death than people who are physically active,” says Olshansky. “It’s the level of physical activity that has been well established in the scientific literature that’s linked to risk.”

However, he believes that step count is just the beginning when it comes to crowd-sourced data that can be translated into empirically verified measures of risk to assist in disease prevention at both the individual and population level.

The problem is that physicians rely on annual physicals to assess their patients’ risk factors for disease “at a single moment in time,” according to Olshansky. Yet, by tracking patients using wearable sensors over the course of a year, physicians could gain valuable insight into hard- to-detect factors contributing to patients’ health problems.

“Longitudinal health data is the gold standard for understanding and evaluating the relationship between various behavioral and environmental risk factors and health,” states the study. “Data from wearable devices can be used to generate charts that track all sorts of health barometers across time—sleep/wake patterns, blood pressure, blood sugar, physical activity, periods of sedentary behavior, and so on.”

This rich source of data from wearable sensors could not only give physicians a more accurate picture of their patients’ health and risk factors, but could be aggregated with other patients’ data for population health management, concludes Olshansky.