Adrian WoizikMay 13, 2023
Ben Adida

I want to give a shout out to exposure notifications, the API Apple and Google created at the beginning of the pandemic. Well scoped. In close collaboration with domain experts. Privacy protecting. Measurably helped, maybe not as much as hoped, but still. And now being responsibly shut down, as promised.

We should praise folks when tech is done right. I want to see more of this responsible deployment of tech.

May 13, 2023 at 1:16amWeb
Show threadEvan ProdromouMay 13, 2023
@ben Ben, aren't there other communicable diseases that we could prevent with this technique, though?
Show threadBen AdidaMay 13, 2023
@evan my sense is it's most useful for emerging diseases or for diseases that have significant asymptomatic transmission. Plus you need a robust testing infrastructure, cause self reporting is weak and leads to false signals. So... Maybe? But i like that they're using it conservatively.
Show threadEvan ProdromouMay 13, 2023
@ben that makes sense!
Show threadFinal Round Player 😷🇪🇺🍸May 13, 2023

@ben @evan The ZOE project in the UK shows that self reported tests still provide good data. The UK NHS app was altered late in the day to allow self reporting of OTC LFTs.

What I'd really like to understand is why the Google-Apple API was limited to one regional health authority. And separate but related, what the server side component is. It looks from the outside like we could have had a competitive marketplace for decentralised UI apps all interoperating via the API. But maybe not.

Show threadFinal Round Player 😷🇪🇺🍸May 13, 2023

@ben @evan Mainly it seems really sad that worldwide, the API is being abandoned, when #covidisnotover

And finally, what potential is there for the same code to be used for other highly infectious (airborne) diseases? Like Flu, RSV, Norovirus.

Show threadBen AdidaMay 13, 2023
@jbond @evan i guess I'm not bought into the idea that there would be good enough signal to noise ratio through self reporting of diseases where we don't do a ton of real diagnostic testing. Who gets a flu test? Most people just say they have the flu when they feel crappy.
Show threadFinal Round Player 😷🇪🇺🍸May 13, 2023
@ben @evan A part of this would be public education that flu (and RSV and Norovirus ) are highly infectious and deeply unpleasant diseases that you should try hard not to pass on. It doesn't matter if the tests and reporting are a bit vague and prone to false positives, if it encourages people to voluntarily self isolate if they feel ill. Via peer pressure from the people who get the warning.
Show threadBen AdidaMay 13, 2023
@jbond @evan that doesn't sound workable to me. People aren't going to prioritize public good if it's too vague or approximate.
Show threadShadSterlingMay 13, 2023
@ben @jbond @evan I think the solution is to make real diagnostic testing readily available to everyone. Maybe have the threshold be in annual deaths. For the U.S. if you set it to 40k the flu is on the line and covid still needs coverage. Not comfortable letting 40,000 people die of a single preventable disease every year? - then set the threshold lower
Show threadKevin RussellMay 15, 2023

@ben @jbond @evan

Waste water monitoring is a data provider, could be even greater. It is privacy respecting. Measurable data.

The biggest problem is what to look for.

Scientists recently discovered thousands of "new viruses" -- thousands, in baby poop.

The discoverers are calling it a "virome" after the popularization of biome (see yogurt ads).

Many of the found viruses are recognizable, the majority are bacteriophages.

Astounding enough, but its taken too many decades to notice.

Show threadFinal Round Player 😷🇪🇺🍸May 15, 2023

@kevinrns @ben @evan I believe the ONS covid survey in the UK did use waste water monitoring as one of it's inputs. But that survey has been closed down.

And of course there's plenty of science still to be done on all the weird stuff in the biological margins.

Show threadKevin RussellMay 15, 2023

@jbond @ben @evan

There were waste water monitoring in the US and Canada

On the biological margin....
https://theconversation.com/thousands-of-unknown-viruses-discovered-in-baby-poo-why-this-is-not-necessarily-a-bad-thing-204419

Thousands of unknown viruses discovered in baby poo – why this is not necessarily a bad thing

Babies guts found to have ten times as many viral species as bacterial species.

The Conversation
Show threadColorblind CowboyMay 13, 2023
@evan @ben We haven’t even ended the pandemic with this one.
Show threadChad BrubakerMay 13, 2023
@ben thanks! :)
Show threadKate Nyhan is changing serversMay 13, 2023
@ben Responsibly shut down as promised? Ok, if we have achieved a threshold of community transmission under which an *opt-in* exposure notification system is somehow *irresponsible,* is there also a threshold that will trigger its reopening?
I mean, I would be thrilled if there is no winter wave, and if there is no new variant with even worse outcomes or an even worse attack rate - but would you put money on it?
Show threadKate Nyhan is changing serversMay 13, 2023
@ben And, if it was scoped so well, what's the rationale for the six feet-fifteen minutes criteria? Sure, I can see how that was a reasonable approach in March 2020. But that was when we thought SARS-CoV-2 spread via ballistic droplets. Time marches on and now we know that is a mischaracterisation of transmission dynamics.
Show threadKate Nyhan is changing serversMay 13, 2023
@ben
"Close collaboration with domain experts" - I guess it all comes down to which experts
Morawska et al. COVID-19 and Airborne Transmission: Science Rejected, Lives Lost. Can Society Do Better? Clin Infect Dis. 2023 Feb 10:ciad068. doi: 10.1093/cid/ciad068. Epub ahead of print. PMID: 36763042. https://doi.org/10.1093/cid/ciad068
COVID-19 and Airborne Transmission: Science Rejected, Lives Lost. Can Society Do Better?

