Catch up on stories from the past week (and beyond) at the Slashdot story archive

 


Forgot your password?
Close
typodupeerror
×
Science

Coronavirus Apps Don't Need 60% Adoption To Be Effective (technologyreview.com) 54

Posted by msmash from the upon-further-inspection dept.
With dozens of digital contact tracing apps already rolled out worldwide, and many more on the way, how many people need to use them for the system to work? One number has come up over and over again: 60%. From a report: That's the percentage of the population that many public health authorities documented by MIT Technology Review's Covid Tracing Tracker say they are targeting as they attempt to protect their communities from covid-19. The number is taken from an Oxford University study released in April. But since no nation has reached such levels, many have criticized "exposure notification" technologies as essentially worthless. But the researchers who produced the original study say their work has been profoundly misunderstood, and that in fact much lower levels of app adoption could still be vitally important for tackling covid-19.

"There's been a lot of misreporting around efficacy and uptake ... suggesting that the app only works at 60% -- which is not the case," says Andrea Stewart, a spokeswoman for the Oxford team. In fact, she says, "it starts to have a protective effect" at "much lower levels." The Oxford models found that "the app has an effect at all levels of uptake" as illustrated by this graph which shows every level of adoption slowing to pandemic to some extent. Because of the way such digital contacting tracing and exposure notification apps work -- by notifying users if their phone has been in proximity to the phone of somebody who later gets a diagnosis of covid-19 -- blanket coverage is preferable. The greater the number of users, the higher the likelihood that it will help at-risk people to self-quarantine before they can infect others.
This discussion has been archived. No new comments can be posted.

Coronavirus Apps Don't Need 60% Adoption To Be Effective More

Coronavirus Apps Don't Need 60% Adoption To Be Effective

Comments Filter:

  • "the app has an effect at all levels of uptake"

    Whatever uptake occurs, fully that many people have accepted the operant conditioning training them to accept constant surveillance.

    It's a tautology, really.

    • Re:Always effective, yes (Score:4, Funny)

      by DontBeAMoran ( 4843879 ) on Monday June 08, 2020 @12:13PM (#60160094)

      I'm not installing this, otherwise someone will come knocking at my door every week to know if I'm still alive.

      I haven't left my house in three months. My food supply may be running low, but the constant junk mail will keep me supplied in toilet paper forever and I tightly roll the surplus to make logs for my wood stove in preparation for the coming winter. A big thank you to companies and stores still stuck in the 1980's for their ads.

    • As if you haven't been trained to already accept constant surveillance.

      You carry your cell phone with you everywhere, don't you? Police have been using that to track people since at least the 90s. You carry a tracker with you everywhere, and yet you complain about installing an app that literally does not track you.

      • Because it's bad, worse is fine.

        Got it.

        • installing an app that literally does not track you.

          Until the next round, when the "value proposition" is changed to "Well, now that you've installed it, it just makes sense for us to upload everything to our servers, right? More efficient and effective that way. I mean, that's so obvious, we'll just go ahead and push you the patch..."

          -That- is the main argument against slippery-slope operant conditioning (which is not a process of rational consent, it's training an animal's subconscious, but that's a whole other debate), which you seem to have missed.

          • Until the next round, when the "value proposition" is changed to "Well, now that you've installed it, it just makes sense for us to upload everything to our servers, right?

            Write your own app. The protocol is open.

            You're constantly being tracked by your cell phone and you don't care. You're just angry for no reason.

            • I am not obligated to write an app, or do anything. That'd be accepting even more parasitism from you and your corporate masters.

              You know my cell phone habits? Impressive psychic powers, but it doesn't add to your argument.

              • You know my cell phone habits? Impressive psychic powers, but it doesn't add to your argument.

                You didn't understand my argument. My argument is that you are a fool and a hypocrite, and you don't understand technology. Your posts are supporting all of that.

                • I understand technology fine, at least enough for a major Fintech firm to call me a "Senior Software Engineer".

                  But then, my company actually creates value, rather than privacy parasiting and rent-seeking. You work for Google or equivalent, I assume?

                  • I understand technology fine, at least enough for a major Fintech firm to call me a "Senior Software Engineer".

                    Then why don't you understand how technology works? Why are you happy to be carry a phone that tracks you, and not install an app that doesn't track you?

                    Also, your entire argument is based on a logical fallacy, the slippery slope.

                    • For the same reason I'd prefer a tracker blocker to block all of them, rather than a few. And I'm not looking for it to block less.

