Book Review
Counter Attack
Are data analysts ending privacy as we know it?
By Charles R. Meyer, M.D.
“In which city is Guerrero Street?” That multiple choice question during my online application for an Allscripts e-prescribing program gave me a mild case of the creeps. Before encountering this startling query, I had entered the usual demographic information—my name and address, plus DEA and medical license numbers. In an attempt to verify that I was the Charles R. Meyer, M.D., who was permitted to prescribe medications, the site walked me through a series of questions (the answers to which were likely only known to me), which Allscripts’ computers had gleaned from publicly available data mere seconds after I submitted my demographics. Not only did they know my daughter had lived on Guerrero Street in San Francisco, they also knew I had bought an Audi in the past several years. Suddenly, I had a tiny, somewhat disturbing glimpse of the data trail I had left over the years.
In The Numerati, journalist Stephen Baker says our digital tracks are much more extensive than the few revealed in my Allscripts application, that in our Internet-centric, credit card-dependent, cell phone-obsessed culture, we leave thousands of them every day. Websites record our “cookies” when we visit them. Credit companies know our buying and borrowing habits. And cell phones have become an organic arm of the Internet, subject to the same snooping. Every move we make in these arenas generates numbers—how many times we visit certain sites, how much we buy, and even how long we linger at given spots. Each month, Yahoo gathers 110 billion pieces of data about its users as each person visiting Yahoo sites leaves a trail of 2,520 clues.
Watching and measuring our every move is a group of mathematically talented analysts that Baker dubs “the Numerati,” who feed eagerly on the data gathered by our digital wanderings and generate models they hope will predict our behavior—and our buying habits.
Baker explores the Numerati’s intrusion into the world of workers, shoppers, bloggers, terrorists, patients, and “lovers” (those using online matchmaking sites). In one example, the Numerati analyze and model grocery shoppers using “smart” grocery carts that record individual choices over time and eventually direct shoppers to their preferences based on a year’s worth of purchase details. These smart carts are automated personal shopping assistants. The paradigm familiar to most of us is Amazon or Netflix, where, after a few book purchases or video rentals, our next site visit starts with “We thought you’d be interested in these books/movies/associated paraphernalia.” Although most of us would like to think we are unpredictable and unique, the Numerati are generating a veritable “mathematical modeling of humanity.”
The Numerati’s calculations are based on clusters of behavior, and precise accuracy isn’t the goal—yet. If they’re watching political behavior, close is good enough if they can categorize accurately 75 percent of potential voters. If, like Umbria Communications in Colorado, they’re parsing language on blogs to sample opinion on products as mundane as deodorant, probability trumps precision. “Truth is not a make-or-break test for the Numerati,” Baker writes. “They triumph if they come up with better, quicker, or cheaper answers than the status quo. Google, for example, doesn’t provide definitive answers.”
One wonders whether such approximations could apply to medicine. Yet Baker describes how Eric Dishbaum at Intel’s Portland, Oregon, research lab, is applying such statistical modeling to medical care. Dishbaum is developing sensors that will “track our pathways in the house, the rhythm of our gait … diagram our thrashing in bed and chart our nightly trips to the bathroom—perhaps keeping tabs on how much time we spend in there. Some of these gadgets will even measure the pause before we recognize a familiar voice on the phone.” His theory is that after a pattern for a given individual is established, a change in gait might signal a weak leg or the shorter steps of impending Parkinson’s. Any practicing physician realizes that a 15-minute look at a patient in the office is a snapshot of that person. Home sensors analyzed by the Numerati might enhance that look and help us discover disease brewing before classic signs and symptoms appear.
Like my Allscripts questions, some of this is more than a bit creepy—Big Brother with not just a calculator but legions of computers. Yet Baker is optimistic that privacy won’t disappear in part because the analyzers, like all of us, also get analyzed. “The people in the best position to exploit our privacy are also gaining an intimate understanding of how their own privacy can be trampled,” he writes. So, as we keep leaving digital tracks, we need to keep track of the trackers. MM
Charles Meyer is editor in chief of Minnesota Medicine.