Harnessing long-term data key to meeting fitness goals

Conducted by researchers from the University’s School of Information Technologies, the study examined the use of physical activity trackers by 21 FitBit users who averaged nearly two years of tracking (~23 months). On average, the participants wore their Fitbits on 68 percent of the days, although six wore them nearly every day, some for three years.

Researchers discovered that although all the participants had been using their activity trackers over a long period of time, they hardly made any use of their years of data.

“While all of the participants had long-term health goals, it had been just too hard for them to get a meaningful view of their data,” said study co-author Professor of Computer Science Judy Kay.

“Just six of our participants viewed data that was older than a month, and none had an easy way to see the long-term data in a form that matched with their goals.”

Interestingly, the study also revealed participants often over- or underestimated their long-term activity levels.

“On average, participants’ step estimates had a 20 percent error and half of the participants had no idea if they were more or less active on weekends than weekdays,” said study co-author Lie Ming Tan, a PhD candidate in the School of Information Technologies.

“Somewhat surprisingly, the participants who saw very consistent activity values – day after day, most days – recalled their activity level at a similar accuracy to the participants who did not wear the FitBit regularly.”

In the study, the authors introduced a new custom interface – called ‘iStuckWithIt’ – that visualises a person’s long-term activity levels.

The study demonstrated that iStuckWithIt enabled people to discover new insights from their long-term physical activity – even when data was incomplete.

For example, one study participant realised that they had become less active when they moved from a central location to a suburb. That prompted them to consider moving back to the city.

“Devices like a FitBit do a good job of helping a person monitor each individual day’s activity and that may act as a trigger to do more activity. But our work explored how to create a useful interface for long-term data since it has the potential to play other important roles in reflection, goal setting, monitoring and planning,” said Professor Kay.

“Long-term data has the potential to enable people to self-reflect on their activity levels achieved over the long term, exploring patterns and features to gain insights into the factors that may have impacted their behaviour.”

iStuckWithIt can show various forms of long-term data and it takes account of when a person wears their activity tracker and for how many hours, while embedding activity data – in the form of steps or active minutes – to support achievement of goals.

“A user could use iStuckWithIt to become aware of their recent and long-term activity levels, accounting for changes and patterns over months, seasons and years,” said Mr Tang.

The study co-authors said this long-term data could also be used to support interactions between users and medical practitioners or health advisers.

They also highlighted opportunities to extend the current interface through integration with a user’s email and calendar to access more contextual information (such as work hours versus free time), or through features to record or highlight special periods (such as training for a marathon).

“The design of iStuckWithIt is very flexible. It also could be used to display information such as a user’s heart rate and sleeping patterns,” Professor Kay said.