Categories
科技報導

Why did Apple and Google’s blessing of the new crown tracking app fail in the United States?



Lack of coordination, inadequate testing, and distrust of technology put seemingly promising innovations in trouble. After the new coronavirus ravaged the United States this spring, Jodie Pond, the health director of Teton County, Wyoming, hopes to deploy a new tool to control the pandemic. Technicians are racing against time to develop applications that can quickly and quietly identify people who have been in contact with other people infected with the new coronavirus.

Traditionally, this “contact tracing” process is a difficult manual task: contact the infected person, trace their whereabouts in the past few days, and then find anyone who might have been nearby at the time. In the eyes of local health officials like Pound, this is nothing new. But the new coronavirus epidemic means tracing contacts on a larger scale, which may be the largest in history. Therefore, the talents in Silicon Valley recommend using our smart phones that we never leave our body to automatically identify people who may have been exposed to the virus.

To Pound, this scheme seemed very useful. Teton County is twice the size of Delaware and is home to 25,000 people. But Teton County is adjacent to Yellowstone National Park. In summer, 40,000 tourists visit Teton County every day. A smartphone application can help Pound’s contact tracking team to find people who have been exposed to the virus more quickly.

But, soon Pound discovered that this matter was not that simple. At the beginning, she chose an application developed by researchers at MIT, which uses GPS signals to track people’s trajectories. However, this application is not accurate enough to identify whether a person contacted with the virus has been in contact with other people. Therefore, in July, when the number of infections in Teton County surged, Pound’s small department had to return to the traditional manual method of tracing contacts one by one. “It’s totally beyond our ability to track contacts,” she said, “I feel we can’t control it at all.” She later hired more trackers, and she kept making calls on weekends. Later, Teton County issued a mandatory mask order-the only mask order in the state. Pound’s office is also working hard to arrange more virus tests in local hospitals. Finally, the epidemic was brought under control.

Why did Apple and Google's blessing of the new crown tracking app fail in the United States? 1

Pound still hopes to find an application before the winter ski season arrives. At that time, many people will gather together again. But in the following months, digital contact tracing changed. At this time, the digital contact tracking technology combines the system developed by Apple and Google, and can use more accurate Bluetooth technology. Pound’s developers intend to use this technology. Therefore, she also specially trained her tracking staff, readjusted the office’s record keeping system, and encouraged local residents to download the application.

Then, last month, she ran into trouble again. At a press conference in Cheyenne 400 miles away, the state health department responsible for contact tracing in some counties in Wyoming (not including Teton County) signed another app called “Care19.” Pound had never heard of this application. However, according to Apple and Google regulations, Care19 is the only application in Wyoming that can use Bluetooth on Apple and Android phones. Months of preparation were in vain. Pound’s team had to start again to prepare. She didn’t know if she could still complete the task now.

Why did Apple and Google's blessing of the new crown tracking app fail in the United States? 2

“It’s too difficult. It took a lot more time than we thought,” Pound said. “We are just a small health department. We just want to do our best.”

Pound’s difficulty is a microcosm, summarizing why digital technology has still failed to prevent the worst epidemic in this century. In the United States, the drawbacks of digital contact tracing are related to random responses. The result of this hasty response is more than 180,000 deaths and more than 6 million infections. Contact tracing relies on a large number of cheap and fast tests so that infected persons and contacts can be quickly identified. However, most parts of the United States do not have such capabilities. Contact tracing also depends on the public’s trust in the government and health system, as well as a social safety network for those who cannot work or need child care assistance while in isolation. But these are not enough.

The fragmented response of the US government is its own stumbling block. In European countries, it is straightforward for the whole country to respond to the pandemic, release applications and integrate with the existing public health system. But in the United States, policies, attitudes, and software systems vary from state to state, and even within the state. Conflicts are common-such as conflicts between state and local mask policies or different guidelines for who can be tested. Sometimes, the state government will handle contact tracing. Sometimes, local counties, cities, or university campuses handle this by themselves. More often, like in Wyoming, it is a mixed model. Without a unified national strategy, the local health departments in trouble can only make their own plans.

“The problem is not the app, but the fragmented nature of contact tracing and political issues. It’s a complete mess,” said Tim Brookins, the lead developer of the Care19 app. Currently, Wyoming and North Dakota are using the Care19 app.

Why did Apple and Google's blessing of the new crown tracking app fail in the United States? 3

Change strategy

When Apple and Google released their respective technologies in April, the two companies did not intend to intervene too much. Almost all smart phones in the United States are either Apple’s iOS system or Google’s Android system. The two companies will provide the underlying technology and guidance to the states. But states can choose their own developers to design and release applications-and decide how to integrate the applications into their local health systems. Originally, this was to allow states to have full flexibility, but never thought that digital contact tracing would become a political game and a headache for public health officials. Only six states-Alabama, Arizona, Nevada, North Dakota, Virginia and Wyoming-agreed to do so. The other 44 states did not agree.

