After months of darkness, the nation received a glimmer of light as pictures of frontline workers getting vaccinations hit social media. But that small ray of light is dimming as nearly all states struggle to administer the vaccine with no federally coordinated rollout plan, and faulty technology guiding their efforts.
The vaccine has been distributed to hospitals in all 50 states, with rollout plans for first responders and healthcare workers. But what happens next? How do vaccines get into the arms of a nation of 328 million people who are strongly divided and frequently wary of government? We are already witnessing technical failures from California to Maine as users struggle with finding vaccination locations and signing up for time slots. In Florida, several counties have turned to using the invitation website Eventbrite for vaccination scheduling, as they don’t have their own scheduling capability.
According to someone involved in Operation Warp Speed who asked to remain anonymous due to the politically charged environment, there is no plan for the Warp Speed team to develop a system for last-mile delivery. Operation Warp Speed successfully helped develop a vaccine in record time, and relied on logistics experts from the Department of Defense to move the vaccine from production facilities to hospitals while ensuring the vaccine stayed refrigerated—not an easy task, but one that falls well within government’s expertise. But the project’s scope of work doesn’t include getting the vaccine from hospitals into people’s bodies. This means that any vaccine distribution or immunization tracking will need to be state-based.
There is a solution. Smart use of data and technology could save the day. Unfortunately, decades of technological neglect in government means that federal and state agencies are woefully ill-equipped to handle the tremendously complex task that looms in front of them. So here we are, in the early stages of what could easily develop into a full-fledged technological disaster, leaving huge portions of the nation unvaccinated or semi-vaccinated (the vaccine requires two doses) for far longer than necessary while the pandemic rages on.
Any vaccine rollout plan must surmount enormous challenges, and one way to do that might include carefully thinking through how to apply the right tech tools at the right places to achieve well-planned-out goals. But a standard response to such a crisis—across both public and private sectors—is to throw any old technology at the problem.
When I (Hana) worked with the Transportation Security Administration (TSA) at the United States Digital Service—a federal team of technologists—the agency was recovering from the stinging fallout of its Randomizer app, which purported to reduce racial profiling by TSA agents. Instead, it was a $47,000 app that randomly generated an arrow that pointed left or right. When I came on board, the agency was facing another problem—security lines so long that they stretched out of the airport and caused people to miss their flights. TSA leadership proposed solving the problem with . . . an app. This time the team conducted user research, which determined that an app was not the right solution, and the plan was scrapped.
Nearly everyone appears to be falling victim to the “throw technology at the problem” path that so many government institutions and private sector companies have charted. The Centers for Disease Control and Prevention has rolled out two separate systems—Vaccine Administration Management System (VAMS) for users to register for vaccines and V-safe to track vaccine side effects and get reminders for a second vaccine dosage. Both systems require excessive personally identifiable information that may discourage participation.
States have been offered the option to use VAMS for their own tracking. But the portal can be onerous for users: clunky, time-consuming to use, and weighted down with unnecessary questions like workplace name and work address. The registration process alone can take even tech-savvy users 10 minutes to get set up. (Yes, we timed it.) Only then can users attempt to schedule a vaccination appointment.
This approach is the opposite of how most users will approach the system, which is likely with the question, “Where and when can I schedule an appointment?” There is no reason for systems meant to speed the vaccine rollout to slow users down by attempting to collect the amount of information someone might anticipate offering up as a clinical trial participant. That degree of data collection will be impossible for the general population. It is also unnecessary for vaccine recipients beyond healthcare workers. For the general public to consider using it, the system will need to be completely rethought and redesigned. We need to cut out those invasive questions. This is what we have done with contact tracing and exposure notification apps that I (Ramesh) and my team have built for six U.S. states and territories with PathCheck Foundation.
Currently, 19 states and the District of Columbia have signed on to use VAMS. Others are considering VAMS as a backup, and are looking to expand the use of existing vaccine tracking databases, called immunization information systems (IIS). Ten states have said they will use PrepMod, a system developed by a Maryland nonprofit that some states already use to track vaccinations. PrepMod takes users through nine different screens in order to register for a vaccination. In other words, it is not designed for emergency situations like the one we are facing today. In a demo on YouTube, the woman demonstrating the system refers to it as “a little glitchy.”
New York City just expanded vaccination candidates to include anyone over the age of 65, in addition to frontline workers. But registration requires two-factor authentication just to set up an account in New York’s system, and then a six-step process to set up an appointment. The comments on New York City’s tweet announcing the vaccine registration page illustrate how much trouble people are having already, with some saying they spent four or more hours trying to register and find a place with available vaccination times, only to give up in frustration.
How could this oncoming chaos be averted? We must apply modern techniques like data analytics, user research, and usability testing to learn about the vaccine and immunization process from the perspective of different Americans, all while preserving privacy and people’s right to remain anonymous. These data points could be applied to the population in two critical ways that could speed the nation’s healing and rate of immunization—and help stabilize the economy.
