Skip to content
 

Do these data suggest that UPS, Amazon, etc., should be quarantining packages?

Doug Davidson writes:

I just wanted to draw your attention to this paper [Aerosol and surface stability of HCoV-19 (SARS-CoV-2) compared to SARS-CoV-1, by Neeltje van Doremalen et al.] that used Stan.

They are concerned with how long the virus remains viable on different surfaces, including packaging material. I think this will become more important as time goes on, as people rely (even) more on delivery than they do now.

I guess one question is whether these estimates are credible. If they are, delivery services (Amazon, UPS, etc.) could change policy so that packages are routinely quarantined for a period corresponding to the upper bound of the estimates. I don’t know enough about decision theory, or all of the factors involved to even pose this question in cost-benefit terms, but maybe some of your commenters do!

The above-linked article states that “the virus can remain viable and infectious in aerosols for multiple hours and on surfaces up to days.”

I’ve not tried to evaluate this work but I’m sharing it here. There are two questions here: first, the relevance of these data to real-world disease transmission; second, the propagation of these estimates from virus counts to exposure probabilities to effects on the transmission network, to get a sense of the potential benefits of quarantining packages. Or maybe they could just wipe everything down? I dunno.

56 Comments

  1. MichiganWater says:

    Or have the packages sit in a room/run through a tunnel with sufficient UV light?

      • From the article:

        but were converted to be non-infectious after 90-, 60- and 30-min exposure at 56 degrees C, at 67 degrees C and at 75 degrees C, respectively. Irradiation of UV for 60 min on the virus in culture medium resulted in the destruction of viral infectivity at an undetectable level.The survival ability of SARS coronavirus in human specimens and in environments seems to be relatively strong. Heating and UV irradiation can efficiently eliminate the viral infectivity

        I’m not sure 60 minutes of UV irradiation is a reachable goal for every package going through their pipelines, and those sorts of durations for heating seems unreasonable too.

        All of which leads me to believe that it’s probably just safest and simplest to wet down the exterior of the package with a soap solution upon receipt. Would be useful if someone actually tested soap concentration / type / and time on various surfaces.

      • What is the likely probability if a carrier handles paper money and is transmitted to another party picking up the virus if done within hours?

    • Anoneuoid says:

      Came here to say this. UV is great for this type of thing. Just put it over a conveyor belt at some point.

      • UV equipment Costs a bunch, if done wrong could provide a false sense of security (ie. too short duration, not the right wavelength, bulb burns out, etc) hazardous to workers, only works for the exterior of the package (but the plastic wrap inside is apparently the longer lived material right?)

        So, I agree UV sterilization of packages could be a good idea in theory, but I’m not at all confident that it’ll provide a strong practical benefit. If it can be done cheaply, I think great.

        After seeing these kinds of results publicized in the news, now when I receive packages, I’m wetting them down with a solution of ~ 1tsp liquid dish soap in ~ 1L water (in a spray bottle), and letting them sit for an hour before I open them. I’m guessing this is going to dramatically reduce viral load given that it’s a lipid enveloped virus. I’m also leaving the box outside and just bringing in the contents, which I wipe down with a cloth using the same solution… a little over the top I think personally, but it has about zero cost, and potentially significant benefit, so it’s the right thing to do.

        I also plan to wash down my produce with a similar solution, using a dye and scent free soap.

        • Anoneuoid says:

          UV equipment Costs a bunch

          I rigged up my own UV thing for a mouthguard I couldn’t get the smell out of for under $20. It worked great.

          Obviously at industrial scale it will be more than that but so is the budget.

      • Thanatos Savehn says:

        UV is increasingly used in the food processing industry. It’s essentially a quality assurance measure whereby you fiddle with the Neyman Pearson controls until you get to your desired level of risk and then insure around the prospect of the occasional dead Granny who caught listeriosis from your veggies (not saying I like it but that’s the way the world works). The problem with something like SARS-CoV-2 is that all it takes is one contaminated package (aka fomite) to restart the conflagration all over again. Since perfection isn’t possible in this world we’re not likely to see this problem go away until either SARS-CoV-2 has swept the world or an effective vaccine against it has been developed and deployed.

        • Phil says:

          A fluorescent lamp without the phosphor coating emits quite a bit of UV, enough to deactivate bacteria in a 4-inch-deep stream of water in a few seconds. if the lamp is hanging over the water. UV systems for pools and hot tubs and such are usually a lot more expensive because the lamp is immersed in the water and needs a quartz window rather than glass to avoid blocking the UV, but systems suspended in air can be very inexpensive indeed. If it takes a long time to work (for coronavirus) then it’s the time and maybe space that’s the problem, not the UV source.

    • Heck I’ve been spraying the packages with Sevent Generation Disinfectant Spray. LOL

  2. jim says:

    The question is how much infection does virus from surfaces cause?

