[UPDATE] Could symptom-tracking app data help predict potential COVID-19 cases?

Researchers have described their development of a mathematical model informed by data gathered from the COVID-19 Symptom Tracker mobile app, which may be able to help predict the likelihood of an individual having the virus based on the symptoms they present.

Like 0 Comment

Jump to:


Want regular news updates direct to your inbox? Become a member on The Evidence Base®, for FREE, now>>


[UPDATE] Could symptom-tracking app data help predict potential COVID-19 cases?

Published in Nature Medicine, researchers from Massachusetts General Hospital (MA, USA), King’s College London and ZOE (both London, UK) describe their development of a mathematical model informed by data gathered from the COVID-19 Symptom Tracker mobile app, which may be able to help predict the likelihood of an individual having the virus based on the symptoms they present.

The COVID Symptom Tracker app was launched in the UK on 24 March and subsequently in the USA on March 29; it has had more than 3 million downloads worldwide.

The app requires individuals to self-report daily on any symptoms they may be experiencing, or lack thereof; it was designed to track the onset and progression of COVID-19 symptoms, to help identify individuals most at risk of developing the diseases and how it spread.

In this study, researchers analyzed data from 2,618,862 app users from the UK and USA; approximately one-third of these users had reported experiencing symptoms known to be associated with COVID-19. Of these, 18,374 users reported having received a lab-based test for the virus, with 7178 individuals having tested positive.

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,”

– Professor Tim Spector, King's College London

By integrating and comparing information on reported symptoms with positive results from traditional lab-based test results, investigators determined which symptoms commonly associated with COVID-19 were most likely to be associated with a positive test result and developed a mathematical model based on this to help predict whether an individual who presents with certain symptoms is likely positive for the virus. This may be particularly useful in areas where testing, or access to testing, is limited.

Investigators observed a wide range of symptoms to be associated with COVID-19, cautioning against focusing only on fever and cough; for example, anosmia – loss of taste and smell – was a common symptom observed in two-thirds of app users who reported having received a positive test for COVID-19.

The team's model utilizes information on individuals’ age, sex, BMI and presentation of four symptoms – anosmia, severe or persistent cough, fatigue and skipping meals – to predict the likelihood of COVID-19 infection with close to 80% accuracy.

The model was subsequently applied to 805,753 app users who had self-reported COVID-19-associated symptoms, but who had not received testing for the virus. The model predicted that 17.42% of these individuals, likely had COVID-19.

Limitations of the study include the self-reported nature of the app data, which is not as accurate as physiological evaluation. Further, lab-based testing is likely to only be performed for individuals who present with severe symptoms, which could impact the generalizability of these findings.

The study authors suggest that the combined widespread use of the app and model could help identify individuals who are likely to be infectious as soon as their earliest symptoms start to appear.


Sources:

Menni C, Valdes AM, Freidin MB et al. Real-time tracking of self-reported symptoms to predict potential COVID-19. Nat Med. doi:10.1038/s41591-020-0916-2 (2020) (Epub ahead of print); www.kcl.ac.uk/news/new-ai-diagnostic-can-predict-covid-19-without-testing


COVID-19 Symptom Tracker provides valuable, real-time epidemiological data on the disease, report suggests

Published in Science, an international team of researchers led by Andrew Chan (Massachusetts General Hospital (MA, USA) describe their analysis of early data from the COVID-19 Symptom Tracker mobile app, highlighting the potential of the app to provide real-time epidemiological data on the disease that could help inform policy and resource allocation, and supplement information about COVID-19 that scientists have struggled to glean from lab-based procedures alone.

The COVID Symptom Tracker leverages the established digital backbone of an app used for personal nutrition studies and was launched in the UK on 24 March, and subsequently in the USA on March 29. Developed by the COronavirus Pandemic Epidemiology (COPE) consortium – including researchers from Massachusetts General Hospital, the Harvard T.H. Chan School of Public Health (both MA, USA), the healthcare science company Zoe Global Ltd. and King's College London (both London, UK) – the app aims to track the onset and progression of COVID-19 symptoms, to help identify individuals most at risk of developing the diseases and how it spread.


Read more about the launch of the app here>>


By Utilizing pushed software updates, the app developers are also able to add or modify the questions users are required to answers daily in real time, enabling them to test emerging hypotheses about COVID-19 symptoms and treatments.

In this report, investigators analyzed data on symptoms from individuals who self-reported their data on the app during its initial launch period. Researchers observed that complex symptom presentation – with a cough and/or fatigue, in combination with at least one additional symptom such as diarrhea and anosmia – was more common amongst users who reported testing positive for the virus compared with those who reported negative test results.

Interestingly, researchers obserbed that anosmia – loss of smell – was a more common symptom presented by individuals who tested positive for the virus compared with fever, suggesting anosmia may be a more sensitive symptom for virus positivity.

Based on such symptom data from 2 million initial app users, researchers generated a weighted prediction model.

This work has led to the development of accurate models of COVID-19 infection rates in the absence of sufficient population testing,”

– Andrew Chan, Senior study author (Massachusetts General Hospital).

The model was utilized to successfully predicted two spikes in the number of confirmed COVID-19 cases in South Wales; this was achieved an average 5–7 days in advance of confirmed cases reported by public health authorities.

In the study, the authors stated: “These results demonstrate that this app prospectively captures the dynamics of COVID incidence days in advance of traditional measures, such as positive tests, hospitalizations, or mortality.”

Currently, researchers are planning additional studies of individuals who will undergo uniform COVID-19 testing, in order to validate their approach to such symptom-based modeling of disease incidence.

“These data demonstrate compelling evidence for the potential predictive power of our approach, which will improve as more data are collected to inform the model. Further, they highlight the potential utility of real-time symptom tracking to help guide allocation of resources for testing and treatment as well as recommendations for lockdown or easement in specific areas,” the authors concluded.

Chan commented: “This work has led to the development of accurate models of COVID-19 infection rates in the absence of sufficient population testing. For example, the UK government has acted upon these estimates by providing advanced notice to local health authorities about when to expect a surge of cases.”


Sources:

Drew DA, Nguyen LH, Steves CJ et al. Rapid implementation of mobile technology for real-time epidemiology of COVID-19. Science. doi:10.1126/science.abc0473 (2020) (Epub ahead of print); www.massgeneral.org/news/coronavirus/research-tracker-app

Go to the profile of Ilana Landau

Ilana Landau

Assistant Editor, Future Science Group

184 Contributions
2 Followers
0 Following

No comments yet.