To hear more audio stories from publishers, like The New York Times, download Audm for iPhone or Android.

On the evening of March 4, James Cai, a 32-year-old physician assistant, was languishing on a cot, isolated in a small, windowless room on the emergency-room floor of Hackensack University Medical Center, when the television news caught his attention. Before that moment, Cai had been in a strange medical limbo, starting midday on March 2, when he left a medical conference in Times Square because he had a bad cough. Instead of heading to his home in Lower Manhattan, he texted his wife that he was going to spend the night at his mom’s place in New Jersey. His mother was out of town, and if he had the flu, he could spare his wife and their daughter, a cheerful 21-month-old who clung to him when he was home, the risk of catching whatever it was. That was Cai: cautious, a worrier, overprotective, the kind of medical professional who liked to rule out the worst-case scenarios first.

At his mother’s home that evening, he waited until about 8 o’clock., when he thought the urgent-care facility nearby would be relatively empty, then headed over for a flu test. If it was not flu, he could think about going home. He put on a mask before the doctor examined him and learned that his heart rate was elevated, which did not surprise him: He could feel the palpitations. He got a flu and a strep test and asked for a Covid-19 test as well, only because they might as well be exhaustive; but the doctor told him he did not have the test, and neither of them thought much about it after that.

On March 2, many doctors on the East Coast still saw Covid-19 as an ominous but distant threat. Although several elderly people had died by then of complications from coronavirus in Washington State, the outbreak seemed mostly contained to that part of the country. Only two people on the East Coast had tested positive: a health care worker from Iran and a lawyer from New Rochelle, N.Y., whose results were reported the same day Cai went to the doctor. At the urgent-care center, the doctor reported that his chest X-ray looked normal, and the flu and strep tests came back negative. But the doctor was worried that Cai’s symptoms — that cough, surprisingly powerful for something that had kicked in so recently, and an elevated heart rate — were consistent with a possible pulmonary embolism, a clot in an artery in his lung that could prove fatal. He advised him to go immediately to the nearest emergency room, H.U.M.C., where they could give him a CT scan, which would provide a more detailed picture. Cai drove to the hospital and waited for his scan on a cot in a hallway. Not long after, he was moved to the small, windowless room, where he started to feel even worse: short of breath, feverish. He had diarrhea, and the brief walk to the bathroom nearby left him exhausted. He took a video of himself to show his wife, and in it he looks a little wild-eyed; he is breathing fast, as if he has just been chased and whatever was chasing him is right outside the door.

The next morning, March 3, not long after his CT scan, a nurse came to give him a Covid-19 test. The nurse was wearing full personal protective equipment, which typically includes eye protection, a respirator mask, gloves, a long gown and a head cap. The hospital had not tested him earlier because the C.D.C. guidelines at the time suggested that testing should be reserved for those who had recently traveled to China or come into close contact with someone believed to have the virus. Cai had not been there for years and to his knowledge had not been in contact with anyone who had tested positive. Now he thought they were just being thorough.

The following day, an infectious-disease doctor, Bindu Balani, a calm woman with a gentle delivery, came to see him in his room, also wearing P.P.E. She explained to Cai that he did not have a pulmonary embolism, but that they could see on the scan that he did have pneumonia, and that she would start him on antibiotics. Also, it became clear to Cai that something about the CT — a shading in one lung — had given them cause to test him for coronavirus.

Balani was measured when she explained this plan — they wanted to rule it out — but after she left, Cai started Googling symptoms from his stretcher in the small room and asked his wife to do the same. He saw that his symptoms, which had intensified since earlier that day, matched up almost perfectly with those of Covid-19: cough, heart palpitations, fever, diarrhea, chills, fatigue, shortness of breath. It would almost have to be coronavirus, except that there was no way it could be coronavirus: What were the odds that he, James Cai, 32-year-old mediocre basketball player, doting father, conscientious physician assistant, intrepid foodie, would be the first person in all of New Jersey to come down with it?

