Could Google predict your next sick day?

Your next sick day could come from a cold, or stress, or allergies, or anything. The unpredictability of being derailed is what makes sick days stressful.

You won’t know what will hit you, necessarily. Thanks to Google’s technology, however, we might have a shot at predicting when it will be.

According to CNBC, the Google Brain team is teaming up with hospitals to sort through millions of medical records using machine learning, looking for clusters of reasons and times that people get sick. The institutions reportedly removed personal information  from the files, making it impossible to pinpoint patients’ identities from the documents.

Google’s technology just might make it easier to tell when illness is coming your way in the future, which would already be a huge advance in knowledge.

But where Google’s machine learning could come in handy as well is help employers and employees understand the eternal problem of sick days.

Sick days have a significant impact on the U.S. economy. The U.S. Integrated Benefits Institute reported in 2012 that bad health sets the US economy back $576 billion, and that 39% ($227 billion) of the money lost is due to a decline in productivity: people being out of work because of sickness or lowered performance.

That is, with a few catches.

How it’s going so far

Katherine Chou, the head of product at Google Brain, told CNBC more about the technology in an interview, only promised that the technology can “improve predictions,” not that it will be all-knowing.

Just in case you’re scrambling to see when this technology will be readily available, you might want to pump the brakes for a bit. CNBC added that until the data is officially sifted through, the outcomes won’t be revealed.

Workers avoid taking sick days

Another catch: most people who get sick enough to stay home from work aren’t actually sick enough to go to the hospital. So Google’s technology may be limited in its usefulness for employees and employers.

In fact, people are more likely not to stay home at all. It’s common for people to work through their illness at work in the office.

A 2016 survey by NPR, the Robert Wood Johnson Foundation and Harvard T.H. Chan School of Public Health provided data on how many people work through their sickness instead of not showing up for work.

Researchers interviewed 1,601 employees in the U.S. who work 20 or more hours per week.

The study found that “adults in low-paying jobs are more likely to say they go to work when sick,” adding that 65% of participants reported going in “always or most of the time” when battling the flu. Fifty-five percent of participants with jobs with average salaries reported doing so, and 48% with high salaries did.

But it looks like people in specific positions that are heavy on face-to-face interactions also come to work sick anyway.

“Half of restaurant workers and more than half of workers in medical jobs say they still go to work always or most of the time when they have a cold or the flu. Many workers have also had experiences in caring for family members who were seriously ill, injured, or disabled while working at their current job,” the report said.

That could be for a variety of reasons: because employees are afraid of losing pay, since companies can already skimp on the sick days they allow employees. There are no federal legal requirements for paid sick leave in the U.S.

Whether come in sick or would rather stay home, Google Brain’s technology might be an option for thinking about your health concerns someday.