After more than three weeks on the ventilator, after battling weakness and delirium on the general medical floor and a stay at the long-term rehab hospital where she rebuilt the strength to walk again, my patient had made it home. The dark shadows beneath her eyes were fading. Her skin was tanned. The persistent shortness of breath had finally abated, and she had recently run four miles to commemorate four months since she was diagnosed with Covid-19.
Four months. I closed my eyes and found myself once again in those early days of the pandemic, clustered outside her room with a team of doctors and nurses. Nearly two weeks in, she still needed high levels of support from the ventilator and we were starting to talk about the impossible decisions we might face if her lungs never improved. But we waited, because this was a new virus and we did not know its course, and because we had the resources to do so. And now there she was, in clinic - months later, doing far better than I would have predicted.
I have been surprised by similar recoveries in the past weeks. People we thought could die, or at least end up significantly impaired, have made it home. But there is something troubling about this too. It is clear to me that there was no one specific therapy that determined the outcomes of our sickest coronavirus patients in the intensive care unit (ICU).
On the contrary. While even the best possible treatment couldn't save everyone, those who survived did so because of meticulous critical care, which requires a combination of resources and competency that is available only to a minority of hospitals in this country. And now, even as we race towards the hope of a magic bullet for this virus, we must openly acknowledge that disparity - and work to address it.
Since the beginning of this crisis, conversations about death from Covid-19 have revolved around patient characteristics - men are more likely to die than women, as are people who are older or obese, or those with co-morbidities. But we now know that the hospital matters too.
In a large study recently published in the journal Jama Internal Medicine, a team of researchers examined hospital mortality rates in over 2,200 critically ill coronavirus patients in 65 hospitals throughout the country. Their findings? Patients admitted to hospitals with fewer than 50 ICU beds - smaller hospitals - were more than three times more likely to die than those admitted to larger hospitals.
Though they were not able to study factors like staffing and hospital strain, these likely contributed. In fact, a recent investigative piece in The Times examined mortality data for hospitals in New York City - and found that at the peak of the pandemic, patients at some community hospitals (with lower staffing and worse equipment) were three times more likely to die than patients in medical centres in the wealthiest areas.
Knowing first-hand what it requires to keep critically ill Covid-19 patients alive, this does not surprise me. Though the public has largely focused on new treatments - with excitement and controversy swirling around remdesivir and dexamethasone and convalescent plasma - none of these is any use without the people and systems to deliver critical care, a laborious and resource-intensive process.
In the ICU, we must interpret and react to each indicator. Our nurses are frequently at the bedside, attuned to the most minute change. We make constant small tweaks to the ventilator and to our medications to support blood pressure. Though it looks passive in a way - a comatose patient in a bed - and is not at all glamorous, critical care is an immensely active process.
We are all familiar with the images of Covid-19 patients lying on their chests, and we know that prone positioning saves lives. But the simple act of turning a critically ill patient is physically strenuous and, if done hastily, treacherous. Breathing tubes and intravenous lines can become dislodged. The head must be repositioned every two hours. At my hospital, during the height of the pandemic, we formed a dedicated "prone team" of respiratory and physical therapists who were available 24 hours a day. This spared the bedside nurses and kept patients as safe as possible. Even so, breathing tubes became kinked, and on at least one occasion, we had to urgently replace a breathing tube - a risky procedure.
This is why in some hospitals, prone positioning might not have been offered at all. Indeed, the Jama study found rates of prone positioning to range from just under 5 per cent at one hospital to nearly 80 per cent at another. Patients would have suffered as a result.
Anyone who has cared for a coronavirus patient knows how quickly they can crash. Thick mucus blocks airways and endotracheal tubes. Oxygen levels plummet. Heart rhythms go haywire. As a doctor, I'll admit that we are rarely the first to intervene in these moments of crisis. Instead, we rely on nurses and respiratory therapists. More times than I would like to count, I have watched with gratitude as their interventions - suctioning, repositioning a breathing tube, increasing the dose of medications to raise blood pressure - avert certain disaster. It is humbling to realise that had our nurses been spread too thin, these relatively small events would have turned catastrophic.
Perhaps most importantly, because we had the resources to do so, we were able to give our patients time for their lungs to recover. I think of one man so sick that he was dependent not just on the ventilator but also on a heart-lung bypass machine. These machines, and the staff who know how to manage them, are a truly limited resource. Large academic centres have five of them, maybe 10. Some community hospitals do not have any.
This man had been on the machine for weeks, encountering one complication after another. He bled, we stopped blood thinners, and then surgeons had to rush in overnight to replace a part of the machine when it clotted off. There seemed to be no way out. But then, even as we prepared to say enough, his lungs started to improve. I remember standing outside his room one overnight, amazed, as his stiff lungs began to work with the ventilator once again.
He has now left the hospital. On the night of his return home, his son sent me a note: "Finally family is back, and that is the best feeling of this world." You might say he was lucky. But so were we. He was able to return home not because of any 11th-hour save on our part, but because we were able to watch and wait. And we could afford to do so only because here in Boston, we were busy but never underwater.
Of course, we made mistakes, miscalculations and errors in judgment as we learnt about this new disease. But we were in a privileged position. It could have been far worse. And as the pandemic tears through rural areas of the country with even less access to resource-rich hospitals, I am worried that the inequities of this virus will only become more entrenched.
Just as we devote resources to finding a vaccine, we must also devote resources to helping hospitals deliver high-quality critical care. Maybe that will mean better allocating the resources we do have through a more robust, coordinated system of hospital-to-hospital patient transfers within each region.
Maybe it means creating something akin to dedicated coronavirus centres of excellence throughout the country, with certain core competencies.
Maybe it will mean expanding the reach of experienced critical care hospitals through telehealth. This will not be easy. But as this virus will be with us for the foreseeable future, it is our duty to try.
As the video visit with my patient ended that day, she reminded me that she had been transferred to us from a small hospital in the western part of our state. "If I hadn't been transferred, I would have died," she said.
I paused, reflecting on that. What had we done for her, really? We had never enrolled her in a clinical trial. There was no mystery diagnosis to be solved, no high-risk procedure performed. We simply did our best to minimise damage to her lungs and keep her other organs functioning while we waited - which makes it even more painful to admit that she might be right.
• Daniela J. Lamas is a critical care doctor at Brigham and Women's Hospital in Boston.