Cai hated being in that room on the emergency-room floor, where all night long he could hear people crying out in pain or wailing in grief. He tried to tune it out, to take in the consoling texts from friends who knew he was at the hospital. His fever kept spiking to around 102 degrees. The isolation only made him feel worse.

The next day, the hours passed slowly as Cai awaited the results of his test — until that evening, when Cai looked up at the television in his room. The evening news was showing a large image of a post that had just shown up on the Twitter feed of the governor of New Jersey, Phil Murphy. “Tonight, Acting Governor @LtGovOliver and I are announcing the first presumptive positive case of novel coronavirus, or #COVID19, in New Jersey,” the tweet read. “The individual, a male in his 30s, is hospitalized in Bergen County.” Cai’s heart rate, already too fast, sped up, and he felt the chill of his own sudden sweat. Please, God, don’t let that be me, he thought.

He held up his cellphone, shaking a little — from fever, from shock — and took a photo of the image of the tweet on the television news. He was sure that the governor was talking about him, and yet he was praying that he wasn’t. Soon after that, an emergency-room doctor came in and told him what he’d already known in the deep part of his psyche that always prepared for the worst: Cai was in fact the first patient in New Jersey to test positive for Covid-19.

Cai worked about six days a week for a medical practice that had four offices around the metropolitan area, most of them in heavily Chinese and Chinese-American neighborhoods like Flushing and Chinatown. Many of his closest colleagues and friends were immigrants and medical professionals like him. As soon as he saw the television news, Cai had texted the photo he’d taken to one of them, his close friend Yili Huang, a cardiologist in private practice and affiliated with Mount Sinai. “It can’t be,” his friend wrote back. Now Cai let him know that it was true: The test was positive. Earlier, Balani, trying to reassure him, said that even if he had it, he was most likely past the worst phase of a coronavirus infection: the first two days. “She didn’t lie to me, right?” he asked his friend. Huang tried to be comforting, “Of course not,” he wrote. But now that Huang knew that his friend really had tested positive, it dawned on him that Cai was alone in a room facing what could be a life-threatening virus, in a hospital where no one had ever encountered it.

Cai and Huang met five years earlier at a professional dinner. Each came to the United States when he was young, Cai at 14, Huang at 11. They instantly bonded over their love of the Shanghai waterfront and their similar accent (“a charming accent, very smooth,” is how Huang describes it). Huang had, among many of his friends, a reputation as a big-brother type — someone who followed up to see how your mother was feeling if she had been ill; who always finessed picking up the check; who lent money to his friends if he thought he could help them with a good investment. Cai called Huang his brother and considered him part of his extended family.

Just a few weeks earlier, Huang and Cai were catching up on the phone when the subject of the coronavirus came up. Huang, an optimist, reassured Cai that he didn’t think Covid-19 would ever be a crisis in this country, an opinion many of their colleagues shared. SARS, Ebola, MERS — none of them ever posed a public-health threat here. And soon it would be warm, when many viruses seemed to disappear. Cai was relieved to hear Huang’s assessment, but at the urging of his wife, he prepared for the possibility that the pandemic would reach the East Coast and do real harm. As early as late February — when people in New York were still flying around the globe, clutching poles on the subway, hugging friends hello — Cai made two trips to Costco in Brooklyn to buy provisions: frozen vegetables. Frozen fruit. Twenty pounds of rice. Protein shakes, just in case. Huang might have been sanguine, but his former supervisor at Mount Sinai, Paul Lee, a cardiologist, had posted warnings about what was to come. Many of Cai’s friends who were fellow Chinese immigrants were also stocking up. Like Cai, their family connections and exposure to Chinese media drove home how dangerous the disease was and how quickly it spread. If the virus became prevalent in New York, Cai knew what his family would do: They would lock down for two full months. No one would have to leave the house for anything.

