Innovative Solutions Predictions

Predicting Hot Spots

Because COVID-19 has an average 5-day incubation period (ie humans can spread for 5 days before showing symptoms), it’s given epidemiologists quite the headache to stop spread before it happens. It’s been imperative and incredible to watch how epidemiologists have worked with other industries to predict future COVID19 hot spots.

Twitter and Google: Scientists found a way to predict COVID19 cases and deaths hotspots 2-3 WEEKS BEFORE they happen. How? Through Twitter and Google internet activity. If the number of tweets with the words related to COVID19 (like covid, corona, epidemic, flu, influenza, face mask, spread, virus, …) start increasing, we can predict cases 3 WEEKS in advanced and deaths 4 WEEKS in advance. They also found that by counting the number of people that searched for “fever” in google can predict COVID19 deaths 22 days prior to a spike.

Kogen et al. (2020). An Early Warning Approach to Monitor COVID-19 Activity
with Multiple Digital Traces in Near Real-Time.

Other innovative solutions? I posted previously on a few methods that predict hotspots 5-7 days BEFORE they arrive:

Fecal matter: In sewage systems. Epidemiologists in New York found that high levels of the virus in sewage can predict outbreaks 7 DAYS BEFORE the outbreak occurs. AND, not only outbreaks but can predict an increase in hospital admissions. In other words, if we test sewage systems, we can tell a week beforehand whether there will be an outbreak in that city. Most recently, this has been an effective method in airplanes and cruise ships too.

Symptom tracking: Some of my colleagues are using a symptoms to predict outbreaks 5 DAYS BEFORE they start. They are simply asking if people are well or sick through an app, and if they are sick asking what symptoms do they have. This strategy has been highly effective in the UK.

Google mobility data: I’ve posted phone data extensively in the past. But, most recently, we were able to predict the recent second “spike” in the southern states by seeing how people’s movement to non-essential businesses changed.

There are other ways I’ve heard, like credit card useage, but I haven’t seen published science on this yet.

Without effective drugs or vaccines, these are really useful strategies to deploy public health efforts BEFORE a hotspot hits. Can’t wait to hear what else scientists come up with!

Love, your local epidemiologist

Important Note: Figures come directly from the article. I did NOT create these.

Data Sources:
Google/Twitter article:…
Sewage article:…/10…/2020.05.19.20105999v1.full.pdf
Symptom tracking article:…

Leave a Reply

Your email address will not be published. Required fields are marked *