Last week, Apple and Google tactfully acknowledged the flaws in this way of allowing states to choose. Now, they no longer rely on states to develop their own applications. Apple will provide built-in options in iOS, and Google will provide custom applications for states. However, each state still needs to choose to join, but technology companies will do more technical things. Nevertheless, the use of these applications on a large scale in order to bring real help to public health officials still needs to overcome the scattered pandemic response situation across the country. “Without public health infrastructure, everything else is useless. Unfortunately, we don’t have such a public health infrastructure to control the pandemic,” said Barack Obama’s CTO of the 2012 campaign. Said Harper Reed.

Last week, one thing Apple and Google did not change was their attitudes towards privacy and security. Their system design does not recognize individuals or track their location. The device shares the randomly generated code with compatible devices of the accessory via Bluetooth. When a user’s virus test result is positive and verified by the local health agency, the user will receive a string of keys sent to the central server. Then, their unique code will be broadcast to other mobile phones, so users who may have been in contact with a confirmed patient can receive a reminder message. In theory, no one will know who or when you were exposed to the new coronavirus.

At the end of April, in order to downplay the impact of the virus, Apple and Google changed their focus from “contact tracing” to “risk notification.” To avoid confusion, the two companies stated that the government can only grant a license to use the agreement to a development agency in each state or county.

But this system is not perfect. A study by Trinity College Dublin found that applications using Apple-Google technology did not perform well on buses and trams due to interference issues. The researchers suggested changing the threshold for “risk” from 15 minutes within six feet of an infected person to 10 minutes.

Even so, some countries are still anxious to develop applications that use this system. Germany and the United Kingdom abandoned their previously developed applications in order to adopt new applications based on the Apple-Google framework. In Ireland, about 40% of smartphone users have installed the country’s Covid Tracker app. Gar Mac Criosta, product manager of the Irish Health Service Administration Group, said that since the system was launched on July 7, more than 800 risk notifications have been issued.

Creosta said that Ireland’s work encountered privacy and security issues, but the government resolved the issues with a unified national plan and consistent information. He said that people responsible for designing and developing Covid apps across the country will meet regularly on Zoom to discuss ideas, issues and success. “People from all corners of the country gathered together on Friday night,” he said. But none of them are Americans.

In Switzerland, 35% of smartphone users have downloaded the country’s contact tracking application. Switzerland’s national digital contact tracing is handled by Carmela Troncoso, an assistant professor at the Swiss Federal Institute of Technology. Tronkoso said that recently, after receiving a reminder from the app, 26 people tested positive for the virus and then went into quarantine. Others may have also received the reminder and then chose to self-quarantine.

On average, Switzerland reports hundreds of confirmed cases every day. Tronkoso and her colleagues believe that the application can cut off some important new infection routes. Epidemiologist Salathé Marcel, an epidemiologist involved in the country’s anti-epidemic work, said: “Isolating one more confirmed case will reduce the risk. Because if there is no timely isolation, any confirmed case will be It may cause the next super spread event.”

It is not clear how many people will need to use the app to be effective. In April, an epidemiological model given by researchers at the University of Oxford showed that it would require 60% of the population to use contact tracking applications to effectively prevent an outbreak. But the study also shows that even if the usage rate is not so high, such as the level of Ireland and Switzerland, new infections can be controlled.

Tronkoso said that Switzerland’s anti-epidemic experience shows that the government’s support for contact tracing is very important. For example, after some employers tell their employees not to use the app to avoid delays in work, the government can launch a plan to compensate for absent workers. “It’s hard to say whether Switzerland has achieved success,” Troncoso said, “but we are working in this direction.”

Political intervention

In the United States, political factors affecting other aspects of the epidemic have also hindered contact tracing. Take South Carolina as an example. The state announced in May that it would use the Apple-Google framework to develop applications. At the time, there were only three states that made such a decision, and South Carolina was one of them. But the action is fast and the end is fast. In June, just before the summer adjournment of Parliament, lawmakers added new content to the Covid-19 spending bill, prohibiting state agencies from using digital contact tracking applications. Mark Sweatman, a lobbyist for the Medical University of South Carolina, said: “Some people worry that this contact tracking platform cannot be shut down and they may be tracked by mobile phones.” The Medical University of South Carolina originally planned to do so. The state develops a risk notification application. But by mid-July, the state’s confirmed cases surged, which exceeded the ability of public health officials to track cases manually.

This suspicion also partly reflects bad experiences elsewhere. For example, Utah invested millions of dollars in April to develop GPS-based tracking applications. But few people downloaded the app, and the county public health department refused to use it, and then Utah shut down the app’s location tracking function. Other scheduled applications are either not accurate enough or are found to send location data to third parties. Opponents in South Carolina do not believe that the Apple-Google system will do better than GPS-based applications in terms of privacy protection.