    Someone please correct me if I’m wrong, but my understanding is that the primary transmission mode by is through air, and that the effectiveness of transmission through the air increases dramatically as the concentration of virus increases and the length of the exposure time increases. This is why health care workers are so heavily impacted, right? They’re continuously bathed in a high concentration of airborne virus produced by very sick people breathing.

    OK, now you think about the virus count in a square foot on a surface – from a place where there isn’t a large concentration of sick people – vs in a cubic foot in an air volume; and a single exposure (touching) which still doesn’t have a direct path to infection, vs continuous exposure (breathing), coupled with the fact that on a surface the virus starts to disintegrate almost immediately; I don’t really think there’s much to be concerned about. I mean if someone can show that surface-borne virus can generate some reasonable chance of infection, OK, fine.

    But even then packages themselves – what can actually be exposed to UV – might be a relatively small problem. The big problem might be what’s likely plastic inside the package, which can’t be exposed to UV during shipping and which is probably plastic which helps the virus survives longer.

    I’d be interested to hear people’s thoughts on that, maybe I’m wildly wrong but that’s what seems sensible from what I know.

    • I think the goal of this paper is to help address the question of whether surfaces could be a reasonable route of infection. It provides some mechanistic measures of survival times.

      >coupled with the fact that on a surface the virus starts to disintegrate almost immediately;

      that’s exactly what they were trying to test, and NO it doesn’t. for several hours to days the virus is viable… the test they use is to infect cells in tissue culture with virus washed off the surface of the material, and see how many of the cells die. So it’s not just detecting the existence of the virus, but some kind of actual viability.

      My take on all this stuff is that *yes* it’s *possible* that fomites (objects with viruses on them) could be a problem with this virus, as it’s relatively stable for hours to days. It suggests that in a healthcare setting, daily washing of the floors, light switches, desk tops, keyboards, etc is probably a good idea.

      If it’s easy enough for you to wash down the outside of your packages, and wipe down the products you remove from them, given the environment where you receive them… do it. If that’s really hard, then at least do a good job of washing your hands after handling and unpacking packages.

      • Clyde Schechter says:

        In addition to the above, I am wearing disposable gloves when I handle the packages and unpack them. Maybe we’re both over the top here, but I don’t think so. Paranoia used to be a thought disorder, but now it’s a key to survival.

      • jim says:

        “No viable virus could be measured … after 24 hours on cardboard for HCoV-19…”

        The upper row of charts shows that on cardboard (shipping packaging) concentration decreases by 10^-1 to 10^-2 in four hours. I’d say that’s pretty fast. So, yeah, if the FedEx guy is coughing directly on your package when he drops it on the doorstep, then you have a serious concern.

        The caveat is the temperature conditions of the study: since many viruses prefer cold conditions, the half-life might be longer for packages being transferred in cold vehicles.

        • jim says:

          Anyone have any ideas or reference on the details of the experiment?

          but beyond that:

          What’s the concentration of virus in the breath of a highly contagious person? If a very contagious person blows out a breath of air, say 500 ml, say 1 ft above a surface, how much of the virus will settle to the surface, and how long will it take?

    • dzhaughn@yahoo.com says:

      For viruses in general, I will trade my face mask for your hand-sanitizer any day.

      https://www.chicagotribune.com/opinion/commentary/ct-five-myths-about-colds-20151125-story.html

      Healthcare workers may contract it mainly through the air, but that’s conditional on their keeping hands clean.

      • jim says:

        I read today that the primary mode of transmission of covid-19 is close-quarter coughing, like within 6 feet. This implies that people get it on their hands and pass it by touching each other.

        So I’ll grant you the need for hand sanitizer

        • I really don’t think this is well understood yet, but it does seem very clear that MANY infections come from asymptomatic people, *not* from coughing. so it’s either regular breathing or hand-hand or hand-object-hand.

          • Rahul says:

            Daniel:

            What’s your source for this?

            Seems opposite of what I seem to recall in some quite authoritative guidelines.

          • jim says:

            I said “I read” because it was easier, but the information above is from an NBC interview with Fauci. I lost the link. He said specifically that there is “some chance” of aerosol transmission but downplayed it a lot.

            But now that you mention it I remember the reason I thought it was aerosol in the first place: I read about a study in Germany where they took throat cultures from asymptomatic children and found very high virus counts, so they believed it could be transmitted aerosol asymptomatically.

            Maybe I saw that on MR? Don’t recall.

            I agree though it does seem to be moving way to fast to be only by direct one-to-one transfer, seems like it has to be one-to-many.

            • Anoneuoid says:

              I agree though it does seem to be moving way to fast to be only by direct one-to-one transfer, seems like it has to be one-to-many.

