Officials in the Biden administration responsible for communicating with the public have repeatedly said this has become a “pandemic of the unvaccinated.” Such words may not ring true to people who got their shots and still became sick.

A growing number of studies suggest that coronavirus vaccines continue to provide strong protection against severe disease and hospitalization, but their ability to prevent mild illness is less robust today than the original clinical trial studies demonstrated a year ago.

The virus has mutated. The delta variant is rampant. Some elements of immunity may be gradually waning for people who got shots many months ago. New research studies in the United States, Israel, Britain and Qatar have shown a partial erosion in the effectiveness of vaccines against mild to moderate infections. Immunocompromised people are winding up hospitalized despite being vaccinated.

People should expect to roll up their sleeves again. Boosters are coming. The Biden administration is finalizing plans to authorize boosters for people who are eight months past their last shot.

The vaccines remain remarkably effective at the crucial task of keeping vaccinated people with breakthrough infections out of the hospital or the morgue, according to infectious-disease experts. The vaccines are clearly the best weapon against the virus.

But vaccine proponents say the public should be realistic about what vaccines can do. The shots don’t create an impermeable protective sheath. Instead, the vaccines prime multiple elements of the immune system to recognize and respond to coronavirus infections and rapidly clear the virus. Getting a little bit sick post-vaccination doesn’t mean the shots didn’t do their job.

“We could have done a much better job at setting realistic expectations for this vaccine,” said Paul A. Offit, a pediatrician and vaccine expert at Children’s Hospital of Philadelphia. “And I think that’s hurt us. Because I think people get disappointed. They think the vaccine isn’t working.”

This reframing of what success looks like is a daunting messaging task for public health officials who have repeatedly described the coronavirus in alarming terms. They are now emphasizing protection against severe disease and death rather than blanket protection against infection.

No one knows how many breakthrough cases there are, nor their degree of severity. While some breakthrough cases will be asymptomatic or only mildly symptomatic, others could flatten a person for days with flu-like symptoms but still fall short of requiring advanced medical treatment.

The Centers for Disease Control and Prevention stopped documenting mild to moderate cases May 1, focusing instead on hospitalizations and deaths. An internal CDC document first published July 29 by The Washington Post offered a national estimate of 35,000 breakthrough infections a week but did not provide documentation for the estimate.

In recent weeks, federal officials have backed away from their earlier assurances that almost all hospitalizations from the coronavirus are among unvaccinated people. CDC Director Rochelle Walensky has twice been asked this month by reporters if she stood by an earlier statement that 97 percent of people hospitalized by the virus are unvaccinated, and both times she acknowledged the data was out of date. Her agency has yet to produce new estimates.

What the country is left with is a series of muddled messages from federal health officials and confusing research studies about breakthrough infections. The nation’s 169 million fully vaccinated people are forced to navigate the summer surge of cases, fueled by the delta variant, with limited information about protection against the virus and the durability of their immunity.

When the first clinical trial data on the Pfizer-BioNTech and Moderna vaccines were published nearly a year ago, they showed stunning levels of vaccine efficacy against symptomatic illness — about 95 percent. But effectiveness, when measured in clinical trials, may be influenced by highly controlled conditions and may not have captured asymptomatic or mild cases.

It turns out the vaccines may provide somewhat less protection against mild to moderate sickness than many vaccinated people presumed when they first got their shots. During the initial months of vaccinations, many news reports, including in The Post, focused on the rarity of breakthrough infections. But the data on breakthroughs was sketchy, and it is likely that people with mild cases did not get tested for the coronavirus and instead assumed they had a cold.

Meanwhile, the delta variant has become dominant — primarily by taking advantage of the large fraction of the population that has resisted getting vaccinated. In the Deep South and other pockets of the country, the virus is spreading at alarming rates and filling intensive care units again. All that virus in circulation, combined with less guarded behavior, inevitably has led to a rise in breakthrough infections among the vaccinated.

“Delta has high prevalence in the U.S. and has moderate impact on vaccine effectiveness. The vaccines are likely to still provide protection against severe disease,” Heather Scobie, a member of the CDC’s Covid-19 Response Team, said at an Advisory Committee on Immunization Practices meeting Friday. She said more data is needed, especially for the Johnson & Johnson vaccine and “the potential for waning immunity.”

Scobie summarized reports of how the vaccines fared against delta in Canada, Qatar, Israel and England plus Scotland. In those nations, vaccine effectiveness at preventing infection ranged from about 40 percent up to nearly 80 percent. Effectiveness against severe illness was consistently higher, in the ballpark of 90 percent for vaccine recipients overall. Differences in methodology could explain that variation in study results, Scobie said.

