Covid-19 Self-Reporting App

[Andy Kirkham]

Apologies if this has already been mentioned, I'd like to spread the word about this app created by King's College London which allows people to report their own Covid-19 symptoms or the lack of them. https://covid.joinzoe.com/ 

 

Obviously, the more people sign up, the more accurate the results will be.

 

Analysis of data obtained so far has produced a very striking result in the form of an estimate of the number of people (in Britain) with symptomatic Covid-19 on each day. According to their estimates, the peak was reached on April 2nd with a figure of about 2 million - way, way above official figures derived from testing -  and this has fallen by about 70% to date.

 

 

Edited April 18, 2020 by Andy Kirkham

[Reorte]

Bit annoying that it's a phone app instead of just a page to report on, I can't take part as it is.

 

Very interesting results, I always thought that the testing should at least track the absolute number reasonably well even though it wouldn't give any idea about the total.

 

Edited April 18, 2020 by Reorte

[Nearholmer]

2 hours ago, Andy Kirkham said:

Obviously, the more people sign up, the more accurate the results will be.

What I couldn’t understand from a quick skim is how many people are using it currently, which, together with the demographic of users and their geographic distribution will give a feel for whether it is collecting enough, and sufficiently distributed, data to be “believable”.

[burgundy]

30 minutes ago, Nearholmer said:

What I couldn’t understand from a quick skim is how many people are using it currently, which, together with the demographic of users and their geographic distribution will give a feel for whether it is collecting enough, and sufficiently distributed, data to be “believable”.

This link will give you a geographic distribution, if not the level of granularity that you might like or the demographic.  

https://covid.joinzoe.com/post/covid-decreasing-uk

For my area, it shows 4,300 contributors to the app, for a local authority that has a little more than 200K electors (from memory) - so something like a 2% sample. Someone better than me at sums may like to comment on the statistical significance of that sample size.

It may not be perfect, but it seems to be the best way to generate data ahead of the curve, whereas most Government announcements are after the event - hospital admissions and deaths. 

I am no expert, but it seems to me that we shall need some kind of early warning system like this in order to relax lockdown. It looks as though there will inevitably be 2nd, 3rd, 4th waves as controls are relaxed and the sooner flare ups are identified, the sooner constraints can be brought back for those specific areas. For those planning the mother of all parties when this is finally over, I do not believe that there will be a clear moment when we declare "victory". It will be a long process of attrition until a vaccination is available - and probably for some while after that. 

I would strongly encourage everyone to download the app. and contribute to the data.

Best wishes 

Eric 

Edited April 18, 2020 by burgundy

[Andy Kirkham]

1 hour ago, Reorte said:

Bit annoying that it's a phone app instead of just a page to report on, I can't take part as it is.

 

Very interesting results, I always thought that the testing should at least track the absolute number reasonably well even though it wouldn't give any idea about the total.

 

I don't have a smartphone either but I use an Android emulator called Bluestacks on my laptop https://www.bluestacks.com/

[Reorte]

re: my earlier comments about no phone, according to the FAQ on the app's site they're intending to launch a web-based version ("as soon as we have the resources").

 

[Nearholmer]

Nearly 2.5M participants.

[Ozexpatriate]

The University of Stanford conducted an COVID-19 antibody study in Santa Clara County (Silicon Valley area). 

 

The study tested 3,300 adults and children. Based on their findings, infection rate may be 50x what official testing shows.

 

CNN:

Quote

The study estimated that 2.49% to 4.16% of people in Santa Clara Country had been infected with Covid-19 by April 1. This represents between 48,000 and 81,000 people, which is 50 to 85 times what county officials recorded by that date: 956 confirmed cases.

 

[monkeysarefun]

Australia is leading the world in testing numbers on a per capita basis - currently over 1% of the population has been tested, with 406000 tests completed. Given we've got 6500 cases, thats an infection rate of 0.016%, which is probably accurate given that most of our cases are in clusters around identified infected people.

General community transmission seems to have been avoided happily due to the governments surprisingly  (for a pro business conservative party) quick response and not dilly dallying with herd immunity schemes like the UK or having a leader who kept dismissing it as the flu and a democrat hoax, like the US.

 

The government is currently touting a soon to be released voluntary  app that will track everyone who has been in contact with an infected person, but even Barnaby Joyce (National  ie conservative party MP)  is having a bitch about it over privacy  so not sure how successful that'll be, especially given that things here seem generally positive anyway with the daily countrywide new cases rate under 50 per day for over a week.

Edited April 18, 2020 by monkeysarefun

[Andy Kirkham]

Two new features have been added since yesterday:

1. A form where you state to what degree you've been able to self-isolate

2. The ability to report on multiple people.

[Chris hndrsn]

23 hours ago, Reorte said:

Bit annoying that it's a phone app instead of just a page to report on, I can't take part as it is.

 

Very interesting results, I always thought that the testing should at least track the absolute number reasonably well even though it wouldn't give any idea about the total.

 

 

We need to be mindful that only those who are ill enough, paranoid enough, or required to give a sample will be tested, and as we know, not everyone gets symptoms serious enough to notice. Those who are not showing symptoms, that is those who are asymptomatic,  most likely are unwittingly passing the virus on, which is why we are locked down here in Australia, as you are in the UK.

 

Mobile phone apps like the one discussed are an excellent tool for the public health officials in identifying and predicting hotspots and as an aid to contact tracing when someone is later identified as contagious. Even though the user contact data is anonymized it helps the virus carrier/patient during their contact tracing interview/s, confirm their movements and likely contacts.

 

The BBC4 program from 2018 “Contagion” was an excellent predictor and example of how a pandemic spreads so quickly and the use of smartphone apps with Bluetooth. https://www.bbc.co.uk/programmes/p059y0p1

 

Cheers,

Chris

 

Edit: My apologies, I conflated this Kings College app, which is solely self reporting of health conditions with the Imperial College app used in the BBC4 program Contagion: Pandemic.  You were right in addressing your issue of lack of desktop/laptop reporting, they understandably based their decision on the ubiquity of smartphones. The BBC/Imperial College app was a tracker using Bluetooth to record the Bluetooth handshake when two or more phones are in close proximity, which can show potential growth and spread of a pandemic very well and “live”. Our government here in Australia are trying to sell this sort of app to the public at the moment and understandably are having trouble. 

Edited April 19, 2020 by Chris hndrsn
Error in context about the app