              Or it was already spread months ago. On the other hands ILI has been on the rise the last few weeks: https://www.cdc.gov/flu/weekly/index.htm#ILINet

              In a note at the top of the page the CDC suggests it is due to:

              The COVID-19 outbreak unfolding in the United States may affect healthcare seeking behavior which in turn would impact data from ILINet.

              • Your own data you posted recently shows on the order of 10-15% of tests are positive. It’s still a relatively uncommon disease even among people with symptoms. It DIDN’T spread widely months ago. just get that out of your head.

              • Anoneuoid says:

                Your own data you posted recently shows on the order of 10-15% of tests are positive. It’s still a relatively uncommon disease even among people with symptoms. It DIDN’T spread widely months ago. just get that out of your head.

                Remember these are not antibody tests, they only test for current infection. Anyone who already recovered will test negative. So 10-15% would be huge if it was a random sample, but instead that is only 10-15% of “suspected cases”.

              • 🙄

                use your model to make some predictions and see if they match the data… in particular why are hospital admissions following an exponential curve with constant coefficient?

              • Anoneuoid says:

                use your model to make some predictions and see if they match the data…

                Of course. Where is the hospital admission data?

              • https://ourworldindata.org/coronavirus

                in particular: https://ourworldindata.org/coronavirus-source-data

                you can use confirmed cases and deaths instead.

                Here is data on testing:

                https://ourworldindata.org/covid-testing

                Please go ahead and make a model which includes the idea that more than say 20% of the world or even europe+north american population has had this virus already and show how your model explains the testing and detected cases.

                There is nothing that is going to get you exponential growth except an immunologically naive population and a small prevalence.

            • Anoneuoid says:

              Please go ahead and make a model which includes the idea that more than say 20% of the world or even europe+north american population

              I couldn’t find hospitalization data at that link… To be clear, I don’t know what measures were taken in other countries.

              But in the US, I do not see why it just started to cause a hysteria and exponential increase in cases within week of when they started rolling out testing after months of delays with almost no effort at screening international passengers, etc. And now the number of cases has been pretty close to a constant proportion (fluctuating between 10-17%) of the number of people tested (according to the covidtracking.com data).

              However, the linear relationship between tests and cases shows up again in this chart. Almost all countries are between 5-20% positive tests:
              https://ourworldindata.org/grapher/tests-vs-confirmed-cases-covid-19

              There could be some combination of false positives/negatives, selective testing, etc common around the world that results in that factor of 2 around ~10% rate, or it really is that ~10% of people have this virus already.

              Also, I happened to come across this:

              Interestingly, nearly one in ten (9%) people went as far to admit they believe they have coronavirus right now.

              https://www.studyfinds.org/national-nightmare-1-in-5-americans-expect-theyll-be-diagnosed-with-coronavirus/

              I’d love to see a survey like that actually reported correctly instead of as clickbait, but still interesting that the numbers match up.

              Please don’t think I have some dogmatic belief in this explanation, but I don’t think it can be dismissed out of hand either.

              • Anoneuoid says:

                Almost all countries are between 5-20% positive tests:

                Looking again 1-20% would be a better range from that chart. Also, when I looked at the raw data that seems to be comparing # tests/cases at different timepoints for each country… so it is very messy data.

                I also see “hospitalized” just got added to this data, which seems to be the cleanest so far: https://covidtracking.com/api/us/daily

              • Testing is predicated on having a credible case for having it. I don’t think that 10% holds at all for Korea or Singapore etc where there is much more testing. It makes perfect sense that 10% of people that doctors think might have it do have it… It doesn’t make sense that 10% of everyone has it right now. The thing you’re looking at is basically doctors are about 10% good at guessing who has it

              • Anoneuoid says:

                Testing is predicated on having a credible case for having it.

                Yes, and confirmation also depends on false positive/negative rates of the test. Deborah Brix said a few days ago (Mar 17) in one of the press conferences that we are going to learn in the next few months that the tests being used in other countries were not very good: https://www.youtube.com/watch?v=HbhnBEiqGGk&t=1h17m16s

                I don’t think that 10% holds at all for Korea or Singapore etc where there is much more testing.

                South Korea was 8,652/316,664 = 2.73% but there was no testing data for Singapore from that source.

                It makes perfect sense that 10% of people that doctors think might have it do have it… It doesn’t make sense that 10% of everyone has it right now.

                That makes sense, but it doesn’t explain:

                why it just started to cause a hysteria and exponential increase in cases within week of when they started rolling out testing after months of delays with almost no effort at screening international passenger

                I think everyone agrees that some percent of the increase in cases is due to the testing. The question is how much.

              • >I think everyone agrees that some percent of the increase in cases is due to the testing. The question is how much.

