The week before Christmas, hundreds of medical residents at Stanford University Hospital joined an emergency Zoom call. They had been brought together by shared outrage at their administration’s allocation plans for its first 5,000 doses of the newly authorized vaccine for Covid-19, the pandemic that had defined their past year. Only seven of those shots were reserved for residents, the lowest-ranking physicians, even though they’re more often exposed to patients infected with the coronavirus than other employees whose work had been almost entirely remote. But some of those employees—including hospital executives and dermatologists who’d only seen patients virtually—were nonetheless ahead of them in line.

For Angela Primbas, an internal medicine resident at Stanford, it was the last straw. She and her colleagues had been putting in 80-hour weeks caring for Covid patients, often missing out on their program’s educational curricula to pitch in and staff wards instead. They’d also recently gotten word that a shipment of the N95 masks they’d been using had been defective, leaving the young doctors vulnerable to exposure. They were more anxious than ever to get the vaccine, only to discover they’d been left out.

“There was just a lot of emotional and physical stress that had been piling up over the course of many months, and then to be just completely excluded from participating in the one bright spot—the light at the end of the tunnel—was so upsetting,” Primbas recounted by phone. She and the other residents decided to take action. The hospital wanted to kick off its vaccine rollout with a public-facing photo op, and they were going to get one.

The following morning, hundreds of residents, physicians, nurses, and other supporters staged a major protest at Stanford Medical Center, demanding that workers with the most contact with patients be first in line. Spokespeople told multiple media outlets that they took full responsibility for the problem and would right it immediately, blaming the error on a flawed algorithm for determining who’d get vaccinated first.

Residents have indeed been vaccinated since their headline-grabbing demonstration, but not before hearing from friends in programs elsewhere whose experiences paralleled their own. In hospitals like University of Chicago and Johns Hopkins, physicians told me, work-from-home PhD students in their 20s were routinely offered vaccines they believed would be better off given to patients.

Snafus across the country have gone well beyond snubbed hospital residents. Since vaccines were sent rapidly out to states, high-profile screwups have dominated media coverage of the effort. Federal contracts with CVS and Walgreens to vaccinate nursing homes dragged well behind schedule. Spanish-language sections on enrollment websites spouted misinformation. Hundreds of hopeful recipients camped out at rumored distribution sites only to leave without jabs. Untold numbers of unused doses wound up in dumpsters, while vaccine targets nationwide fell millions short.

Such disasters reflect the immense challenges of implementing the largest mass vaccination program in US history, which until recently was helmed by a federal government actively hostile to it. As the Biden administration settles in and vows to ramp up coordination of and financial support, state and local efforts will scramble to make up for lost time. Their ability to do so will depend on their willingness to reach the patients that the 21st century has left behind.

From the early days of the global coronavirus pandemic, societies have, to varying degrees, adhered to measures like social distancing and school and business closures. The goal was to “flatten the curve”—to slow the virus’s spread to avoid overwhelming hospitals’ intensive care units—in hopes that a vaccine would soon be available. And not long after, it was: largely thanks to investments of billions of dollars from the US and German governments, pharmaceutical giants Pflizer and Moderna both produced vaccines that boasted around 95 percent efficacy in clinical trials, greatly reducing symptom severity in vaccinated patients. These results clinched emergency use authorizations from the Food and Drug Administration, and the vaccines began making their way into arms less than one year after the novel pathogen arrived on American soil—an absolutely astonishing timeline.

After nearly a year of incalculable loss—topping 400,000 American deaths, not to mention countless hours with friends and family deferred—the vaccine is a ticket back toward normalcy. But that normalcy may elude us until upwards of 90 percent of people develop antibodies against the virus, either through vaccination or infection. Given the unknowns about how long protection from infection lasts, reaching so-called “herd immunity” will require getting shots into nearly everyone in the country.

There’s really no precedent for that. While mass vaccinations have played a key role in United States public health policy, they’ve tended to be somewhat targeted by geography or age: Specific neighborhoods or cities were vaccinated against smallpox in response to outbreaks in the late 19th and early 20th centuries, a few million doses of polio vaccine were administered to grade schoolers in the 1950s and 60s, and a slate of childhood immunizations are still delivered on a routine basis today. But rolling out a vaccine to hundreds of millions of adults in a relatively short time is an entirely different situation: while children almost always have institutional relationships with schools and pediatricians, their ties to potential service providers can loosen with age and allow them to fall through the cracks. Getting adults to the right place at the right time, despite varying schedules, care responsibilities, access to healthcare, relationships with the state and levels of trust in medicine, is a formidable project.

