There are no big winners in the annual influenza lottery. The best possible outcome requires three correct matches. Researchers must accurately predict the strains that will appear in the next year’s flu. Correct vaccine formulations must be ready in time in sufficient quantities to protect the population. During the flu season, accurate location tracking and prediction gives the best chance for adequate medical preparation and response. So far, no smartphone apps predict virus composition or aid vaccine production. There is hope, however, for smartphone help in predicting where and when the flu will strike next.
Researchers at the University of Iowa’s Carver College of Medicine found that the data reported by smart thermometers via smartphone apps can track and predict flu activity in the U.S. even faster than the CDC. In a study published in Clinical Infectious Diseases, the research team used more than 8 million anonymized temperature readings from nearly 450,000 Kinsa Smart Ear Thermometers. We wrote about the Kinsa Insights data tracking system just last month, but the University of Iowa (UI) study tops that story with documented results from a vastly wider sample. According to the study, CDC data trails real-time flu activity by nearly two weeks at best. UI found that smart thermometer data can detect clinically relevant symptoms before a person goes to the doctor. Using data from August 30, 2015, to December 23, 2017, the UI researchers were able to predict flu activity at least three weeks before the outbreaks, the study reports. By adding data from smart thermometer users who allowed sex and age data in addition to temperature readings, the researchers were able to predict which age groups were most stricken during the season. Cross-checking the UI data with CDC information found a high positive correlation (r>0.95) between the two datasets.
When smartphone apps are paired with smart sensor-based health devices, they can go beyond their initial user reporting function and contribute insights about the health of national or even global populations. The potentials for health and medical discovery and prediction are mind-boggling.
