Many public health experts have warned that the coronavirus will put a tremendous strain on the nation’s intensive-care units. Now, new estimates show that the country’s ICU capacity — as measured by the number of beds per capita — varies greatly, with as much as a tenfold difference from region to region.

But Adam Sacarny of the Columbia University Mailman School of Public Health, who conducted the research, and other experts caution that the number of ICU beds is only one of many factors used to gauge a region’s health-care infrastructure and the readiness of its hospitals to respond to the coronavirus.

Let’s start with the data (you can look up ICU capacity in your city in the table at the bottom of this story). The map above divides the country into “hospital referral regions” — basically regional health-care markets as defined by medical referral patterns for major procedures like heart and brain surgery. Each region has several hospitals, some big and some small, and nearly all of them submit regular cost report data to Medicare. These reports typically include ICU bed counts.

Sacarny analyzed data from 2016 to 2018, the most recent years available, and found that hospitals reported approximately 93,000 ICU beds in the United States. That figure does not include neonatal or pediatric intensive-care beds, which typically would not be available to the adults who make up the bulk of hospitalized coronavirus patients. The number also excludes beds at government-run facilities such as Veterans Affairs and military hospitals. Researchers have previously estimated that those types of facilities account for about 2 percent of all ICU beds.

Areas with greater populations naturally have more ICU beds than less-populated areas. For a better sense of capacity, The Washington Post calculated the number of beds per capita in each hospital referral region and expressed the findings by the number of beds per 10,000 people older than 15.

The United States has about 3.6 ICU beds per 10,000 people aged 16 and older; that number is quite a bit higher than in most other countries. When it comes to ICU capacity, “we are actually quite well bedded,” said Jeremy M. Kahn, a professor of critical-care medicine at the University of Pittsburgh who has published several studies on the topic.

Kahn notes that we don’t have one national health-care market but rather hundreds of regional ones. And the data shows huge variations in ICU capacity between markets. The Fort Collins, Colo., hospital referral region, for instance, has just over one ICU bed for every 10,000 people. The Slidell, La., region, on the other hand, has nearly 11 beds per 10,000 people.

“This level of variation might seem weird,” Sacarny said, “but it’s pretty standard to see big differences in health-care use across areas.” He notes that the map shows patterns similar to those seen in measures of health-care utilization, such as hospital admissions and Medicare reimbursements per enrollee.

“All else being equal, regions with more ICU beds per capita will be better equipped to handle a surge in demand” related to the coronavirus, Kahn said. But he stressed that more beds aren’t always better. Studies have shown that an excess of beds, and the pricey infrastructure and staff surrounding them, would either sit vacant and waste money, or entice hospitals to direct noncritical patients to the ICU even if they don’t need it.

Sacarny agrees that more ICU beds don’t necessarily translate into better care. “Places that provide more health care do not have better health outcomes, on average,” he said. “They are providing a lot of care but maybe not the right kind of care.”

“If you have a lot of ICU beds, you end up filling them with lower-severity cases,” Kahn said. Indeed, one crucial piece missing from the data is the number of beds currently occupied. Nationwide, one recent estimate pegged the ICU occupancy rate around 62 percent. But that figure will vary wildly by region.

The nation’s overall capacity suggests that a fairly high number of existing ICU patients may have milder conditions that would allow them to be moved to regular beds, freeing up space for coronavirus patients. “We have the ability to take a lot of patients in ICUs and care for them in other settings,” Kahn said.

Nevertheless, Kahn said, he is “extremely” worried that the virus will still overrun ICU capacity in some areas of the country. In a 2010 paper he co-wrote, he estimated that if 0.02 percent of the country were to suddenly require ICU care as a result of a pandemic, ICUs would be overrun in roughly 17 percent of the nation. If the share of critically ill rose to 0.05 percent, capacity would be exceeded in 90 percent of the United States.

So if more beds aren’t necessarily the answer, what is? Kahn recommends against building field-hospital-style ICU facilities in places like fairgrounds and universities, which is what’s happening in parts of Italy. “Why would we take the most severe patients and move them into these ad hoc facilities?” he asked. Hospitals should focus instead on which of their current ICU patients can be moved to regular beds, freeing up more ICU space for those who really need it, he said.

Long term, he says, we can prepare for the next epidemic, in part by collecting better data on our ICU capacity. Though the data presented here is the most recent available, it is still several years old.

He also says our current health-care system, which incentivizes competition between hospitals to attract patients, isn’t a great way to allocate critical resources in the time of a pandemic.

“Philadelphia has five Level 1 trauma centers,” Kahn said, by way of an example. “No city needs that many. We lack the capacity for regional coordination among hospitals to get patients out of hospitals” and into other facilities, like skilled nursing facilities, that can care for them.

Though regions with more ICU beds may end up faring better during the pandemic, in other words, their concentration of resources in ICU capacity may hamper their ability to save lives at other times.