The medical offices where Cai worked had put up a sign directing patients with a cough or fever to wear a mask, and to self-quarantine for two weeks if they had traveled recently to China. Cai never failed to wear a mask and gloves at the office. And yet he still did not see the virus as an imminent threat: He made plans to attend a medical conference and took the subway around the city to his various offices without wearing any protection. He and some of his Chinese-American friends, most of them first-generation, wore masks in public starting in January, reminded that it was a common-sense precaution by the devastating news from Wuhan. But then, in early February, a video ran on the local news showing a man violently attacking an Asian woman who was wearing a mask near a subway turnstile in downtown Manhattan. Cai — and many of his friends — stopped wearing them.

Now he felt he had let down his guard, and the worst had happened: He had tested positive. He felt real terror, as did the rest of his family. His father, who lives in Shanghai, reached out through various connections to doctors who had managed the illness there. His wife’s family was doing the same. Huang also was getting in touch with everyone he knew who he thought might be able to help. “I called up all my pulmonary friends, I.C.U. friends, infectious-disease friends — people I hadn’t spoken to in 10 years,” Huang says. He spoke to Chinese doctors from Shanghai who had been deployed to Wuhan, all of whom painted a dire picture of the damage the virus could do. He came to understand that many people recovered quickly on their own, even after a long illness; but he also knew that the disease could go from progressing slowly, seemingly harmlessly, to moving unfathomably quickly, even in otherwise-healthy people. The antibiotics Cai was given might help with a secondary infection, but they could not fight the virus. And there was no way to know what course Cai’s case would take.

Cai was anxious, and it seemed to him that the doctors were trying to keep him calm. They assured him that he was a young, healthy man. He remembers many telling him this would feel like a bad flu. But by March 6, his fifth day at the hospital, this no longer felt to Cai like any other flu. By then, he’d been moved to the third floor, into a negative-pressure isolation room — a room whose atmospheric pressure was so low, air outside flowed in, theoretically preventing any potentially contaminated air from flowing out. He had a pulse oximeter on his finger and could keep an eye on his own oxygen levels. He could see that they were unstable, sometimes dropping momentarily to a unnervingly low level of saturation — 85 percent — before shooting right back up. In a healthy individual, saturation levels typically remain above 95 percent. “I have difficulty breathing now, too much phlegm,” he wrote to Huang. Especially when he lay flat, his oxygen levels fell. “I need to get up and take a deep breath.” He felt as if he had been swimming under water, then surfacing to try to get relief — but his breaths were never deep enough to provide it.

The care he was getting was not always comforting. A nurse came in at some point to take his blood pressure and temperature, but his voice was fearful. “Turn your face away,” he told Cai. He placed a thermometer on the tray and told him to use it himself.

But his main anxiety was that his condition would deteriorate — that his lungs eventually would be so compromised that his oxygen levels would drop to a degree that endangered his life. The mechanism is both complicated and simple: If not enough oxygen reaches the organs, the intricate gears and motors of the human body start to fail. He frequently texted Huang. He was scared, he told him. He asked for reassurances that his friend would not let him die there. Of course not, Huang replied. Huang hoped he was telling the truth.

Cai’s world was reduced to the size and reach of his phone. To pass the time, he watched videos of his daughter over and over and stared at a picture of her in his arms. He would have yearned to video-chat with her but was afraid that it would be too upsetting for her — and maybe for him. They could never explain to a young toddler where he was and why he couldn’t come home, and so he and his wife decided not to tell her anything. He knew she had to be confused and suffering, and the thought of that was bound up with his own confusion and suffering.

The evening of Saturday, March 7, Cai was afraid to go to sleep. He was barely able to talk without collapsing into coughing fits. Earlier that day, he started receiving oxygen from a tank through a nasal cannula, a flexible tube that sits just inside the nostrils. But as he monitored his oxygen levels from his bed, he could see they were dropping. Even with the extra oxygen, his saturation level was as low as 88 while lying down, suggesting his lung functioning was weakening. He started to worry about acute respiratory distress syndrome. From there, he knew intubation could follow, a procedure that involves putting a tube down a patient’s throat and connecting the lungs to a ventilator. Cai knew that the I.C.U., where the ventilators were kept, was on a different floor; if he started to crash — if his vitals indicated that his organs were in imminent danger of starting to shut down — how were doctors going to intubate him and transport him to the ventilator in time to save his life? He’d seen patients die from respiratory failure in less than 10 minutes.