                Sure, just write a differential equation for identified cases… first actual cases is growing exponentially with a logistic slowdown as it reaches the total population

                dN/dt = c*N*(1-N/Pop)

                then tests each day are let’s say 10% of the absolute increase in N, with saturation at a testing capacity:

                dT/dt = (1-exp(-dN/dt * scale))*Cap;

                where solve for scale to get dT/dt = .1*dN/dt at t=0

                Then the probability P of a test being positive increases with the case load N, some kind of logistic function perhaps

                dPos/dt = P(N)*dT/dt

                Then integrate the whole thing in time using a numerical integrator, and see how it looks…

          • Martha (Smith) says:

            My impression from what I’ve heard on NPR is that a lot of the spread is hand-surface-hand, which is why using hand sanitizer and wiping surfaces with disinfectant are being promoted.

            Looking it up just now, I found https://www.nih.gov/news-events/news-releases/new-coronavirus-stable-hours-surfaces , which “suggests that people may acquire the virus through the air and after touching contaminated objects”.

            • jim says:

              In the interview I saw Fauci downplayed the idea of non-human surface transmission. He emphasized the “6-feet” rule – that’s about the distance droplets can travel from coughing, and hand-shaking and person-to-person touching.

    • I spray all components of a package with Seventh Generation Disinfectant Spray and don’t handle it further for ten minutes b/c it takes at least 10 minutes to kill the viruses and bacteria. I use eco-friendly products for the most part. They require more time to work.

  3. Doug Davidson says:

    From what I have read, the CDC FAQ mainly refers to the earlier concern about packages shipped from China:

    https://www.cdc.gov/coronavirus/2019-ncov/faq.html

    “There is still a lot that is unknown about the newly emerged COVID-19 and how it spreads. Two other coronaviruses have emerged previously to cause severe illness in people (MERS-CoV and SARS-CoV). The virus that causes COVID-19 is more genetically related to SARS-CoV than MERS-CoV, but both are betacoronaviruses with their origins in bats. While we don’t know for sure that this virus will behave the same way as SARS-CoV and MERS-CoV, we can use the information gained from both of these earlier coronaviruses to guide us. In general, because of poor survivability of these coronaviruses on surfaces, there is likely very low risk of spread from products or packaging that are shipped over a period of days or weeks at ambient temperatures. Coronaviruses are generally thought to be spread most often by respiratory droplets. Currently there is no evidence to support transmission of COVID-19 associated with imported goods and there have not been any cases of COVID-19 in the United States associated with imported goods. Information will be provided on the Coronavirus Disease 2019 (COVID-19) website as it becomes available.”

    However, if community transmission is occurring, some fraction of packages could be handled by infected individuals within the US.

    If the estimates are realistic for conditions like those in a shipping warehouse, it seems reasonable to me to update the CDC FAQ – if nothing else, leaving non-essential packages alone for a day or two doesn’t seem like a bad idea.

  4. Adede says:

    Just wash your hands after you open a package from Amazon. Problem solved.

  5. Nick Adams says:

    Coughing expels droplets of varying size. Large ones fall immediately onto surfaces, smaller ones float in the air for some time. Most viruses can’t survive in small droplets for any length of time so human transmission is via touching infected surfaces. Coronaviruses can survive in small droplets because it has a thicker than normal envelope, so they could potentially be transmitted by both touching infected surfaces and by inhalation of infectious airborne droplets.
    To cause infection viruses have to enter a human cell and to do so they use the pre-existing receptors on the cell surface. Because they are very simple and compact organisms (with very little functional redundancy) each virus usually targets a single receptor type. The receptor type they target determines what kind of disease they cause e.g. hepatitis A uses receptors found on liver cells.
    Beta-coronaviruses such as SARS and covid-19 target ACE-2 receptors which are found in the alveoli of the lung but not the upper respiratory tract. This suggests that inhalation of small airborne droplets which can reach the alveoli are more important for infection transmission than large droplets on surfaces (which are transmitted by touch to the mouth and nose).
    I’m not saying that you shouldn’t wash your hands or clean your UPS packages if you want to, but I suspect that these precautions are likely ineffective.

  6. steven t johnson says:

    Delivery services that screen packagers are as safe as the screening is effective. Gloves and sneeze guards over packing tables should be a relatively inexpensive backup. Hand washing immediately after opening the package is yet another safeguard. This assumes that freshly opened packaging materials are not apt to carry any C19. This doesn’t seem to be much of a problem. Of course, an unlikely series of events that lead to spread is inevitable in an enormous number of trials.

    That said, it seems to me the most likely problem in materials transmitting C19 is money. I mean the physical money exchanged between two people in commercial transactions, like paying for a drive-up meal and maybe getting change. The phrase “money laundering” should take on a new meaning?

Leave a Reply to Dzhaughn