Offit and some other prominent experts feel the CDC and news media have botched some of the messaging on breakthrough infections. They cite the Provincetown, Mass., outbreak in July that led to more than 300 breakthrough infections and spurred the CDC to revise its mask guidance for vaccinated people. The significant fact, they argue, was that only four of the vaccinated people who became infected wound up in a hospital.

That’s a sign of vaccine success, not failure, said Amesh Adalja, a senior scholar at the Johns Hopkins Center for Health Security.

“Breakthrough infections were expected. Because vaccines are not bug zappers. They’re not force fields,” Adalja said. “The goal of the vaccines was to make breakthrough infections clinically inconsequential, and that’s what they’re doing.”

He’s vaccinated, and unworried.

“I suspect that maybe someday I’ll get a breakthrough infection. But I’m okay with that, because I’m vaccinated, and I know the breakthrough infection will be mild,” he said. “This is an endemic respiratory virus. Everyone’s going to encounter it.”

Offit goes further in reframing the discussion. He thinks the use of the word “breakthrough” by the CDC and the media to describe mild or asymptomatic infections is misleading.

“The word ‘breakthrough’ makes it sound like the vaccine didn’t work. It was the wrong word to use,” Offit said.

Vaccine effectiveness compares risk for people who are vaccinated with risk for people who aren’t. Researchers use this to assess the reduced likelihood of infection, sickness, hospitalization or death in an immunized person vs. the likelihood of those outcomes in a vulnerable person, in real-world settings.

The higher this value, the better the vaccine is working relative to not having it. An effectiveness of zero percent against infection would mean a vaccinated person has the same chance of infection as an unvaccinated person. An effectiveness of 95 percent means the person who’s vaccinated has 5 percent of the risk an unvaccinated person does at catching the coronavirus.

Because effectiveness must account for real-world variables, it’s not carved in stone. It’s a moving target. Many factors, some obvious and others less so, are distilled into this number.

One factor is the virus itself. It is steadily mutating and through natural evolution finding ways to become more “fit” as a pathogen. Virtually all new infections in the United States now involve the delta variant, the fittest version of the coronavirus yet to emerge. Scientists say it is at least twice as transmissible as the early strains of the virus. Whether it causes more severe illness is not well-established.

Vaccine effectiveness can also be altered by waning individual immunity. Although it is possible to get an antibody test, a person’s immune system is multilayered, and protection against a specific pathogen is not easily measured. Some people are immunocompromised and when vaccinated may not mount a significant immune response. The CDC said Friday that 40 to 44 percent of patients hospitalized with breakthrough infections have compromised immune systems. The CDC recommends an additional shot for people who are immunocompromised.

Another major influence on vaccine effectiveness is the amount of community transmission of the virus. That can be affected by changes in behavior, such as engaging in riskier activities. If a vaccinated person is in a community where lots of people are not wearing masks and are breathing out the coronavirus in crowded indoor settings — offering more chances for the virus to spread — effectiveness against infection can be lower than it is for a person who lives where people are more cautious and the virus is rarer.

“We need to consider: What does the local level of transmission look like?” said Robert A. Bednarczyk, an epidemiologist at Emory University’s Rollins School of Public Health.

Different researchers have reported different levels of vaccine effectiveness. The differing estimates may be attributable in part to idiosyncrasies of national vaccine programs, the CDC’s Scobie said. Countries employ different lags between doses, and different brands.

One study that has not gone through peer review, compiling data from the Mayo Clinic Health System in multiple states, found that in July, vaccine effectiveness against infection for the Pfizer-BioNTech vaccine dropped to 42 percent. The Moderna vaccine was higher, at 76 percent. Both vaccines robustly protected against hospitalization, at 75 percent or above in July.

Another recent study, by researchers in Qatar and also not yet peer-reviewed, found effectiveness against infection was 54 percent for Pfizer and 85 percent for Moderna. Protection against hospitalization and death were much higher. A paper in the New England Journal of Medicine by British researchers found vaccine effectiveness for two Pfizer doses was 88 percent against symptomatic delta infection.

Some of the best data has come from Israel, which has a high level of vaccination. New data from Israel’s health ministry, released Monday, shows a significant drop in protection against hospitalization among people age 65 and older who were vaccinated in January and February.

The Israeli data has been cited by U.S. officials as one reason the Biden administration is preparing to recommend boosters for people eight months after being fully vaccinated. Israel has already started beginning giving booster shots to people older than 50.

Laurie McGinley and Tyler Pager contributed to this report.