How Your Smartphone Data May One Day Predict How Long You Will Live
If you could know when you were going to die would you want to? In the movie "Don't Look Up," tech conglomerate Bash has smartphone software with algorithms that can read users' health data, fitness routines, and mood. Bash founder Sir Peter Isherwell reveals to professor Dr. Mindy that he knows more about him than Mindy knows about himself. User data even allows him to predict when and how users will die based on data collection, including buying habits and lifestyle patterns. His prediction about Mindy was wrong, but still, it shows what the power of data can do. And, armed with the right data, we could make lifestyle changes that affect how and when our lives will end.
While predicting your own demise may sound like Sci-Fi fodder, Wired says smartphone apps already record much of our daily activity and lifestyle habits that could affect health. And fitness trackers and apps, of course, track fitness activity. According to WebMD, accelerometers that track movement are built into even the cheapest smartphones. And soon, researchers may be able to use motion sensors and other data from your everyday activities to predict mortality. A recent study, published in PLOS Digital Health, shows how your smartphone data may be able to predict how long you will live.
How smartphone data might predict health and mortality
The study gathered data from 100,000 adults aged 45 to 79 in the U.K. Biobank who wore wrist sensors continuously for a week. Researchers analyzed user data from 12 consecutive walking sessions of 30 seconds each, totaling six minutes per day. Bruce Schatz, Ph.D., an expert in medical informatics at the University of Illinois and a co-author of the study, says, "It's well known that people [who] move more — and move more vigorously — live longer. We ended up trying to see what you could tell from walking motion that had some medical significance." (via WebMD).
Using walking intensity as a yardstick, researchers predicted the risk of death in each following year for every participant. Over the next five years, predictions were quite accurate compared to U.K. death records. But there are caps on what this tech could do. Schatz says, "It's not giving you, personally, 'You have 5 minutes to live,' [rather] what's the likelihood that you'll die in 5 years, or in 2 years?"
The study has some limitations, including racial diversity and socioeconomic status (per Science Daily). And the same data that could predict mortality could also wind up in the hands of employers, insurers, and other agencies that could severely impact your life, causing privacy concerns for many would-be users (via American Medical Association). So while predicting your own death could become feasible in the future, it seems there are many kinks to be worked out just yet.