I asked Jason Schwartz, assistant professor of health policy at the Yale University School of Public Health—an expert in vaccine policy who spends every day of his life thinking about this stuff—if he’d ever imagined what exactly a nationwide mass vaccination program would be like. He told me he hadn’t. “This is so far beyond our vaccination playbook that it explains why so much of this work is being envisioned, imagined, and implemented in real time,” he said. “We have so few lessons to draw on, other than imperfect analogies to other aspects of vaccination.”

Public health departments across the country have had hypothetical plans for mass vaccinations on the books for years—a tendency that ramped up in the 21st century in response to concerns about germ warfare in the aftermath of post-9/11 anthrax attacks. But since those were written, public health department funding has been gutted; as Kaiser Health News reports, Great Recession–era austerity measures still haven’t been fully reversed over a decade later. Today’s state and local health department budgets are 16 percent and 18 percent smaller than they were in 2008, and employ nearly 40,000 fewer people. Further compounding the problem is that these plans have rested on a reasonable assumption that simply isn’t true of the Covid-19 pandemic. “The plans that were developed prior to 2020 pretty much all assumed we would have federal leadership and federal financing, said Lindsay Wiley, director of health law and policy at the American University Washington School of Law. “There wasn’t really a plan in place where the idea was ‘what free commercial event planning software can we use if we get zero federal leadership and support?’”

Wiley’s framing is hardly an exaggeration: Mere days after the inauguration of President Joe Biden, reports began to circulate that there was no existing federal Covid-19 vaccine distribution plan for the new administration to inherit. This punted responsibility to state and local health departments, whose long-awaited funding for vaccine distribution, passed through Congress’s December stimulus bill, only just began to trickle to recipients in mid-January—months after it would have been most useful. “We’ve been calling for funding—screaming for funding—for months,” Claire Hannon from the Association of Immunization Managers told me by phone. “Obviously, it’s better late than never, but it’s difficult to only get funding after the vaccine has been rolled out.”

When I asked Hannon about how the delayed federal cash will affect the overall rollout, she said she’d always been much less worried about the first stage of vaccinations than she was about the waves that come next. Phase 1a—made up of the highest-priority vaccine recipients, according to the CDC—was arguably the easiest part. As states move on to Phases 1b and 1c, debates have raged over who should get shots first and why. But as it turns out, vaccine prioritization is less a philosophical question than a logistical one.

Even as the Trump administration left public health departments adrift, with no federal support or coordination for months on end, it had little hesitation about funneling resources into the private sector. Multimillion-dollar contracts were awarded to Walgreens and CVS to administer doses to the country’s 3 million nursing home residents, who along with 20 million health care workers comprised Phase 1a of vaccine allocation. But the drugstore giants lagged weeks behind schedule in state after state, with Oklahoma, Michigan, and Mississippi going so far as to beg the federal government to allow them to reassign nursing home vaccinations to other pharmacies or public health officials. Aharon Adler, a nursing home manager in Chicago, struggled to get information from CVS before they arrived to vaccinate workers and residents. He’d been prepared so inadequately for the big day that he hadn’t even been told that shot recipients had to stay for observation in a socially distanced room, and the only space he’d designated was too small, slowing down the process by several hours. When we talked by phone, Adler still hadn’t been able to confirm with CVS when exactly they’d return for the second dose. Notably, the only state that didn’t work with CVS or Walgreens—West Virginia—also became the first to successfully vaccinate all of their nursing homes.

While CVS and Walgreens were woefully botching the nursing home rollout, high-profile incidents like Stanford’s allocation algorithm began stoking outrage on social media. New York’s Governor Cuomo endeavored to combat such unfairness by threatening stiff penalties for institutions that vaccinated anyone “out of order,” which reportedly spooked some hospitals into throwing unused doses in the trash instead. Meanwhile, relatively substantial numbers of people included in Phase 1a reportedly declined the vaccine, or preferred to take it later once they’d seen others do so safely.

The combined impact of these mishaps was that the early stage of the rollout underperformed projections by several million doses. Those meager numbers—as well as anecdotes about undeserving recipients and overemphasized but rage-inducing images of shots piled up in garbage cans—began fueling a backlash against what was by January being widely characterized as a disaster.

