SEATTLE — Half of new coronavirus infections in Washington are now occurring in people under the age of 40, a marked shift from earlier in the epidemic when more than two-thirds of those testing positive were in older age groups.

A new analysis finds that by early May, 39% of confirmed cases statewide were among people age 20 to 39, while those 19 and younger accounted for 11%.

The trend is concerning and should be kept in mind as more counties begin to ease restrictions and reopen businesses, said Seattle epidemiologist Judith Malmgren, who is affiliated with the University of Washington and is lead author of the report.

Though younger people are less likely to die or be hospitalized with the virus, they can still suffer serious illness —as underscored by recent reports of a rare, life-threatening inflammatory syndrome in children. And even if younger people don’t get sick, they can pass the virus on to others who are more vulnerable.

“Younger people are the most likely to be socially active, they are the most likely to work in essential professions and have more contact with the public,” Malmgren said.

Malmgren and her colleagues don’t attempt to tease out all the reasons for the shift in age distribution, but one factor is obvious, she said. “Being a Seattleite, just walking around and seeing so many young people congregating without wearing masks, I thought: ‘This is interesting.’”

Malmgren was also intrigued by the epidemic’s overall trajectory in the state, with cases peaking on March 22, declining for a few weeks, then hitting a plateau with an average of about 200 cases a day for more than a month. She decided to dig more deeply into those numbers to see what patterns might emerge.

It’s the kind of analysis that will become increasingly important as society reopens and health officials will need to quickly identify new cases, hot spots, and groups at high risk in order to keep the virus from exploding again, Malmgren said.

“While disease modeling was helpful at the beginning of the epidemic, a more granular approach to profiling case characteristics is paramount to evaluate where the disease is now in the population and to plan accordingly,” she said.

The report is posted on medRvix, a preprint server for health sciences research. It has not been peer-reviewed, but is based on publicly available data from the Washington Department of Health.

Washington State Health Officer Dr. Kathy Lofy confirmed the trend.

“I believe this means that older individuals who are at higher risk of infection are doing a great job social distancing and protecting themselves,” Lofy said in a statement. The age shift also reflects the success of efforts to fight outbreaks in long-term care facilities, and an increase in outbreaks flaring at work sites across the state, she added.

Uneven testing through the course of the epidemic complicates the picture as well. Early on, testing was limited to people with known risk factors, including being over the age of 60. Since then, capacity has expanded greatly. But Malmgren and her colleagues found the increase in cases among younger people was not a result of differential testing levels.

While some have argued that allowing more young people to become infected will lead to herd immunity — a situation where enough of the population is immune to stop the disease from spreading — Malmgren says that’s a terrible idea.

Herd immunity is usually achieved through vaccination. With no vaccine, at least 70% of the population would need to be infected —a level of contagion that would overwhelm hospitals and result in high mortality. It’s also not clear how long immunity to the novel coronavirus will last.

According to estimates based on modeling and results from a surveillance testing program, fewer than 2% of people in King County have been infected so far.

Dr. Jeff Duchin, health officer for Public Health for Seattle and King County, said about a third of new infections in the county are still occurring among older people in care facilities. But it makes sense to expect the age distribution to begin skewing younger as people begin to move around more and return to work.

“The pattern in who’s getting infected over time reflects people’s behavior,” he said. “If we’re seeing a drop in age, that means younger people are doing things that place them at a higher risk of transmission compared to people who are older who might be staying home more reliably.”

The age shift is reflected in the rate of hospitalization for COVID-19, which has dropped more precipitously than the number of confirmed cases. Even at the height of the epidemic, people under 40 accounted for fewer than 15% of hospitalizations. According to DOH data from May 10, 12.3% of the nearly 200 people hospitalized for coronavirus-like symptoms statewide were under the age of 40.

Since young people are less likely to seek care or be hospitalized, the shift in infection demographics could make it harder to gauge the true incidence of the disease, the report says.

According to Washington Department of Health data, no children have died of COVID-19 in the state, and only 1% of deaths have been in people aged 20 to 39; 90% of those who have died were 60 or older.

The absolute numbers of new infections statewide are declining in every age category except one: People 19 and younger. Though still a small proportion of total cases, the youngest sector recently overtook the oldest, with new cases among children and older teenagers now exceeding those among people 80 or older.

“I think this is a very interesting observation, and we definitely need to follow up on it,” said Dr. Janet Englund, an infectious disease specialist at Seattle Children’s. Very few kids are currently hospitalized with the virus, and only recently has testing been widely available for youngsters, she added.

Malmgren and her co-authors say health departments, especially in counties with young populations, might need to develop new messaging that makes it clear anyone can become infected.

“The purpose of this analysis is to alert the public and government officials and institutions that this is a higher risk group for transmission and infection,” she said, “despite the popular misconception that children, teenagers and young people are not at risk.”