Donate to Science & Enterprise

S&E on Mastodon

S&E on LinkedIn

S&E on Flipboard

Please share Science & Enterprise

Symptom App, Algorithm Predict Covid-19 Infections

Android phone

(Pexels, Pixabay)

11 May 2020. Data from millions of smartphone app users in the U.K. and U.S. reporting on Covid-19 symptoms show they could estimate a population’s Covid-19 infections. Results of the analysis by researchers at Kings College London and the company Zoe Global Ltd. in the U.K., and Massachusetts General Hospital in Boston are reported in today’s issue of the journal Nature Medicine.

The Covid Symptom Study tracks reports of Covid-19 symptoms from smartphone app users. The free app became available on 24 March and is now used by more than 3 million individuals, according to its developers at Zoe Global. As reported by Science & Enterprise on 31 March, people who download the app from the Apple or Android stores are asked to give a summary of their overall medical condition, then answer a few more questions each day about their health. The researchers believe there are more symptoms to Covid-19 beyond the coughing, fever, and breathing problems mentioned most often. Other symptoms including headaches, muscle pains, loss of smell or taste, and diarrhea are also being reported by some patients.

The Nature Medicine paper reports on an analysis from data contributed by 2.5 million users in the U.K. and some 168,000 in the U.S. as of mid-April. Of the total app users sending in data, about one in three reported symptoms associated with Covid-19 infections. Among all app data reporters, 18,400 also took a Covid-19 diagnostic test — the so-called gold standard RT-PCR SARS-CoV-2 test — with about 7,200 or 39 percent of those testing positive.

One symptom reported by app users registered a particularly high association with a positive Covid-19 diagnostic test: anosmia, or loss of smell and taste. About two-thirds of people testing positive also reported anosmia as one of their symptoms, while less than a quarter in the U.K. and less than one in five in the U.S. testing negative also reported these symptoms. Persistent cough, chest pains, and skipped meals indicating loss of appetite were reported by four in 10 to half of app users testing Covid-19 positive.

Zoe Global that developed the app with the academic researchers was founded as a nutritional science company collecting data from blood tests and glucose monitors to track metabolic responses to various food inputs, including glucose and fat levels in the blood. The company’s nutritional and test data are used to produce algorithms that predict individual responses to different foods, and in this case researchers applied those kind of algorithms to predict a person’s likelihood of Covid-19 infections, based on symptom reports.

A statistical model developed by the team analyzed symptom reports and diagnostic test results, and identified a set of symptoms beginning with anosmia, but also fatigue, persistent cough, and loss of appetite, combined with age and sex, that predict Covid-19 infections. When applied to the nearly 806,000 app users experiencing symptoms, but did not take a diagnostic test, the analysis indicates about one in six (17%) were likely to have a Covid-19 infection, which researchers say translate to about five percent of the population at large.

“Our results suggest that loss of taste or smell is a key early warning sign of Covid-19 infection and should be included in routine screening for the disease,” says Kings College epidemiology professor and senior author Tim Spector in a university statement. “We strongly urge governments and health authorities everywhere to make this information more widely known, and advise anyone experiencing sudden loss of smell or taste to assume that they are infected and follow local self-isolation guidelines.”

The researchers acknowledge shortcomings in the source data, notably self-selection and high proportion of female participants, about two-thirds of the total. The team is planning two clinical trials with the symptom reporting app combined with antibody tests, including one study in Boston among high-risk participants.

More from Science & Enterprise:

*     *     *

3 comments to Symptom App, Algorithm Predict Covid-19 Infections