An axiom in medicine is that good judgment depends on experience, and experience depends on bad judgment. Basically, one way doctors refine their care is through a learning curve resulting from inexperience and lack of judgment.

A fitting example during the COVID-19 pandemic was the worldwide experience with ventilators, which are used to support lung function in patients with serious COVID-19 pneumonia.

Viruses or bacteria that invade the lungs and cause pneumonia impair the body’s ability to take in oxygen. When this becomes severe, it is known as hypoxemia (this is what those finger oximeters measure). In advanced hypoxemia, patients experience shortness of breath. If it worsens they will be unable to breathe. At that point, doctors will insert a tube in the windpipe and attach the tube to a ventilator, which takes over a patient’s breathing until the lungs heal.

In March 2020, with COVID-19 ravaging Italy, western Europe and New York City, deaths from COVID-19 pneumonia began increasing exponentially. Body bags were stacking up in New York hospital morgues, and health officials nationally became concerned about an anticipated shortage of ventilators, which became a political issue with mayors and governors jousting to acquire more ventilators.

“This is a story about doing the impossible,” then-New York Mayor Bill de Blasio boasted at the height of the crisis. “We’d never made a ventilator before — and so we made thousands. We learned it would take a year — and so we did it in a month.”

General Motors and Ford also started a World War II-type manufacturing campaign to deliver new supplies of the devices suddenly deemed critical to the COVID-19 effort.

One problem: No one had any idea how many ventilators would be necessary. Estimates were all over the map.

“There is a broad range of estimates of the number of ventilators we will need to care for U.S. patients with COVID-19, from several hundred thousand to as many as a million,” several physicians and public health experts wrote in the New England Journal of Medicine in April 2020. “The national strategic reserve of ventilators is small and far from sufficient for the projected gap. No matter which estimate we use, there are not enough ventilators for patients with COVID-19 in the upcoming months.”

For a short time, there were temporary ventilator shortages in certain regions, but the estimates from the New England Journal of Medicine turned out to be too high. The situation soon changed dramatically with two unexpected developments.

First, doctors noticed that COVID-19 patients did not always improve as anticipated when placed on ventilators. Both in the U.S. and Europe, these patients were actually dying at a higher rate than expected. Ventilators may have been applied too quickly, and it is now believed they actually caused some of those deaths, either as a result of the aggravating damage done by forcing air into the lungs or from other microorganisms entering the lungs more easily down breathing tubes.

Ventilators undoubtedly saved many patients, and many of those who died would have died anyway, but it is likely that of the nearly 2.5 million worldwide COVID-19 deaths in the first year of the pandemic, several thousand were directly the result of ventilator complications.

The second development was that, for poorly understood reasons, some patients with COVID-19 pneumonia and severe hypoxemia did not experience shortness of breath in the same manner as patients with other diseases. Doctors were astonished to see patients with otherwise life-threatening low oxygen levels walking around their hospital rooms or talking on their cellphones.

This gave physicians latitude to use alternative treatments to inserting breathing tubes and ventilators. Some medical centers in Europe and North America (UChicago Medicine was one of the first) successfully began using high-flow nasal cannula systems, turning patients on their stomachs, and employing masks similar to those used in sleep apnea, often successfully raising oxygen levels to safe values.

While ventilators are still essential for some COVID-19 patients, the pattern of use has changed dramatically. In the early days of the pandemic, data from the Centers for Disease Control and Prevention shows that 25% of all hospitalized COVID-19 patients were placed on ventilators. Today that figure is 5%. There are many causes for the decrease: vaccination, a less severe virus and (with the early pandemic deaths of the most susceptible) fewer vulnerable patients. But there is no question that a learning curve and change in practice was part of that equation.

Recently in New York City nearly 3,000 ventilators, which originally cost taxpayers $12 million, were sold for less than $25,000. A Long Island junk dealer used 28 trucks to haul off many of the ventilators Mayor de Blasio touted three years ago.

In medicine, what seems true today might not be true tomorrow, and no one knows whether it will be true the day after tomorrow. In retrospect, the 2020 controversy over having enough ventilators proved mostly unwarranted and resulted in unnecessary expense. But we couldn’t know that at the time — we had to err on the side of having too many rather than too few ventilators.

Still, the virus threw us surprise curveballs that resulted in dramatic changes in the treatment for hypoxemia in severe COVID-19 pneumonia. Planning for future pandemics is not as easy as it is sometimes portrayed. The take-away for future pandemic strategists is to factor in a good degree of flexibility — because nature never runs short on curveballs.

Dr. Cory Franklin is a retired intensive care physician. Dr. Robert Weinstein is an infectious disease specialist at Rush University Medical Center