The latest project to emerge from the initiative is a huge new study that was tracking physical activity of people from over 100 countries via anonymized smartphone data. The huge dataset allowed researchers to analyze step data according to country, gender and community type.
“Big data is not just about big numbers, but also the patterns that can explain important health trends,” the researchers say. “Data science and modeling can be immensely powerful tools. They can aid in harnessing and analyzing all the personalized data that we get from our phones and wearable devices.”
Big data, big healthcare
The data was harvested via the free Azumio Argus app, which is used for tracking physical activity via the sensors built into all smartphones. In total, some 68 million days of data were generated, with researchers given minute-by-minute insight into the activity of over 700,000 people from 111 countries.
“The study is 1,000 times larger than any previous study on human movement,” the authors say. “There have been wonderful health surveys done, but our new study provides data from more countries, many more subjects, and tracks people’s activity on an ongoing basis in their free-living environments versus a survey in which you rely on people to self-report their activity. This opens the door to new ways of doing science at a much larger scale than we have been able to do before.”
The team used the Gini index that’s traditionally used to calculate income inequality to measure activity inequality in each country. The hope was that understanding activity distribution would provide crucial insights into the health diversity of a country.
Across the world, the average person recorded 5,000 steps per day, with a clear correlation emerging between activity levels and known health issues in a country, such as obesity. Individuals in the five countries with the greatest activity inequality are nearly 200 percent more likely to be obese than individuals from the five countries with the lowest activity inequality.
Some other interesting trends emerged. For instance, whilst activity levels averaged the same in Mexico and the United States, the US had a much wider range of activity, which the researchers believe explains the higher obesity levels in the US.
Such insights could prove invaluable in helping to target interventions. When the researchers tested this hypothesis via a computer simulation, it resulted in a four times greater reduction in obesity compared to un-targeted approaches.
Interestingly, the levels of activity variation also seemed to coincide with the the activity levels of women in the country’s population. When women were largely inactive, high activity inequality was common. When both sexes were similarly active however, inequality was often much lower.
The researchers wanted to examine the potential impact of the walkability of a city on activity levels.
“If you must cross major highways to get from point A to point B in a city, the walkability is low; people rely on cars,” they say. “In cities like New York and San Francisco, where you can get across town on foot safely, the city has high walkability.”
When analyzing data from the United States, there was a clear correlation between the walkability score of a city and the levels of activity inequality. Higher walkability is associated with significantly more daily steps across all age, gender, and body-mass-index categories. However, the researchers found that women recorded comparatively less activity than men in places that are less walkable.
It’s a great example of how simple, user generated data can be used to make healthcare smarter and more targeted. Whilst things like user privacy and data governance are hugely important, the promise is clear should these crucial steps be undertaken correctly.
The Morning Email
Wake up to the day’s most important news.