Smart Thermometer Data Speed Flu Forecasting

Smart ear thermometer and app (Kinsa Inc.)

9 February 2018. Data collected from electronic thermometers are shown by researchers to track and report the spread of influenza cases in the U.S. faster than federal health authorities, as well as provide more up-to-date forecasts. A team from University of Iowa in Iowa City, working with the company Kinsa Inc. in San Francisco, report their findings in yesterday’s issue of the journal Clinical Infectious Diseases (paid subscription required).

The U.S. is currently in the midst of a widespread influenza outbreak that shows no immediate signs of diminishing. The flu is a contagious respiratory illness caused by infections from a family of viruses that result in symptoms ranging from mild to severe. Young children, older people, and pregnant women are often at greater risk of contracting the flu or suffering complications. According to the latest report at press time from Centers for Disease Control and Prevention, dated 27 January, nearly every state, U.S. territory, and District of Columbia, are reporting higher than normal (statistical baseline) numbers of flu cases.

The Iowa team of internist and epidemiologist Philip Polgreen and computer science postdoctoral researcher Aaron Miller are seeking faster and more detailed data on influenza outbreaks, to achieve as close to real-time reporting as possible. Faster status reports can aid public health authorities forecast trends in the spread of the disease, as well as alert hospitals, clinics and residents. Reports from the CDC, however, describe the status of outbreaks about 2 weeks after data collection, usually from medical records of patients at hospitals and clinics, well after symptoms become evident.

Polgreen and Miller enlisted the help of Kinsa Inc., a maker of electronic fever thermometers. Kinsa offers a line of thermometers for adults and children, cleared by FDA, that connect through hard-wire or Bluetooth to smartphones, with apps that record and track users’ temperatures. The company says more than 1 million people use Kinsa thermometers, recording more than 25,000 readings per day during the flu season.

Kinsa also enables its customers to share their readings for studies like those done by Polgreen and Miller, and for alerts to health authorities. The data are captured directly from the devices, says Kinsa, and thus they provide direct measures of current outbreaks, rather than relying on indirect indicators, such as Internet searches or social media posts. The data are encrypted when shared, with identifying information removed.

For the Iowa team, Kinsa provided data from more than 450,000 ear thermometers, collected from August 2105 to December 2017, totaling more than 8 million individual temperature readings. Polgreen and Miller constructed statistical models to track current suspected influenza outbreaks, and forecast trends in activity up to 3 weeks in advance. In addition, the researchers tracked the duration of fevers and their spread within households.

The authors found the data from connected thermometers correlate directly to CDC’s flu surveillance. The findings show data from Kinsa devices have a very high correlation with CDC’s reports — greater than 0.95, with 1.00 indicating a perfect match. In addition, high correlations also occur across regions and age groups. Moreover, the smart thermometer data could accurately forecast CDC’s findings up to 3 weeks in advance, as well as providing real-time reports.

“Our findings suggest that data from smart thermometers are a new source of information for accurately tracking influenza in advance of standard approaches,” says Polgreen in a university statement. “More advanced information regarding influenza activity can help alert health care professionals that influenza is circulating, help coordinate response efforts, and help anticipate clinic and hospital staffing needs, and increases in visits associated with high levels of influenza activity.”

The data also offer guidance for individual households. According to the authors, flu episodes last 3 to 6 days in individuals, with cases of the flu returning to the same person occurring more often during the flu season, increasing the risk of complications. “Because the device gives users the option of voluntarily reporting anonymized information on sex and age,” adds Miller, “we were able to estimate age-group specific incidence. Our results also suggest that influenza-like-illness moves more frequently from children to adults rather than vice-versa.”

More from Science & Enterprise:

• Low Carb Diet, Biomarker Tracking Help Reverse Diabetes
• FDA Approves Wearable Epilepsy Seizure Detector
• Infographic – Artificial Intelligence in Smartphones
• Little Data to Support Wearable Device Health Benefits
• Medical Tech, Phone App Companies Partner on Wound Care

*     *     *