Cai’s family and friends were continuing to communicate with doctors in China and passing on their advice and suggestions. It was common practice during the Covid-19 outbreak there to give patients a second CT scan to provide a clearer view of the progression of lung damage; the so-called ground-glass opacities on the lungs common with Covid-19 could easily be missed on an X-ray or mistaken for something else.

Earlier that morning, Cai told the infectious-disease doctor on call that weekend that he wanted a second CT scan, a suggestion made by top doctors in China, who thought they could help his doctors in New Jersey understand the progression of the illness. The doctor seemed disinclined. They would determine treatment based on oxygen levels, which they were keeping an eye on. Simply moving Cai to the scanner risked exposing health care workers to the virus. Decontaminating the room that held the scanner would also take time, during which the scanner could not be used. (H.U.M.C. did not make some doctors involved in Cai’s care available for comment but responded in an email that they followed “C.D.C. and/or evidence-based protocols” that were “different from protocols physicians from China were advocating.”)

At around 10 a.m., Cai’s phone rang. His friend Huang wanted to talk with the infectious-disease doctor on call. He spoke to her on speakerphone so that Cai could hear. We are formally requesting a second CT scan, Huang told her. She explained, as Cai recalls, that it wasn’t necessary and most likely wouldn’t change the course of treatment, whatever the results. He pressed her on how confident she was about their treatment — and if so, on what basis? She had never treated a Covid-19 patient. How could she dismiss the collective wisdom of doctors in China who had seen thousands? Cai’s oxygen levels were not getting better, despite the antibiotics; Huang had the sense that the doctors at Hackensack did not fully appreciate how quickly patients could take a turn for the worse. The doctor said she would bring it up with Cai’s physicians.

Cai’s boss, Dr. George Hall, also made a call, not long after Huang spoke to the infectious-disease doctor on call. He spoke with another doctor on Cai’s caregiving team, a hospitalist named Danit Arad. Arad had agreed to share her phone number with Cai’s mother, who had passed it on to Hall. Hall, who is 64, studied at one of the most prestigious medical schools in China before immigrating to the United States in 1987 and opening up four medical centers throughout the city. A father figure to Cai, he, too, had been in touch with contacts in China, including a nephew in Yangjiang, who ran an infectious-disease hospital, to get insight into Cai’s case.

Hall explained to Arad that the Chinese National Health Commission had just published the seventh edition of guidelines on how to treat coronavirus. It was true that they were based more on clinical experience than on published studies, but he urged Arad to follow some of its protocols, which included prescribing two drugs that were commonly given to patients in China soon after they showed symptoms like shortness of breath: chloroquine, an antiviral drug once used to treat malaria, and Kaletra, another antiviral that had once been used to treat H.I.V.

At the time Hall and Arad were speaking, practitioners were struggling to gauge the utility of treating coronavirus patients with chloroquine or a derivative called hydroxychloroquine, which is used to treat autoimmune diseases like lupus. Since then, the picture has hardly become more clear. Two small studies from Marseille, France, published in March found that hydroxychloroquine and azythromycin, an antibiotic, yielded encouraging results in patients with advanced disease; but a close replication in Paris, published soon thereafter, found the drugs ineffectual. Yet another study, this one from China and published online March 30, found that patients who were mildly ill and took hydroxychloroquine fared better than the control group of mildly ill patients who did not receive the drug. When Trump called hydroxychloroquine “a game-changer” on March 19, many researchers considered his enthusiasm premature and possibly dangerous. Practitioners started stockpiling the drug, and doctors worried they would not to be able to provide it to autoimmune-disease patients who relied on it. On March 28, the F.D.A. approved the emergency use of chloroquine and hydroxychloroquine in treating patients with Covid-19, but European regulators are awaiting more data.

As for Kaletra, a study in March in The New England Journal of Medicine found it did not help patients suffering severe illness related to coronavirus, though researchers left open the possibility that it might be more effective earlier in the course of treatment.