At this point, a growing chorus began chucking the baby out with the bath water. People were right to be angry at how the first month of the rollout had gone. But instead of blaming players like a callous federal government and drugstore giants who’d failed to deliver on promises, many onlookers ascribed the mess to the concept of vaccine prioritization itself. Dictating what groups get the shot first, they argued, straitjackets the process, when we really need to just get shots in arms. As Phase 1a finally drew to a close, the far harder work loomed. And for Phases 1b and beyond, the argument went, public health departments ought to broaden eligibility beyond vulnerable subgroups and focus simply on speedy injections at a massive scale.

The title of an essay from bioethics think tank The Hastings Center put it succinctly: “Ethics Supports Seeking Population Immunity, Not Immunizing Priority Groups.” Just before Trump left office, his administration endorsed this view, stipulating that anyone over age 65 should now be eligible to receive a vaccine. “We’re telling states today that they should open vaccinations to all of their most vulnerable people,” Health and Human Services Secretary Alex Azar said on January 12. “That is the most effective way to save lives now.” Several states, including Florida and Louisiana, have followed this directive. In Texas, state health officials went so far as to spike a Dallas plan to start vaccinating high-risk communities of color in favor of a broader, non prioritized program.

That shift may make intuitive sense, but it doesn’t hold up to scrutiny. While it’s good to allow some flexibility in vaccine distribution guidelines to avoid unforced waste, prioritization schemes are far less of a limitation than the fact that states are still struggling to build up supportive infrastructure to do this—not to mention the scarcity of doses in the first place. Adding tens of millions of people to the list of now eligible recipients doesn’t make that any easier—it would be like addressing long waits at the grocery checkout by doubling the number of people in line, instead of opening up more cash registers.

After all, “prioritization” isn’t just a matter of making a list with the power to magically summon arms in a particular order—it requires an active outreach strategy. Given how long the largest mass vaccination program in American history will take—perhaps nearly a year, per conservative estimates—it makes sense to strategize how to target both those patients most likely to die from the virus and those most likely to transmit it. Counterintuitively, doing away with prioritization in favor of speed and efficiency will actually do less to save lives, because the people most able to navigate the process of accessing the vaccine are overwhelmingly the least at risk. Figuring out how to enroll through a website or care provider, being able to take off work, and traveling to a vaccine site for two different doses are all rendered easier by class privilege.

This dynamic has already borne out starkly: in Washington, D.C., the number of early vaccine signups in a given neighborhood is directly correlated with how wealthy its residents are. In Chicago, race has proven a predictive factor. Unsurprisingly, wealth and race also correlates to employment in high-risk frontline jobs, affliction from debilitating comorbidities and residence in overcrowded housing most likely to drive infections. In other words, a passive approach of “first come, first serve” practically guarantees that the people who are safest from the virus—richer, whiter, more connected people who work from home or can otherwise afford to hide there—will comprise the early wave of vaccination, as the people most likely to die or spread it remain unprotected. Beyond being unjust, that ensures the societal benefits of vaccination will be as minimal as possible.

But what exactly does effective outreach strategy look like? When I asked experts who’s doing it right, I kept hearing the same surprising answer: Perhaps the best model for vaccine distribution in the country right now is happening in Central Falls, R.I.

Home to around 20,000 people, Central Falls is a city of superlatives: It’s the most densely populated city in the state, the poorest, and the only one with a majority of residents of color. It was also the most affected by the coronavirus epidemic, with case rates per capita doubling those in hard-hit areas of New York.

Overwhelmed by his duties as the public health commissioner of Central Falls, Dr. Michael Fine began researching other countries’ coronavirus mitigation strategies last spring, as infections surged in his own community. “When you look around the world,” Fine told me by phone, “it’s very clear that the places that have done best with coronavirus have been places that put people to work and invest in a lot more public health presence than the United States does.” He set out to apply those insights, using money distributed to the city to hire 15 so-called “health ambassadors” from both Spanish and English-speaking communities within Central Falls to implement the city’s pandemic response measures.

Since the spring, Fine told me, the health ambassadors have donned bright orange uniforms and maintained a presence at busy spots in town, like outside the Dollar Tree and City Hall. They handed out masks and talked to passers-by about why they were important, eventually driving local mask usage rates from less than 50 percent to over 90 percent. Later, they helped remind locals to get their flu shots, and helped enroll eligible participants in early vaccine trials. As Fine tells it, the health ambassadors became well-known and credible conduits for critical health information, relaying messaging within their own communities in ways officials could not.

