The U.S. is facing its worst flu season in 25 years, according to the Centers for Disease Control and Prevention, which estimates that as of Jan. 3 there have been at least 15 million illnesses, 180,000 hospitalizations, and 7,400 deaths.

Part of the problem may be a new virus strain called subclade K, which has “antigenic differences” from strains used in this year’s vaccine, says Yonatan Grad, professor of immunology and infectious diseases at the Harvard T.H. Chan School of Public Health and the director of the Center for Communicable Disease Dynamics. In the following interview edited for clarity and length, Grad explains why some flu seasons are nastier than others and what we can do to stay safe. 

Why are some flu seasons worse than others? 

The variation in incidence is a function of a couple of things. The first is how much of the population is susceptible to the circulating strain of influenza virus. Immune protection comes from both past infection and vaccination, but both can wane with time and offer less protection as the virus evolves antigenically.

The second is people’s behavior, how much they’re interacting in ways that allow the virus to spread. If you look back at the year of COVID-19 lockdowns, when there was little interaction and little opportunity for the virus to spread, there wasn’t much influenza at all.

What should we know about the strain we’re seeing this year? 

The two main classes of influenza are influenza A and B. There used to be two strains of influenza B that circulated each year, one from a lineage called Victoria, another from a lineage called Yamagata. Yamagata seems to have gone extinct during the pandemic.

Within influenza A, H3N2 and H1N1 are the two subtypes that have been circulating for nearly 50 years.

Over the past 100 years or so, a few different subtypes of influenza A have circulated in people. The 1918 flu pandemic was the H1N1 subtype. It circulated seasonally until 1957, when there was an H2N2 pandemic, with H2N2 replacing H1N1 as the seasonal influenza virus. In 1968, an H3N2 pandemic then resulted in that subtype becoming the seasonal flu, in turn replacing H2N2, which we have not seen since. In 1977, H1N1 reappeared, but rather than taking over, it has been co-circulating with H3N2 ever since.

Usually, one influenza A subtype dominates in a season, with influenza B circulating as well. This year, H3N2 is dominating. Flu B tends to be milder, though it can still cause severe disease. And while H3N2 is thought to be more severe, we don’t really know why.

I went through the history of flu A because there’s some indication that the first flu you’re exposed to may influence your flu responses all your life. So one hypothesis for why H3N2 appears more severe is that people born before 1968, when H3N2 started circulating, may on average have less protection against that strain of that subtype.

When we hear that this year’s flu vaccine didn’t match up well with the composition of the flu that’s circulating this year, what does that actually mean? 

The H3N2 strain used in the vaccine, a subclade J.2 virus, appears antigenically distant from the circulating H3N2, which is from subclade K, raising concern that the antibodies elicited by the vaccine might not offer as much protection against what’s circulating. 

But a very recent paper shows that the vaccine is not as much of a miss as has been described: It seems people do develop responses to subclade K in response to vaccination. Moreover, the early estimate for flu vaccine effectiveness this season is about where we see it for H3N2 strains regularly. 

How do researchers decide what strains of the flu to use in the year’s vaccine? 

The decision about which strains to include in the vaccine takes place many months before the flu season; for the northern hemisphere, it’s usually in February. The reason it’s so early is that it takes a long time to grow the year’s stockpile of vaccines in chicken eggs, which is how most of our seasonal influenza vaccines are manufactured. Subclade K emerged in the spring of 2025, after the strains for the season’s vaccines were set.

This raises the question of whether technologies that can make influenza vaccines faster could help us avoid this kind of situation by allowing vaccine strain selection later in the year. MRNA vaccines for influenza might have been one potential solution, but federal support for mRNA vaccine research has been cut, so we likely won’t know any time soon.

Is it too late to get a flu shot this year? 

No, it’s not too late, especially as we expect the flu to stick around through the next few months. It’s worth noting that it takes about two weeks after vaccination for the protection to kick in, and then protection wanes over the course of the following four to six months or so. We usually recommend getting vaccinated around Halloween to maximize protection through the peak of the flu season, which is usually mid-winter.

To reduce your chances of getting the flu, I also recommend the general precautions for protection against respiratory viruses that many have become familiar with from COVID-19: masking, hand hygiene, and avoiding crowded, poorly ventilated spaces. 

There was a small uptick in COVID in early fall this year, but it hasn’t surged as much this winter. What do you make of that trend? 

It’s likely due to the level of population protection against the circulating strains of SARS-CoV-2, reflecting past infections and vaccination, much like for the flu. I would expect that as population immunity wanes and new variants of SARS-CoV-2 emerge, we’ll be at risk for another wave of COVID-19, and we’ll move into a pattern of periodic surges.