Arad knew at the time that neither drug had been through extensive clinical trials or had F.D.A. approval. She listened patiently to Hall and expressed her concern that his suggestions did not conform to standard medical procedure or C.D.C. guidelines.

Hall understood the need for evidence-based medicine as well as she did, he told her. But this was life and death. Under those circumstances, sometimes you don’t wait for standard procedure, he said. If it came to it, he was sure Cai would assume the risk. Hall suggested that he could provide Arad with a full translation of the guidelines, which had not yet been published in English; Arad, Hall said, took him up on the offer.

Lying in bed that night, Cai feared that he would close his eyes and never wake up — that he would slip away, essentially drowning in his sleep. He was being given oxygen, but even still, he saw his numbers trending downward — in the 80s. Concerned, he messaged a WeChat group that included his father and a doctor his father knew in Shanghai, who had been advising that Cai be put on a high-flow nasal oxygen cannula, a device that allows for a more intensive and stable delivery of oxygen into the lungs. Cai requested that treatment, but the nurses on duty said that they didn’t have the clearance to make that decision. Cai called Hall to ask for help in getting a doctor’s attention. Hall contacted a prominent local doctor, Henry Chen, who oversaw a sprawling network of community-health doctors in New York, in the hope that he could get in touch with someone at the hospital. Chen says he was told that because he did not have admitting privileges, he would not be put through.

Cai had never felt more alone. He repeatedly called for the nurse, and when she arrived, he spoke as harshly as he ever had to a fellow medical professional. “I am not going to sleep until I see a respiratory therapist,” he told her. He wanted closer monitoring; and he wanted the expert care of someone who could provide a higher level of oxygen dispersal. He dropped Chen’s name, even though he knew the name likely meant nothing to the nurse; he reminded the nurse that he was a physician assistant and could judge for himself his risk. Finally, at around midnight, a respiratory therapist arrived with a Venturi mask, providing a treatment that was not as powerful as high-flow but that still provided higher concentrations of oxygen than Cai had been getting. The therapist also took blood for a test that would assess Cai’s lung functioning.

Once he received the oxygen treatment, Cai allowed himself to drift off, though his dreams kept him on high alert. Sometimes he dreamed that he’d woken up — it was morning, and he was alive, which he knew because he was staring at the clock on the hospital wall. Sometimes he dreamed that the Chinese experts were telling him that they had seen the results of the blood test and that the numbers were not good. All night, he drifted between consciousness and slumber, his very dreams trying to make out whether he was going to live or die.

The morning of Sunday, March 8, Cai woke up. He knew he was alive. There was the clock. There was his phone with the photos of his friends and family, the beeping machinery above his head. And yet he was still afraid. He prayed to God; he prayed to Buddha. He bargained: He would save so many lives if only his own could be spared. He would stop working so hard so he could be a better father to his daughter. He read over and over the cards that friends had sent him, tangible objects from the outside world that let him know that he had not been forgotten. He continued to text with Huang, who by then was having his own anxieties. He was worried about his friend, but also about the new cases cropping up every day. “The reality was setting in,” he says. “We will become Wuhan, Milan.”

Later that day, around noon, Hall sat down in the study of his Long Island home to translate the Chinese medical guidelines. It was no small task, but he was not aware of any other translation, and he believed it was important. “No one had any experience here,” he told me. He opened a Microsoft Word document and started translating: the symptoms, the signs of mild cases, severe cases, the course of the disease, the methods of oxygen delivery, the recommendations for follow-up. Just before midnight, having worked for close to 12 straight hours, he sent it off to Arad. His sense of urgency extended beyond Cai’s case. If a health care professional like Cai could not be saved, he explained, his patients — many of whom speak almost no English — would feel they had no possibility of surviving the virus, should they catch it and experience complications.

