By Helen Branswell

It can be jarring to see placards advertising "Flu Shots Today" in late July or early August in 80-degree weather.

But those signs may be more than just an unwelcome reminder that summer's days are numbered. Mounting scientific evidence is raising questions about whether vaccinating people that early may actually be undermining the effectiveness of the nation's massive flu vaccination program.

Studies from the US and Europe have detected a steady decline in vaccine protection in the months after vaccination. The most recent research , published just last month by scientists from the Centers for Disease Control and Prevention, showed that the vaccine's effectiveness was reduced by more than half for a couple of strains of flu, and had diminished almost entirely for another by five or six months after vaccination.

More research is needed to confirm the findings. But Michael Osterholm, director of the Center for Infectious Diseases Research and Policy at the University of Minnesota, said there's enough evidence now to suggest early vaccination efforts -- part of an attempt by commercial pharmacies to capture a bigger piece of the flu vaccine pie -- should be discouraged.

"I have been concerned for some time that we have gotten into the marketing of influenza vaccine versus the effective use of influenza vaccine. And we've got to reconsider that," Osterholm told STAT.

"Until we get more data, frankly I think the very best approach is to try to make sure we get flu vaccine into people just before flu activity starts, not something convenient to when the marketers want to get people in the door of department stores and grocery stores."

Other scientists say it's too soon to jump to a conclusion that would require changing recommendations. The CDC currently advises people should be vaccinated by the end of October.

"Is there waning [of protection] within a given season? I do think the evidence is growing and it's growing in a way that suggests that there is something there. But we need more information," said Jill Ferdinands, an influenza epidemiologist at the CDC and the lead author of the most recent article on this question.

Influenza vaccination programs are a major public health endeavor. Although the annual effort always falls short of vaccinating everyone -- the CDC estimated that only 40 percent of people had been vaccinated by early November, its most up-to-date data -- the US uses more than 100 million doses of flu vaccine every year. So far this year, manufacturers have shipped nearly 148 million doses .

Once people are vaccinated, it takes about 14 days for the immune system to generate a protective response, a factor in the debate over timing.

"It is hugely disruptive to try to immunize millions of people in a six- to eight-week period beginning in October or November. So I understand in the context of a universal immunization program, to get the vaccine into all those arms it's nice to be able to start earlier," said Dr. Danuta Skowronski, a flu epidemiologist at the British Columbia Center for Disease Control.

Canada starts its flu vaccination efforts after the US, generally in late October or early November. But Skowronski said at this point that's mainly because it can't get the vaccine from the suppliers sooner.

The biggest challenge for those who plan flu vaccination campaigns is the mercurial virus itself. Flu season is not like fishing season; it does not have a fixed start date.

The fact is health authorities know a flu epidemic will occur sometime between late November and late March, and that the epidemic will spike, then decline, roughly over a 13-week period.

But which strain of the flu will be the major cause of illness? How harsh or mild will the outbreak be? And most crucially, for planning purposes, when will the annual epidemic start? Even the best flu experts can only guess.

(If you're wondering about flu activity this year , it's definitely heating up across the country. If you've been meaning to get a flu shot but haven't gotten around to it, you can still get one -- but you'd be ill-advised to wait much longer. Flu has arrived.)

Some years the worst of the action is in late January and early February. Other years an early season will hit before Christmas. And many years after an initial peak caused by one of the influenza A viruses, H3N2 or H1N1, there will be a second wave of illness triggered by influenza B, in March and April. "It's part of the complexity of the question," acknowledged Ferdinands.

An unusually late flu season in 2011-12 in Europe sparked the latest round of questions about how best to time flu vaccination campaigns. Scientists who study how well the vaccine protects year after year saw a lot of cases of flu among people who had been vaccinated.

A flurry of studies have since followed. A big one from Europe , published last April, looked at data from five flu seasons. It saw protection against the influenza A virus H3N2 decline to virtually nothing by four months after vaccination. By contrast, protection against influenza A H1N1 did not erode and the vaccine's effectiveness against flu B viruses waned substantially, but didn't reach zero.

The CDC study, which looked at data from four flu seasons, showed a more consistent pattern. Protection against all the two influenza A families and influenza B viruses all declined steadily in the months after vaccination, at a rate of about 7 percent per month. (The decline was a bit steeper for protection against H3N2 viruses.) That might seem small, but people aren't starting from 100 percent protection. On average, flu vaccine is estimated to cut the risk of flu by about 50 percent or 60 percent -- and some years, the protection is less than that, when the strains in the vaccine aren't well-matched to those making people sick.

Ferdinands said given the way the studies are designed, one can't be certain the increasing cases late in the season are due to waning protection. For one thing, the circulating viruses might have evolved, which can diminish the effectiveness of the vaccine.

Additionally, people who get vaccinated early may be more vulnerable to flu than those who get vaccinated later. They may have chronic diseases that make them worry more about flu -- but that also could also diminish their immune systems' ability to respond well to a flu shot.

While the evidence is starting to mount, many flu experts appear wary about a change. Ferdinands noted that very few people actually get vaccinated in the summer.

Though summer vaccinations have increased slightly over the past few years, only about 2.4 percent of Americans get a flu shot before September, data from the CDC show. And over the past six years, less than 10 percent of people were vaccinated by the end of September. The peak month for flu shots was October, with between 16 percent and 18 percent of people getting vaccinated then.

Osterholm said public health officials should try to get those people who are getting vaccinated in July, August, and September to delay their shots. "I think that trying to move vaccination back to mid-October to mid-November catches almost even the earliest flu season and it may give us another couple of months of protection on the other end, when flu season continues on into early spring," he argued.

At this point, many researchers, like Dr. Arnold Monto, an influenza researcher at the University of Michigan, don't advocate for changing the recommendations; pushing too close up against the start of early flu seasons doesn't seem wise to them. That said, if you asked Monto when to get your flu shot, based on the emerging studies, you'd get this reply: "I would say: Don't get it too early."