Seniors who get flu shots significantly reduce their risk of death or hospitalization, says a new study that contradicts another recent review on the topic.

Last week, a review in The Lancet Infectious Diseases concluded that the benefits of giving the flu vaccine to elderly people have been greatly overstated, noting that most of the studies suggesting the vaccine prevents deaths are weak.

But a new analysis of 10 years worth of data, published Thursday in the New England Journal of Medicine, finds that, compared to those who don't get vaccinated, elderly people who get the flu shot are: 

•  27 per cent less likely to be hospitalized with flu or pneumonia
• and 48 per cent less likely to die from flu-related complications.

Lead author Dr. Kristin Nichol of the Minneapolis Veterans Affairs Medical Center and her colleagues studied data pooled from 18 groups of U.S. seniors living in the community -- as opposed to those living in nursing or retirement homes -- from 1990-91 through 1999-2000, and two other groups from 1996-97 through 1999-2000.

They found the benefits of vaccination varied from year to year. That's in part because there are dozens of influenza strains circulating in the community and the vaccine cannot protect against all strains.

Each year, the strains that will included in the annual flu shot are selected, based on the best guess of experts as to which strains will be most prominent in the coming months. Some years, the experts get it right; other years, they don't. In the years they don't get it right, the flu shots offer less protection.

After taking into account those year-to-year variations, the authors conclude that flu shots significantly reduce flu-related hospitalizations and deaths in community-dwelling seniors.

They say their findings underscore the need to improve vaccination rates in this high-risk population group.

"Vaccine delivery to this high-priority group should be improved," they wrote.

Part of why last week's review in The Lancet and this latest study came to such different conclusions is that each research group used different statistical methods.

Lone Simonsen, an infectious diseases epidemiologist at George Washington University in Washington, D.C., one of the authors of The Lancet review, takes exception with the way Nichol's team reached their conclusions.

She told The Associated Press Wednesday that Nichol's method "greatly exaggerated" the vaccine's benefit in seniors. She says Nichol's findings translate into one death averted for every 200 seniors vaccinated -- a number that Simonsen believes overestimates the real effect by 100 times.

Nichol responded Wednesday that she disagrees and defended her method.

"Our conclusions are very robust. We know that," she told AP.

In any case, the authors of both studies agree on one thing: the elderly should continue to get flu shots.

"Influenza causes many deaths each year, and even a partly effective vaccine would be better than no vaccine at all," wrote the Lancet authors.

And the New England Journal of Medicine authors conclude their study by writing: "Even as we wait for new vaccines and new strategies, patients, their health care providers, and policymakers should renew efforts to improve the delivery of current influenza vaccines to this high-priority group."