And now, as the city rolls out the vaccine, the health ambassadors’ role is more important than ever. As part of a pilot program to stress test vaccine distribution, the State of Rhode Island opted to focus first on Central Falls—zeroing in on a highly distressed ZIP code, and affording local officials latitude within it. To kick off the program, Fine and the vaccination team—complete with health ambassadors who’d already been stationed there—set up a clinic at the public housing authority, knocking on doors and vaccinating everyone who accepted a shot. Both Fine and Central Falls Mayor James Diossa both got their first shots on-site, to demonstrate the vaccine’s safety. But Fine believes the health ambassadors helped things go as smoothly as possible: “The ambassadors were there with the teams interpreting, and because of their very local presence, I think it was more comfortable for people.”

Once the vaccination teams had worked through the public housing buildings, the ambassadors resumed their stations around town—this time, enrolling and teaching residents about vaccines and when and how to get one, like on one Saturday morning at the Kiwanis Club parking lot. And each morning before the start of their shifts, they have a bilingual Zoom meeting to discuss what they’re hearing about the vaccines, how to get people excited for them, how to assuage anxiety or quell rumors swirling about them. Fine encourages them to discuss their own experiences getting vaccinated to reassure their neighbors. In one meeting I was invited to, one health ambassador described how he was running into fewer and fewer people who had misgivings about the vaccine, and more and more people excited to get theirs.

The Central Falls model offers an effective strategy not only for prioritizing vulnerable people but actually reaching them. That’s what’s missing from discussions about vaccine distribution—as it turns out, the biggest logistical challenge of turning vaccines into vaccinations isn’t maintaining extremely cold storage or even reaching consensus on who “gets it first,” but how to connect and coordinate with patients who are often by definition among society’s hardest to reach.

Models like that used in Central Falls—actually going out into communities, and knocking on doors or setting up tables and clinics—have been successfully deployed by public health departments for diseases like tuberculosis, another deadly respiratory disease that shares Covid-19’s predilection for the poor and vulnerable. For example, I once wrote about an outbreak among undocumented Chinese immigrants in 2013 and 2014. NYC public health officials were able to trace several cases to an Internet café and karaoke bar in Sunset Park, Brooklyn, and sent workers there to test regulars on site, identifying and treating several additional cases. TB caseworkers also routinely bring medicines or administer antibiotic injections to patients in their homes and workplaces, saving them the burden of traveling to the clinic and making the sometimes lengthy treatments easier to adhere to.

Devising ways of reaching people less institutionally connected to the healthcare system has been central to many public health initiatives, Nabila El-Bassel, director of the Social Intervention Group at the Columbia University School of Social Work, told me. Mass vaccination teams should draw on those lessons, she says: “I’m thinking about people who use drugs, people in homeless shelters, in soup kitchens, in domestic violence shelters, or in community supervision programs…. If we want to get into these populations, we’ve got to think about nontraditional sites and strategies. We can’t just wait for them to come to us.”

Experts have long debated how to handle the so-called “last mile problem,” or the logistics of getting a vaccine from the warehouse or hospital into the arm of a patient. Sometimes, the best option is to travel the last mile for them.

After all, vaccines may be the single most life-saving invention in the history of medicine, but no disease has ever been beaten by science. Turning vaccines into vaccinations requires vast amounts of resources and labor: investment in transformative pharmaceutical research, manufacturing operations, shipping and storage, administrative coordination, public messaging, pharmacists and health care workers, clinic supplies and planning, community outreach and ways to keep them all on the same page. How those elements are marshaled, and on whose behalf, aren’t questions that science can answer.

Those fights happen squarely in the realm of politics: As President Biden’s administration sets to work building a federal distribution plan from scratch, and the deposits from the second stimulus bill finally hit state and local health departments’ accounts, we’re finally in place to start catching up to make mass vaccination work. If we do things right, and implement strategies for meeting the most vulnerable people where they are, the amount of sorrow wrought by the coronavirus will be all but stamped out by the time we pass 300 million vaccinations. Should we fail, the outcomes will look more or less like the past year—with sorrow and death doled out to people who deserved shots instead.