In April 2020, 36 experts alerted the WHO about the importance of airborne transmission of SARS-CoV-2. WHO dismissed this advice until it was published in an in

OUP Academic
Show threadBen AdidaMay 13, 2023

@kdnyhan these are all valid criticisms. Here's how I see it:

- we're not testing systematically anymore. So even though I agree the pandemic isn't over, our change in how we approach it makes this tool less useful. So sunset is the right and responsible move.

- agree that criteria for "contact" could have evolved. Still, the system was useful as is. I wonder if it might have thrown too many false positives if they changed it. Only so much Bluetooth can do here.

Show threadKate Nyhan is changing serversMay 13, 2023
@ben Exposure notifications are *more* useful, not less useful, in a world without routine testing. These days, a prompt to test yourself because of an exposure notification is the only pathway by which people with asymptomatic infections might discover them, and by which people might discover their presymptomatic infection while it's still presymptomatic.
Show threadKate Nyhan is changing serversMay 13, 2023
@ben
I get that you are pointing out: with less testing happening, fewer infections are discovered, and fewer exposure notifications are generated. But those exposure notifications that *are* generated can still be useful: they prompt the exposee to test and mask, potentially extinguishing a chain of transmission.
Show threadKate Nyhan is changing serversMay 13, 2023
@ben If we were measuring value based only on the *volume* of notifications, we would have turned it off in states where people opted in to the tool at a lower rate. Lower participation rate -> fewer exposure notices, just like less testing -> fewer exposure notices.
Presumably they left it on, even in low-participation states, because someone thought that it's better to be notified of *some* exposures than to be notified of *no* exposures - which is IMO right.
Show threadKate Nyhan is changing serversMay 13, 2023
@ben These days, the guidance for people exposed (no matter their vax status) is to mask and test - so I don't see how "false positives" are a problem. No one is being forced to quarantine bc of an exposure. So what exactly was the downside to providing continued access to an opt-in exposure notification system?
Show threadKate Nyhan is changing serversMay 13, 2023
@ben Maybe the tech side requires continued investment that isn't available; maybe the privacy risk is higher now than it was before - not being an expert, I couldn't argue with explanations of that nature.
Show threadKate Nyhan is changing serversMay 13, 2023
@ben But basing the decision to remove the tool on the idea that exposure notification is no longer useful, that, I don't get.
Show threadBen AdidaMay 13, 2023

@kdnyhan it's worth a discussion as to whether this system would continue to be useful at this phase... Maybe I'm seeing it wrong.

But again, this being the first time a tool like this was deployed, I think it's worth sticking to the deployment criteria they set out at the start.

Show threadJeri DanskyMay 14, 2023

@ben Logically, I totally see your point re: the lack of systematic testing and its impact on the notifications.

But emotionally, I just feel abandoned by my government (or even more abandoned than I felt before). I almost cried when the discontinuation notice popped up on my phone.

And yes, I still give a bunch of credit to Apple and Google for what they did.

Show threadBen AdidaMay 14, 2023
@jeridansky i get it. The way we're living with COVID is not at all how i expected things to go, and it's taught me a lot, good and bad, about humanity.
Show threadStephan SomogyiMay 13, 2023
Thanks @ben !
Show threadJamesMay 13, 2023
@ben Getting that sunset notification felt like an end of an era. Great product, thankful for what it did.
Show threadkbMay 13, 2023
@ben It is not done right in the slightest, particularly getting rid of it. It is rather disgusting and directly imacts the lives of those around me.
Show threadMarcel SalathéMay 13, 2023

@ben the “not as much as hoped” part was mostly due to bad integration of the tech into public health pipelines and low uptake rates. The evidence is quite strong. Now is the time to double down on further development, even though I agree with the current plan to discontinue V1.

See https://digitalepi.substack.com/p/more-evidence-that-digital-contact or https://www.digitalepibook.com/ch7.html and references therein.

More evidence that digital contact tracing worked

The UK is currently the only country in the world that has systematically collected and analyzed app data. A new study reveals just how effective digital contact tracing can be.

Digital Epidemiology
Show threadMarcus BointonMay 13, 2023
@ben I was involved with some protocol and code audits of the Apple and Google code, and generally we found that they were indeed responsible and responsive to our findings. Various national implementations of the protocols we looked at varied in quality though :)
Show threadGeorgeMay 13, 2023
@ben The main focus wasn't monetization, which was probably the reason. If only everything else was done like that
Show threadxianMay 13, 2023

@ben we had the chance to collaborate with google and apple folks on privacy concerns when I was volunteering with the pathcheck.org foundation — they were unwilling to touch location data with a ten-foot pole so the proximity model won out.

our original goals to support contact tracing without surveillance were mooted by the pandemic spread by then.

Show threadTiffany LiMay 13, 2023
@ben Yes! This project showed that we could have disease surveillance without giving up privacy. Even in a public health emergency. A great collaboration.
Show threadJennifer Stisa GranickMay 14, 2023
@ben Thanks for the heads up. Do you have a link to this announcement that I can share?
Show threadBen AdidaJun 22, 2023

@jennifer sorry i missed this!

https://www.engadget.com/most-states-halt-use-of-google-and-apples-covid-19-exposure-notification-system-152107338.html

I thought I'd seen a more official Google announcement but now can't find it.

Engadget is part of the Yahoo family of brands

Show threadJennifer Stisa GranickJul 2, 2023
@ben Thanks!