For the app to work, users have to tell the app when they're experiencing certain symptoms, such as a cough, runny nose, stomachache, or fever. The app will then go to work identifying patterns between this symptom and the other data the user's Fitbit has been feeding it. 

The next time the app perceives a change in sleep pattern, for instance, that aligns with the user's sleep pattern before the last cold, it will tell the user that he or she is likely getting sick.  

"We expect a user to experience parallel sleep, heart rate, step and active patterns leading up to a sickness," says Michael Mora, CEO of Datapult, in a press release. "Therefore, if their most recent data mirrors their data from when they experienced the same symptoms, our algorithm would calculate a high probability of a repeat illness for this individual calibration. 

The more users interact with the app, punching in every symptom they perceive, the more accurate the app will become at anticipating illness, explains Mora. 

Health apps have been a huge hit among American consumers. Although there are a handful of popular ones that dominate the market, thousands of developers have tried their luck with apps that offer to help users do everything from eat healthier to meditate. 

Some health apps have received critical acclaim, from both technologists and health experts. Many others have been dismissed: One study earlier this year that examined over 1,000 health apps declared that 43 percent of those on iPhones and 27 percent of those on Android phones were functionally worthless.