                      Again, your basic position is overtly irrational, and every unethical person's excuse: "It's bad, so worse is fine."

                      And you don't understand the slippery slope fallacy. It is only a fallacy if the outcome suggested is in fact not likely. With your logic, it is not only likely, it is guaranteed.

                    • I'm sorry, you've missed my point. My point is "this is not bad."

                    • Which would work, if it were true. So you block no trackers in your browser?

                    • So you block no trackers in your browser?

                      Of course I block trackers in my browser. It's immoral and practically foolish to not block ads.

  • I'm not sure why people want to identify some effectiveness "threshhold" for various mitigation strategies, but there seems to be a strong tendency to want do to it. The fact is that every mitigation strategy, from handwashing to quarantine to herd immunity to exposure notification (and subsequent self-isolation) has some effect at reducing the transmission rates. Some strategies, such as harshly-enforced total lockdown or high levels of herd immunity, can be enough by themselves to push the replication rate below one and cause the virus to fade away. But everything you do that reduces the ability of the virus to spread has some effect, causes some reduction in the replication rate.

    The goal of exposure notification apps is to enable more people to find out faster that they have been exposed, so they can self-isolate before they become contagious. If only one person gets a notification and quarantines themselves, that will have reduced the spread. Not enough to be noticeable perhaps, but some. If 5% of people use the app and a few infections every week are prevented, that will help. If you can combine that with other forms of contact tracing, plus mask wearing and lots of handwashing... that could be enough to reduce R0 to less than one, or maybe some other bits are needed as well.

    Herd immunity also reduces the reproduction rate, even when relatively small percentages of the population are immune. At low levels of herd immunity, it's not enough, on its own, to prevent the need for other mechanisms, but it helps. The seasonal flu is a good example; we have some level of inbuilt herd immunity because most all of us have had previous, similar flus (which doesn't make us fully immune but does give us some resistance), and we have a level of artificial herd immunity from vaccination. Neither of these are enough to keep the replication rate below one, but they do help keep it low enough that much of the population is unscathed every year.

    The more of the mild, easy and painless forms of COVID-19 reproduction reduction we can implement, the less we'll have to rely on the economically-destructive ones.

    • Re: (Score:2)

      by ljw1004 ( 764174 )

      I'm not sure why people want to identify some effectiveness "threshhold" for various mitigation strategies, but there seems to be a strong tendency to want do to it.

      I've observed this too. It comes in silly logic "If 6ft is good enough then why do we need masks? if masks are good enough then why do we need 6ft? therefore the guidelines about 6ft and masks are both incorrect." And it comes, especially in the UK, in the form of zealotry/officiousness "You're not doing this to my particular standard and therefore I will publicly shame you."

    • Re: (Score:2)

      by Slayer ( 6656 )

      If only one person gets a notification and quarantines themselves, that will have reduced the spread. Not enough to be noticeable perhaps, but some.

      The missing keyword here is "opportunity cost". Yes, even a single positively identified infected person helps the effort, but these contact tracing apps didn't fall off a tree. Every single country had to develop their own version because of $REASONS, so lots and lots of money was spent on something with barely any effect.

      Herd immunity also reduces the reproduction rate, even when relatively small percentages of the population are immune. At low levels of herd immunity, it's not enough, on its own, to prevent the need for other mechanisms, but it helps.

      There are several countries, which have by now reached these small to moderate levels of herd immunity: USA, Brazil, Russia, UK, Sweden. Not one of these countries has seen any benefit fr

      • The cost of developing such an application should be a lot less than other measures like total lockdown or building temporary hospitals. Even when you factor in "I gave the contract to my buddy".

        As for how things are going in my location (Texas) the preferred mitigation method seems to be "none of the above" - I went to the grocery store to cash my stimulus check today and the proportion of people wearing masks was definitely under 20%, social distancing not being observed either - I was only in there for 5

      • Herd immunity also reduces the reproduction rate, even when relatively small percentages of the population are immune. At low levels of herd immunity, it's not enough, on its own, to prevent the need for other mechanisms, but it helps.

        There are several countries, which have by now reached these small to moderate levels of herd immunity: USA, Brazil, Russia, UK, Sweden. Not one of these countries has seen any benefit from this herd immunity.