First, states clearly need help with their scheduling software. The federal government or the open-source community could easily come to the rescue here by making available a free, open-source tool that is custom-built for COVID-19 immunization. The tool should be designed to allow people to find appointment availability and schedule their appointments in less than three minutes. It should also allow all residents of a state to sign up today, regardless of their priority group. Perhaps this is done by lottery number, perhaps there are other ways of staggering enrollment times. But no one, in 2021, should be struggling with glitchy appointment technology.
Second, we need to know when we have reached the herd immunity threshold. Yes, we can look at how many vaccines have been administered, but without knowing how those vaccines have been distributed throughout the population, a semi-vaccinated population could create a highly dangerous mix, especially given that we don’t yet know if people who receive the vaccine can still pass along the virus to others who have yet to be inoculated. So tracking where the vaccines have been administered and to whom will be critical.
We know from our research with immigrant populations on vaccine rollout that many of them are at risk of not getting vaccinated while simultaneously being more likely to catch the virus due to their work and living environments. People who do not speak English as a first language, or who came from countries where there was little to no trust in the government, may be hard to reach and are unlikely to proactively sign up for a vaccination.
Many immigrant-focused organizations are already working on language access plans for non-English speaking populations. In our work at New America with these organizations, we have found that immigrant populations have good reasons to be nervous about raising their hand for a vaccine. Many families may have undocumented members mixed in with citizens or green card holders. If one person raises a hand to get vaccinated will that jeopardize the immigration status of other family members?
For this reason, it will be critical to understand where immunized people live, while also anonymizing the data so no one’s privacy is violated. Imagine, for example, that the residents of the Upper West Side of Manhattan have all received the vaccine and the majority of the residents of Chinatown have not. That would be useful information to know, but we won’t know it without a data plan for tracking who has received vaccinations. We also won’t know it unless the data can be anonymized because a large percentage of people will be unwilling to provide that kind of identifying information to the government. And without data scientists and technologists working on the problem, we’ll have a long wait for a well-thought-out solution.
We need to start thinking about what a mid-vaccination world looks like. What happens when a significant percentage of the population is immune and another significant percentage remains unvaccinated? This is the challenge that several apps have suggested they could meet. Multiple airlines announced that they are working together to launch CommonPass, which will store COVID-19 test and immunization records digitally. Such an app could mean that states and countries allow only immunized people to board planes, go to basketball games, or eat in restaurants. It takes only a very small leap to see how this could go very, very badly. First, it opens the door for discriminatory practices using the app as a proxy. Second, and perhaps more important, how will the immunization data get into the app?
“Today’s technology allows individuals to control their own data and participate in group computations without exposing personal information,” said Latanya Sweeny, professor of the practice of government and technology at the Harvard Kennedy School and director of the Data Privacy Lab at Harvard University. “Individuals get privacy guarantees and society gets useful answers.” Sweeny recently launched MyDataCan.org, a platform that allows individuals to control who has access to their data and contributes to public health databases without exposing personal information.
But with no coordination from the federal government, we could be looking at a national future where 50 different states have 50 different immunization tracking apps, much as we do with drivers’ licenses. Of course, with paper documentation states have reciprocity—for example, Texas authorities must recognize a Connecticut driver’s license. Will the same be true of an app? Thus far, it hasn’t been, as evidenced by the fact that every city has its own parking payment app. And also its own 311 app. And also its own pothole reporting app. And so many more.
Our current federal government expects each state to develop a vaccine and immunization plan. We can only hope that the incoming administration will feel differently. On their own, the states are unlikely to quickly and effectively tackle this highly complex undertaking. Even worse, they may not know that they aren’t equipped. It’s just technology, after all. Technology will solve it.
At the moment we are on track to live through years of immunization chaos, privacy infringement, and important data lost to time. We beg the states and the federal government: Please do not let that happen. There are entities in the federal government—18F and the United States Digital Service are two—that could work with states and federal agencies to help design and deliver technology that would serve all states best. The solution could be open source, to allow each state to customize as they like.
To the incoming administration: Please develop a coordinated vaccine and immunization plan. Use technology if your research shows that tech is the right solution. If it isn’t, consider a paper-based solution from pandemics of the past. Seize the opportunity to show the world that, with the right structures in place, the U.S. can develop smart technology in the public interest that also protects people’s privacy. This is a chance to start to shake off the backward momentum of the last year. Let’s take it.
Hana Schank is the director of strategy for public interest technology at New America, and the author of the book Power to the Public: The Promise of Public Interest Technology, due out in April from Princeton University Press.
Ramesh Raskar is an associate professor at MIT Media Lab and the founder of PathCheck, a nonprofit dedicated to creating healthy and resilient communities by containing the pandemic, revitalizing the economy, and preserving the privacy of individual citizens.