The scattered response measures of the United States have also brought about various strategic conflicts. In Teton County, local personnel are responsible for most of the pandemic response. Local officials take actions to develop applications themselves, but state-level intervention disrupts everything. On the other hand, at the University of Illinois, researchers designed a Bluetooth-based risk notification application for the campus, but found that the application could not use the Apple-Google system because the state did not grant school developers permission to use the system . Sweetman said that South Carolina Medical University has repeatedly applied to use its application on the university campus, but because the state health department did not provide support, the school has not been able to use the tool.

In addition, states like Massachusetts and New Jersey have shelved or simply abandoned plans to develop contact tracking applications, instead focusing on other more meaningful things. “All states want to work hard to control the epidemic, including very common and traditional contact tracing,” said Margaret Bourdeaux, a researcher at Harvard University’s Kennedy School. Inadequate testing in certain areas even makes traditional contact tracing difficult. Although some states have initiated large-scale contact tracing projects, the follow-up management cannot keep up. Researchers at George Washington University estimate that contact tracing personnel in many states are far from adequate. For example, California. According to the researchers, California needs at least 30,000 contact tracing personnel to keep up with the spread of the virus. However, a survey conducted by National Public Radio last month found that the actual contact tracing personnel in California are less than one-third of what is needed.

Contact tracing personnel stated that their job is not easy. The other party either does not respond or does not cooperate with quarantine or testing recommendations, especially when the tracing personnel cannot provide assistance such as child care. The distrust of the government, especially the distrust of immigrants and low-income people, coupled with the rumors about contact tracing on the Internet, makes everything worse.

Justin Lessler, a professor of public health at Johns Hopkins University, has been studying the effectiveness of contact tracing programs. He said: “This involves a basic issue. Some people are beginning to feel disappointed with virus control measures such as testing, tracking, and quarantine.” He suggested not to be discouraged by people’s disappointment. Rather than being discouraged, it is better to invest more energy and resources to make the plan really work. “The challenge here is to follow up quickly, trust, and isolate and help them protect other family members.”

He said that digital contact tracing can facilitate these efforts and reach more people more quickly, especially those who do not know they have had contact with an infected person. However, digital tracking will prevent disadvantaged groups from using these applications. If these issues are not taken seriously, no amount of effort will only widen the gap. Bourdo said: “When a health crisis occurs, public trust is the foundation of everything. Therefore, anything that may threaten trust-especially when the benefits are not clear-is not worth doing.”

Seek coordination

From the beginning, some people have realized that the United States needs a more national response. “We are not Canada,” said Scott Becker, executive director of the Association of Public Health Laboratories (APHL), which has a national healthcare system and a national contact tracing application. “If there was a national measure, everything would be much simpler. But in the United States, our public health system does not work like this.”

In May, a group of officials, including state health officials and technical experts from the Centers for Disease Control and Prevention, formed a working group to address this issue. Baker, who later joined the working group, said that an obvious question is how to allow the exchange of information between state applications. Although Google and Apple provide common technologies for exchanging keys via Bluetooth, the states still use their own servers to send notifications. This creates trouble for interstate travelers and commuters, who may have to use applications that connect to different systems. (In addition, setting up a separate server for each state is also detrimental to security.) Therefore, APHL, which often serves as a public health data channel between local and federal agencies, intends to set up a general server to facilitate the sharing of data between state applications.

But for states, the bigger technical problem lies in determining their needs for contact tracking applications and finding suitable people to develop applications. The latest announcements from Apple and Google have potentially eased development pressure for various states. But it remains to be seen whether these new applications (called “Exposure Notification Express”) can achieve greater success in controlling the spread of the virus. One big question is how to establish contact with local health officials who oversee traditional contact tracing. “The real problem is the way contacts are traced,” Brookings said.

As Brookings said, the initial plans of Google and Apple seem to have recognized the challenge of this “last mile problem” and intend to rely on developers to develop custom state applications. For example, Brookings took these local differences into consideration when conceiving Care19. In North Dakota, every campus has its own set of guides and resources after the university resumes. Therefore, Care19 allows people to register with school information and receive local guides. Similarly, he said that in Wyoming, due to the jurisdictional complexity of contact tracing, places like Teton County can customize application functions according to the needs of local residents, while allowing public health officials to view more application usage data and The total number of reminder notifications sent by the app.

Pound said she is willing to use Care19, but is still uncertain about the outcome. There is no IT department in her office. At the same time, she must solve many other problems before considering using other applications. In addition, many residents of the county still retain the previous GPS application on their mobile phones. Recently, she also trained her contact tracing staff to use the Care19 application as a storage tool, that is, to understand where the infected person has been before, so that the tracing staff can quickly contact other people who have had contact with the infected person. However, this is not the risk reminder system she wants. She is still worried about the upcoming ski season. However, in the local anti-epidemic work, it is always a good thing to have an extra tool on hand.