To see if wellness applications, mobiles and wearable devices improve the user’s health,
It does not hurt to ask the obvious questions. Do wellness applications, mobiles and wearable devices truly improve our health? Are they life changers, or merely ubiquitous gadgets?
In response, one is likely to get more opinions than answers, and these opinions are often divided. Harvard Medical School credits wellness apps with encouraging consumers to pay attention to what they eat and how much they exercise, saying “consistent use can make a difference in life”. On the other hand, a frequently cited 2016 study of 800 full time employees from 13 organisations in tropical Singapore somehow concluded that “fitness devices don’t improve health”.
So far, so clear. But, in such a data-rich context, opinions must be evaluated with reference to actual use. So, what does the data say? To better understand the effectiveness of wellness programmes, Umanlife, a health and wellness application provider based in France, analysed more than 6 million tracking data points from its 25,000 French users over a period of five years.
Secure data flow
The Umanlife study aimed to measure whether and how wellness apps contribute to health factors like weight loss, sleep, and physical activities.
First and foremost, strictly non-personalised user data is extracted from the backend platform (see Figure 1 below), then aggregated to reflect every user’s health and wellness journey.
Most of the 25,000 users subscribed in the period 2012-2017. The age and gender distributions are largely consistent with the French national demographics. Gender split is almost 50/50, while family subscribers are skewed at age 35+ as expected. For connected devices, 60% integrated the application with smart scales, 12% with running apps, with the rest choosing a variety of options including Garmin and Fitbit.
Over 60% of users reported weight loss after the first one or two months, this level was maintained at month 5,7 and 10 as well. While motivation levels do seem to fluctuate from month to month, consistent use of the application records an overall increasing trend both on the chance and the extent of weight loss. The early 30s and 40s appear to have slimmed down the most (5 kg per person) among all weight losers. (See Figure 2)
Twenty percent of users who uploaded blood pressure data started with high blood pressure risk, ie systole level over 135 or diastole level over 85. Over a period of two years, these users recorded 6 points lower systole per person and 9 points lower diastole per person. In the same period, over 95% of users with normal BP stayed in the healthy range.
Initially, 32% of users slept less than 6 hours per day, below the 7-9 hour range for healthy adults. Over a period of 12 months, these short sleepers recorded more than 1.6 hours of additional sleep per person per night. The application also shows that 85% of the healthy 7-9-hour range remained in the range. (See Figure 3).
In addition to the health indicators specified, the application also recorded changes in self-reported behaviours like smoking and active exercise. Although such data is perceived to be more subjective and less reliable, long-term data tracking reveals findings consistent with monitored activities.
Twenty percent of users are recorded as smokers, compared to French national statistic of 27.7%. The app has not helped many smokers quit smoking completely – however, heavy smokers are generally smoking fewer cigarettes per day.
Sixty-seven percent of physically inactive users (less than 2,000 steps per day) have consistently recorded more steps per day. (See Figure 4)
This analysis is not suggesting that wellness apps are a “silver bullet” for everyone, nor is it about a singular yes/no vote for hi-tech apps or devices. It is about taking an analytic approach to measure how utilising data (after all, data is only useful when used) can help track and measure individual performance, which then contributes to increased self-awareness and self-monitoring, followed by motivation to maintain or change.
Advanced analytics can go further, helping the user and provider to understand “people like me” through user profiling and segmentation, or to predict health trends and formulate directed actions in advance. For example, what are the key attributes of people who did lose weight and people who did not? In this case, a multi-variant predictive model shows that it is mostly down to the human factors: users with the highest statistical chance of weight loss tend to be female, aged 30-50, a longer-time app subscriber who also logs sleep and activity data!
Broader application of these types of user analytics could bring more engagement opportunities for insurance carriers, such as target marketing, risk prediction, claims reduction, in-force wellness management, to name a few. A
Ms Mandy Luo is ReMark’s Chief Actuary and Head of Data Analytics.