Dear Science,

My brother says that a recent article in the Atlantic says that some of the most methodical research casts doubt on whether the flu vaccine is very effective, though that is (as my brother puts it) a "medically unpopular" view. I cannot be bothered to read the article, but what do you think about this? And since when is medicine a popularity contest?

Sister

Science, particularly medical science, is often an art of deriving conclusions from messy information. Any time experimentation must be done on people to gather an answer, ethical and practical restraints limit the raw data available. For flu vaccination, most of the studies in support—claiming that those vaccinated against influenza have about a 50 percent lower chance of dying during flu season (of all causes) compared to those not vaccinated—are indeed flawed. Any time you retrospectively compare two groups—those who were voluntarily vaccinated compared to those who opted against—a whole slew of confounders enter in. Perhaps the same characteristics that make people choose to be vaccinated also make them more cautious—better hand-washers, better coughers, more likely to stay home when sick, and so on.

Shannon Brownlee and Jeanne Lenzer's Atlantic article jumps off of a clever study done by Group Health researchers. If the previous studies compared deaths between those vaccinated and not vaccinated during flu season, why not use those same groups to compare death rates outside of flu season? Indeed, those who opted to vaccinate against flu were 60 percent less likely to die during months when flu wasn't around; the reduced risk of death observed in the previous flu-vaccine studies can be attributed to this effect rather than the flu vaccine itself.

Is this proof that the flu vaccine doesn't work? No. It's evidence that we don't have good data on the flu vaccine. (Please don't cry right now.) The vaccine might work, or it might not work. We don't know. The way to figure it out is to take some of those who volunteer to receive the vaccine and randomize them to either receive the vaccine or a placebo shot—creating a blinded, randomized, and controlled trial. This sort of prospective study—with the randomization at its core—is the best tool we have in clinical medicine to ascertain if the vaccine helps. Science supports such a study and feels it would be ethical.

With all this said, there is good evidence that the vaccine doesn't cause harm and does help provoke an immune response in otherwise healthy people against the flu virus. Based on the (flawed) available information, Science still thinks it's best to get vaccinated—in addition to washing your hands, covering your coughs, and staying home when you feel sick. The latter works for any virus, not just the flu.

Cohortingly Yours,

Science

Send your science questions to dearscience@thestranger.com.