Diffusion of new technologies to reduce the spread of Covid-19
Coronavirus Apps Show Promise but Prove a Tough Sell
This article from The New York Times discusses the launch of a Covid-19 tracking app by Apple and Google. This app tracks where people go and alerts them of exposures to Covid-19 through exposure notifications on a user’s phone. The software is designed to prevent outbreaks by alerting people if they have been exposed so that they can limit their contact with others. However, it is difficult to get people to adopt and use this app due to privacy concerns that many people have surrounding location tracking and disclosure of health information, despite the fact that the app was designed with privacy protection in mind in hopes it would encourage more people to adopt it. This contact tracing app is effective at reducing the spread of Covid-19; the article explains that epidemiologists suggested that if 60 percent of people in a community use an app for digital contract-tracing, the pandemic could be brought under control because the spread would be limited. Epidemiologists also found this would be the case if even just 15% of people used contact-tracing apps in a community. For example, in Switzerland, the use of such an app is more widespread and more easily adopted within communities, and for every 100 people who test positive, the app correctly notifies 24 others who have been in contact. This has proved as an effective means of contact tracing. Additionally, the University of Arizona had widespread adoption of the app, and they believe that outbreaks on campus were reduced with its help. The New York Times points out that the app was likely widely adopted on campus because the population there is mostly young people who are tech-savvy and will trust an app that their university encourages them to get. However, many students on campus did not share their code with the app once they did test positive, which reduced the effectiveness of digital contact tracing.
This article relates to two concepts from class: diffusion and the spread of epidemics. Diffusion of innovations defines how we imitate the decisions of local neighbors rather than the general world around us. This explains why a new idea or technology gains traction among certain groups and how it spreads to other individuals. Diffusion of innovations introduces the concept of a threshold, or the number of local neighbors required for a node to adopt a new idea or technology; the concept of strong and weak ties can also be used to describe this model, because new information comes from weak ties, but new innovation comes from strong ties. That is, for someone to adopt a new idea or technology, a strong tie must have that idea or technology. Furthermore, diffusion can be understood by clusters, groups of nodes that help identify where innovation or adoption stops. If a cluster has a density of one minus the threshold, and no one within that cluster has adopted the new technology, then adoption will not enter that cluster.
Diffusion of innovations is evident in this article in how the use of the contact tracing app spreads across communities. For example, in privacy-conscious communities, the app’s adoption was not popular. This cluster of strong ties that cares about their privacy to the point where they don’t want to use the app must have a high threshold for requiring the adoption of the app, because they have not adopted it. However on college campuses where people are younger and more trusting of technology, they are more likely to adopt the new technology because many other people within their cluster are adopting it as well. In communities where some people are more privacy-conscious and others are more trusting of technology, both groups can make their own decisions about whether to adopt the technology or not. That is, the privacy-conscious group will not adopt and the technology-trusting group will adopt. The adoption of the app will not spread across groups if the threshold for adoption is not met. Therefore the use of the app, or lack of use of the app, stays within clusters.
This has notable implications for society because if people’s social networks are aligned with the clusters in the aforementioned example, then adoption will only exist within certain social groups. Therefore, the groups that use the contact-tracing app will be more informed of possible exposures and be able to distance themselves if they have been exposed, thereby keeping the entire group healthier. Furthermore, the concept of diffusion applied to a new technology like a Covid-19 contact-tracing app offers insights into what kind of communities are more likely to adopt new technologies and how thresholds can be harnessed to get more people to adopt. For example, if clusters that have enough people above the threshold that will use the technology and do not have privacy concerns over location tracking, other people in their cluster (even if they are initially against using the technology) will adopt the app because their number of neighbors using the technology is above the threshold for adoption.
As for the spread of epidemics, they can be understood through the SIR model and the branching model (which is an instance of the SIR model). The branching model starts with an individual or node at the top and branches out into more nodes at each level. The node at the top is known as patient zero and they bring the infection into the network. Each branch off of each node is for k other people that interact with patient zero. This pattern repeats in a branching shape where the number of branches from each node is based on k, the number of people they interact with. On each branch, there is a probability of contagion, p, that predicts if the infection will pass on to the next node. The spread of the infection is based on the product of p and k. That is, the probability of infection and the number of people who have possibly been infected determine the spread of the infection. A high p and high k make it likely that the infection will continue to spread; if an infected person meets with many people and the probability that they get sick is high, the infection will continue to spread. The SIR model reflects social networks more accurately and how they close in on themselves. In the SIR model, nodes can be susceptible, infectious, or removed (if they have recovered and are no longer infected or contagious). In this model, an infected node can infect its neighbor with probability p. Eventually an infected node will recover, and then it cannot get infected again or spread the infection. Therefore when all nodes have recovered or when the infected node has no neighbors that can be infected, the infection stops running through the social network. The notable issues with this model are that recovery and immunity to infection can be temporary, the infection can mutate and therefore infect people again, and social/contact networks change with time making it impossible to determine when the network closes on itself and stops the spread.
The spread of epidemics related to this article is that the contact-tracing app seeks to reduce both p and k by informing app users of potential exposures. If someone is notified that they are infected, they can share that with others so that they know they should limit their contact with others. That way if one of the users who has been contact-traced is positive, they will have reduced their k (number of people they interact with) and therefore the overall spread of the infection. Furthermore, if someone is alerted that they have been in contact with someone who tested positive for Covid-19, they can take precautions like wearing a mask when they see that person, washing their hands after they’ve seen them, and also wiping down any items that may have been handled by the person they were exposed to. These actions work to reduce p, the probability of infection. Ultimately, the app explained in this article works to combat the spread of Covid-19 by reducing k and alerting others if they have been exposed so that they can limit their interactions with others. By reducing k, the spread of Covid-19 is reduced (because the spread is a product of p and k). Additionally, the app reflects a branching model where each person who is exposed or infected has branches going out of them for all the people they interacted with. If someone has interacted or been infected by multiple people, it could be thought of as a SIR model since the network will eventually close in on itself. This app works to draw out these models so that users know if they have been infected and can take proper precautions to reduce the spread of Covid-19.
The implications for the spread of epidemics, the contact tracing app, and Covid-19 are significant. For one, it demonstrates how people need to limit their social bubbles in order to reduce the spread. That is, if one person breaks out of a social bubble, they are introducing more people who they could infect or become infected by. Where a SIR model eventually closes in on itself, if someone interacts with someone new, they are expanding the networks and expanding the people who could be infected, thereby expanding the number of people the infection has to go through before closing in. The spread of Covid-19 stops when the network closes in on itself, so it is important to remind people that every interaction they have has an implication for the greater spread, and eventually die-out, of Covid-19. Furthermore, people have to be willing to share information for contact-tracing to be effective. In order to reduce k, people must put the code into the contact tracing app so others can be alerted of exposure and limit their contact with others. This is why contact tracing and honesty during the times of Covid-19 is so important. Therefore, it might be better to remove the autonomy of the decision as to whether someone gets to disclose their positive test result (or exposure to a positive person) or not. If everyone is alerted of potential exposure (while keeping the source of the exposure anonymous), people will be able to reduce k and the spread of Covid-19.