Skip to main content

It's Sample Size, Stupid

How Apple's new smartwatch could save lives.
(Photo: Apple)

(Photo: Apple)

Apple is intent on showering the global medical research community with an unprecedented amount of data. Yesterday, along with a new health tracking Watch, the tech giant launched an initiative that allows users to voluntarily give out their medical and behavioral information through a simple app: ResearchKit.

In theory, big data—and the access to the daily habits of millions of users—can save lives. Barbara Evans, director of the University of Houston's Center on Biotechnology & Law, argued that the government could have more quickly detected the deadly cardiac side effects of the now-defunct pain medication, Vioxx, had there been a much bigger sample size. (It took 65 months to detect the deadly side effects and pull the drug off the market.)

"With claims data for 100 million people, the problem could have been spotted in fewer than three months," Evans wrote in the Notre Dame Law Review. "If FDA had had the necessary data networks in place to do large-scale observational studies in 1999, all of the people killed or injured by Cox-2 painkillers after August 1999 (i.e., three months after Vioxx went on sale) might have been spared."

Those kind of numbers—in the hundreds of millions—are only possible when most everyone in a given population has the same diagnostic device. Apple has made this statistical fantasy a reality by selling roughly 700 million smartphones.

If Apple's Watch sees a level of success anywhere near that of the iPhone, it'll instantly give researchers more data than they have ever had before.

Apple's first project will be Parkinson's detection. An app asks users to perform a few simple tasks, such as walking in a straight line and saying "ahhhhh" into the microphone. Micro-fluctuations in voice and walking gait are a good indication of illness onset. On stage, CEO Tim Cook said that exercise might help treat Parkinson's, and the new Apple Watch will know exactly how much physical activity each user performs every day.

If Apple's watch sees a level of success anywhere near that of the iPhone, it'll instantly give researchers more data than they have ever had before.

ResearchKit is part of a broader effort to open up more medical data. For instance, Britain's National Health Service is pioneering a controversial government-wide effort to release raw data to approved researchers. Up until now, most medical data has been held up under strict lock and key, for fear that anonymous patients could be re-identified with sophisticated statistics.

Those fears aren't entirely unfounded; Statisticians are getting increasingly clever at finding identities in anonymous datasets. The more data we reveal, the easier it is to identify us. Speaking personally, there are only so many 32-year-old Jews with chronic back pain living in the Mission District of San Francisco.

If I use social media, it gets even easier. A paper recently published in Science found that consumers could be identified from their credit card purchases, especially if they had a public Twitter account.

While few details about ResearchKit have been revealed, it's likely that the data will only be available to select researchers. So, privacy fears are something to worry about in the (distant) future. The benefits, hopefully, will be more immediate.