Over 100-million fitness trackers were sold in 2016 alone, each of them now uploading personal fitness data to the cloud. This has amounted to massive amounts of data about millions of people all around the world.
Companies are already using the feedback from metrics to work out what motivates people best, and how to present the data to increase engagement.
But the implications go well beyond this: they are simply enormous.
Scientific American has summed it up succinctly: “Terabytes of personal health data, amassed daily in stunning quantities. It’s the world’s biggest health study—and nobody’s running it.”
More than just counting steps
The publication goes on to make the key point, which is that we need to take advantage of all this daily individual fitness data and put it in service of health research.
As fitness trackers become more sophisticated and powerful, they are able to collect a wider range of data. The first fitness wearables were little more than pedometers equipped with Bluetooth, they can now measure heart rate, perspiration, the levels of oxygen in the blood, body weight and BMI.
There is another key factor about these devices: because they are wearables, they don’t only collect this data when the wearer is exercising – they do it all the time.
Continuous, real-time global health data collection
This is the single biggest source of continuous daily individual data that has ever been created, and it’s growing exponentially every day.
Of course there are many refinements that need to be made before this data becomes useful. First, the measurement accuracy of the fitness trackers must improve. For instance, the sleep activity tracking is simply based on wrist movement, which is hardly a real indication of sleep state.
Second, the data needs to be aggregated, which means that it must firstly be shared by all who collect it. One can imagine the problems that will first need to be solved, privacy and commercial value being the most obvious.
An ethical and IP minefield
It’s clear that before we are able to make any use of this data we will have to confront a few very tricky questions.
Data ownership is the major consideration. We will either have to spend vast sums of money to procure the data from the companies that have collected it, or we will have to persuade them to donate it to the common good. That’s if we decide that they own the data, rather than the individuals who have uploaded it from their wearables. If the latter do, then it might become a matter of data donation to a common health research information pool.
Finding a way forward
Individual companies are already using telemetry and machine-to-machine communication to monitor and reward individual customer behaviour. Insurance companies, for example, are using in-car devices to monitor driving behaviour and reward people accordingly.
Furthermore, companies like the Google-acquired DeepMind are already working towards a future where machine learning helps to detect the onset on illness before it happens. DeepMind is currently working with the UK’s NHS in implementing its acute kidney injury alerting app, Streams, which uses data from over 1.6 million patients to detect injury onset and alert clinicians in advance.
So we already have a model for the implementation of health Big Data. The immediate focus thus needs to be on who does the analysis, and how. The sooner we can solve the thorny initial issues, the sooner we’ll be able to reap the benefits, globally. The call now should perhaps be, above all, for altruistic intent.
1. Scientific American – 1 Jan 2015
2. Digital Health, 20 March 2017: What does Google DeepMind want with the NHS? (https://www.digitalhealth.net/2017/03/deepmind-mustafa-suleyman-interview/)