        Nonsense. Herd immunity, at any level, decreases R0. Perhaps it only moves it from 2.5 to 2.4, but it still moves it. Note that I'm not endorsing a strategy of "let 'er rip and we'll rely on herd immunity to wipe out the virus". That strategy is completely stupid, both in theory and as proven out in practice. But it's still the case that even at low levels of herd immunity, it does have an effect. This is utterly obvious: If an infected, contagious person interacts closely with 10 people and none of t

  • This article is unmitigated nonsense (Score:3, Interesting)

    by Slayer ( 6656 ) on Monday June 08, 2020 @12:59PM (#60160342)

    The premise of tracing apps is, that if close to everybody uses them, then a single detected infection can be used to track down possible other infections with much better aim than random testing. Since people quickly observed, that even in the presence of strong pressure only a fraction of people download and deploy these tracing apps, the main argument for using such tracing apps quickly evaporated. This article desperately tries to resurrect it.

    However, the argument it makes is not supported by the data shown. The curves in the figure [technologyreview.com] show little benefit for 14% adoption rate, and only moderate benefits with 28% adoption rate. Note, that people in Singapore with lots of social pressure had adoption rates of about 25%. I have no idea, where the author of this article got the idea, that even 1% adoption rate makes a tangible difference.

    The main problems of these tracing apps go far beyond lack of adoption or potential threats to your privacy:

    • - These apps don't know, whether you are inside or outside, or which direction the wind comes from. The app can only be either very conservative, or very cautious, generating either lots of false positives or false negatives.
    • - These apps don't account for masks being worn or not, or whether the interaction between two people was close or not. Again, lots of false positives or negatives
    • - Once a person has accumulated a small amount of false alarms, this person will likely get rid of the app, despite being much higher at risk than people with no alarms.
    • - The app does not account for infections passed on through surfaces or other means.

    I consider this article a last ditch effort to justify all these development efforts being made in vain, and it fails even at that.

    • only a fraction of people download and deploy these tracing apps, the main argument for using such tracing apps quickly evaporated.

      If even only 10% download and use it, that is still a massive benefit to preventing further spread.

      The benefit of enhanced and instant tracing abilities is so large, that even a small use of them provides disproportionately large benefits.

      What is the downside? It's not like there is real cost to deploying and trying to use this tech. Since even a 1% improvement would provide

      • Re: (Score:2)

        by Slayer ( 6656 )

        With a total Australian population of about 26 million people [wikipedia.org], about 6 million [theguardian.com] people downloaded the app. Yet, only one single potentially infected person was identified three weeks later!

        The main problem with tracing apps is, that human contact tracers are just way more effective and precise, because they do know how to account for all these factors I listed above. With a small number of deployed user base, there will either be no benefit at all, or false positives from these contact tracing apps start flo

        • With a total Australian population of about 26 million people [wikipedia.org], about 6 million [theguardian.com] people downloaded the app. Yet, only one single potentially infected person was identified three weeks later!

          That's good, Australia has used a lot of different prevention techniques, including this one, and now they are only seeing 5-11 new cases of Coronavirus per day.

    • These apps don't know, whether you are inside or outside, or which direction the wind comes from. The app can only be either very conservative, or very cautious, generating either lots of false positives or false negatives.

      Except presumably the contact has a GPS position associated with the timecode. So if it's on the street you can use your brain memory to say "Oh that was on the street when I was walking from the restaurant home on the sidewalk. I didn't duck into the store there."

      The curves in the figure show only moderate benefits with 28% adoption rate

      *Looks at curve*. I would call a 33% drop in cases pretty substantial.

  • Pipe dream (Score:5, Insightful)

    by oldgraybeard ( 2939809 ) on Monday June 08, 2020 @01:08PM (#60160376)
    State surveillance is the governments biggest wet dream. We must do it, to combat coronavirus wink wink and for the children.

    Just my 2 cents ;)

    • The biggest thing getting lost in these discussions is that we're already wink-wink-doing-it, not for the children, but so we can check Facebook while sitting at the drive-thru and get on Slashdot while we're taking a shit. Not only do you provide FB/Google/Apple/Samsung/Huawei/Microsoft (not an exclusive list, check all that apply) with everything a contract tracing app might keep, you give them much more than that - anything that's on your phone, in your email, social media accounts, on Slashdot - that is

    • State surveillance is the governments biggest wet dream. We must do it, to combat coronavirus wink wink and for the children.

      Assuming they use the Google/Apple approach, governments will get zero data about citizen location history or interactions. Some governments are unhappy about this, but Google and Apple aren't budging. Note that Google and Apple don't get any surveillance data from it either (though they obviously already have lots).

Slashdot Top Deals

C'est magnifique, mais ce n'est pas l'Informatique. -- Bosquet [on seeing the IBM 4341]

Close