Around the same time that Hall sat down to work in Long Island, Cai, lying in bed in his room in Hackensack, was surprised when a technician arrived in full P.P.E. He was going to get his second CT scan. Two hours later, when Dr. Balani came to see him with the results, Cai listened to her speak with some fascination and a little bit of fear. She seemed different. She sounded scared to him, but like someone trying hard to sound confident; he had the impression she had rehearsed what she was going to say. She was speaking more quickly than usual. And she was telling him that now it was time to take more aggressive measures. Eventually, Cai saw the scan himself. Instead of just that one white spot on one lung — something with the look of a dandelion gone to seed — there were dozens. The onslaught of the virus could be described as a toxic lava flow of infection that ravages the alveoli, the fragile, thin-membraned air sacs where gases are exchanged in each breath. It looked as if close to 40 percent of Cai’s lungs had succumbed in just five days.

Balani said that they were going to try to put him on a drug called remdesivir. The drug, a descendant of a broad antiviral medication developed a decade ago, was tried in treatment of Ebola with little success. It was more effective in inhibiting MERS in infected monkeys, according to a study published in February in The Proceedings of the National Academy of Sciences. The medication, which fools the virus into incorporating a modified building block into its RNA that stops it from replicating, is still in clinical trials. Many doctors have been cautiously optimistic about some promising research, including one case study published in The New England Journal of Medicine in late February. The first patient in Washington State to be found to have Covid-19 was severely ill when Gilead Sciences, the pharmaceutical company that made the drug, provided him with remdesivir for compassionate use — an effort to use a promising drug for people who are gravely ill when no other treatments are available; the patient recovered. Now Cai’s health had deteriorated to the point at which the hospital could apply for remdesivir for compassionate use. Just a few weeks later, overwhelmed by international demand, Gilead announced that it would stop approving new requests for compassionate use but greatly expanded its clinical trials at various hospitals.

That day, Cai was given chloroquine and Ka­­letra; he was also put on high-flow oxygen, that high-concentration oxygen delivered through the nose. The method allows patients at risk of respiratory failure to stave off intubation and ventilators; but because the patient can breathe and talk through the mouth, the oxygen mixes with the virus in the patient’s nose and windpipe and, especially at highest pressures, can be breathed out into the air. Doctors in the United States have been forced to weigh a medical option that might spare a patient ventilation but could expose medical practitioners to far greater risk. Cai — as the first patient in a hospital that would be, weeks later, flooded by other patients in even more dire circum­stances, including their own staff members — received the treatment.

Cai was simultaneously reassured and distressed to see how grave the doctors suddenly looked, how quickly their stance toward his condition seemed to change. Later he learned that the results of that blood test was cause for real concern. They informed him that they had established a plan for getting him to the I.C.U. if need be. They also assigned him a dedicated critical-care nurse. He hated having those conversations about an intubation plan, hated that they had to talk about it as a realistic possibility. If the disease continued to progress at that swift rate in the next few days, he would almost certainly be intubated, his odds of recovery dropping precipitously.

Later that day, March 8, he asked the nurse to bring him some paper. He wanted to write a letter to his daughter about all the things he would want her to know about him if he did not survive this virus. Tearing up, he started to write. He said he was sorry he hadn’t been a better father. He wrote that he understood what it was like to grow up without a father present — his lived in Shanghai — and that he was sorry she would suffer the same fate. He wished he could play with her and her friends, pick her up at school, walk her down the aisle, solve her problems when she had any. He wanted her to know how much he loved her. He carefully folded the paper, slipped it into an envelope and placed it on the bedside tray where he took his meals. He did not have to tell the critical-care nurse what he was doing for her to figure it out. “I’m so sorry,” she told him.

Cai hoped the remdesivir might help. The hospital had made its own request. But he knew that getting approval for compassionate use — which required the manufacturer’s approval as well as F.D.A. approval — could take time, and he was worried that it would be too late. That day, Huang reached out to every Gilead representative he knew and called on all his doctor friends to do the same. His former supervising physician at Mount Sinai, Paul Lee, had already written an unsuccessful email to an associate director at the National Institute of Allergy and Infectious Diseases on Cai’s behalf to try to get him access to the drug. Huang posted on a large WeChat group for Chinese and Chinese-American cardiologists: “My name is Yili, great to meet everyone, unfortunately on this occasion,” he introduced himself. “I usually don’t post, but my good friend, only 32, health care provider, became this first case in New Jersey. Please help me with some inputs.” With Cai’s permission, he included a photo of Cai’s CT scan. He also forwarded the scans to another friend, Felix Yang, a cardiac electrophysiologist.

