App-Based Disease Surveillance After COVID-19

post by RedMan · 2020-04-10T18:52:52.941Z · score: 1 (2 votes) · LW · GW · 3 comments

With the smartphone apps for location and contact tracing being normalized for fighting coronavirus, it opens up space for using the same techniques to fight other diseases.

HIV, Syphilis, and drug resistant gonorrhea are obvious candidates because they do not have non-human hosts.

Instrumentation would be easy. Generate a database entry when an individual is diagnosed with one of these pathogens, acquire phone data via disease surveillance authorities, use machine learning directed at that data to identify plausible intimate contacts, and send notifications with testing requirements, thrown against a database of most recent tests.

Mandate response (in the usa this would be analogous to a subpoena) to a notification (appearing at a clinic for testing and treatment) and enforce it as aggressively as corona isolation.

From an ethics standpoint, if app based corona tracing is ethical, this is ethical too. From an effectiveness standpoint, this would likely aid in the eradication of two scourges (Syph, GC) and the close control of a third (HIV).

Please let me know why I'm wrong, or alternatively, if this could be expanded even further to trace other social ills.

3 comments

Comments sorted by top scores.

comment by ChristianKl · 2020-04-12T18:00:09.147Z · score: 4 (2 votes) · LW(p) · GW(p)
From an ethics standpoint, if app based corona tracing is ethical, this is ethical too.

I don't think you have made any argument for why the reader should believe that. It's like saying if it's okay to measures someone's temperature to see whether they have fiver when they enter an office building it should also okay to require them to strip naked because you want to be able to search them for skin cancer.

The kind of privacy that seems to be Western consensus to give up, is that if someone spends a time next to each other an authority that looks at the mobile phones of both persons and has control over the mobile phones within 14 days of the meeting should be able to determine whether they meet.

Additionally, we are unclear in the public debate about whether only a given health authority should have the encryption keeps to make that connection.

The scenario you describe involves giving up a lot more privacy.

comment by RedMan · 2020-04-12T20:23:04.264Z · score: 1 (1 votes) · LW(p) · GW(p)

I would be surprised if you could not figure out if two people are screwing with moderate confidence using nothing but demographic data and location based metadata dumped into a ML algorithm. The price of false positives is a few unnecessary tests, and is therefore super low, so it doesn't even have to be that good of a system.

Tinder data could be purchased to build out the initial algorithm, and if there are still challenges, volunteers could be solicited for validation data.

Mixing in public social media (instagram) and actual communications content might help, but after validation of the location system, probably isn't necessary, but could be analyzed using robots rather than human review, which is apparently acceptable for other purposes.

Is it morally justified to use location metadata (gps), public social media (instagram), communications metadata (contact lists), and communication content to enumerate close contacts that may have spread respiratory viruses? If so, how could it be wrong to use the exact same dataset to fight other diseases with massive social burdens.

I mean sure, some people might cry about their privacy, but the data isn't theirs, courts have established that it belongs to the communication companies, all of whom are apparently on board with metadata assisted surveillance for security and now public health.

Google and Apple are building the Bluetooth tracker, the Chinese gps app with color coding for exposure risk is a thing, facebook checked instagram to see if people in Italy are social distancing. Nobody is crying about any of these things. This is just a proposal to use the same datasets for the same reason.

Anyone who argues can be labelled pro disease and pushee out of the public debate, just like anyone who complains about flu tracking software can be asked, 'do you want old people to die'?

The initial system could be instrumented with a color coding scheme, and an app. When people go to dr offices, part of the basic vitals check at the start of a visit is the doctor running a database check and suggesting testing for various conditions based on the color code. The app to check your own color code status could be downloaded by interested users. 'Show your color' would become something people just ask each other during intimate encounters.

Most jurisdictions already require that positive tests for certain pathogens (STIs are on this list) be reported to a central authority by doctors, this is a long-standing thing and nobody with an opinion that matters questions it:

https://medlineplus.gov/ency/article/001929.htm

Governments could implement this proposal without much public debate by just rolling out a corona app, adding features for different classes of respiratory disease, then adding features for the rest of the 'reportable pathogens' that are transmitted by different means. The model could be developed in house using already available data (reported tests and location metadata).

We can look back at this post in five years and see how things have moved. Good luck stopping it if you think this is morally repugnant as you apparently do.

comment by ChristianKl · 2020-04-13T07:18:12.125Z · score: 4 (2 votes) · LW(p) · GW(p)

There's no location data stored in the bluetooth tracking that Google and Apple are building. I don't think it's a coincidence that Google and Apple are so late to the party of building tacking apps. They were terrified of the potential PR risk and waited till a decent agreement on how privacy should be handeled. That agreement seems to be no location tracking.

Facebook might have created useful statistics about who spreads Corona, but they seem have decided against integrating contact tracing into their apps because it's too risky for them PR-wise.

The Chinese could do things like you are proposing but likely only for their territory. Especially as it requires cooperation of official authorities.