Yang replied with a question: Can I put this up on Twitter to show the severity of the disease? He had been frustrated in the previous days by other doctors’ refusal to take the possible spread of coronavirus seriously. At a minimum, Yang thought, the scan would show his colleagues just how quickly the disease could move. Yang made a quick video that showed the deterioration from one scan to the next and posted it on Twitter, asking people to help get in touch with Gilead, to help this patient with “sudden, rapidly progressing resp failure.”

Within 12 hours, half a million people had watched the video. C. Michael Gibson, the founder of the open-source textbook WikiDoc and a top cardiologist with nearly a half a million followers, helped by quickly retweeting Yang. Hundreds of doctors from around the world shared whatever they knew in comments; one doctor, an American who had been traveling back and forth to China, paged Yang at his hospital to share with him what she had learned. Yang believes hundreds of individuals tweeted at Gilead to try to get the company’s attention on Cai’s behalf.

Balani had already been laying the groundwork with Gilead to apply for remdesivir from the time Cai tested positive. His condition now made him eligible for compassionate use. Less than four hours after the image was first tweeted out, Gilead informed Cai’s doctors that the company was shipping the medicine out. Bill Pulte, a philanthropist active on Twitter, also posted a video of Cai that night that circulated widely; other media soon followed. (Gilead declined to provide details, saying it could not comment on individual compassionate-use cases.)

Around 3 a.m. on March 10, Balani arrived at the hospital. The medicine had come in, and she did not want to wait until the morning to administer it. With Balani in the room, a nurse woke Cai up so that he could sign the legal papers. Soon after, he was hooked up, intravenously, to the drug.

The next day Cai’s fever, which he’d had for at least nine days, finally broke. Even before he received the remdesivir, his oxygen levels started to stabilize. Now they indicated he was on the mend. He was still so weak in the following days that he could barely speak without exhaustion; every time he tried, he was racked by coughs. But the progress was steady, and about a week later, he was able to speak to his wife more easily, to start to feel confident walking around his room; he began to let himself picture himself back at home. His daughter would come running, he imagined, with his slippers when he walked through the door, as she always did. Now that he was recovering, his wife admitted to him that his daughter had been running to the door with his slippers for the last couple of weeks every time she heard a noise beyond it, then cried in disappointment when her father failed to arrive.

To date, there is no known cure for Covid-19. It is impossible to know what elements of Cai’s treatment — the high-flow oxygen, the medications, the passage of time, the sense of wraparound community support, the powerful injection of last-minute hope — helped pull him through. On March 21, Cai learned that he had now tested negative twice in a row for the virus. His lungs would need time to recover, but he was alive — and the virus was dead.

He left the hospital that day, nearly three weeks after he arrived. During that time, the number of known cases of Covid-19 in New Jersey had ballooned from one to 1,914. Twenty people in New Jersey had died. And in the weeks to come, the number infected would rise to more than 29,000, and members of the medical staff, now treating hundreds of sick patients, would fall ill themselves. “I intubated my colleague today,” tweeted David Zodda, a Hackensack emergency-room doctor, on March 27: “a young, healthy E.R. doc like me.”

The next day, at home, Cai would tweet out his gratitude to the staff of the hospital, thanking many of them by name, including Balani and Arad, “for saving my life.”

Before he left the hospital, he put on a soft gray hoodie, sweatpants and clean socks, all of which his mother had left for him. He put on a mask. As he walked out of the room that had been like a prison, he looked back at the bedside tray where he took his meals, where he had placed that letter to his daughter in an envelope. He left it behind. Someone would throw it away and clean the room, and another patient would take his place.