How Many People Die From Flu Each Year? Depends How You Slice The Data

The Centers for Disease Control and Prevention's says that the 36,000 annual flu death toll number the media's been using might not be so accurate anymore.

Have the media been scaring people enough when it comes to reporting flu's death risk? Or too little?

For as long as I can remember now, we've been saying 36,000 people die each year from the flu.  When we've asked the Centers for Disease Control and Prevention for updated figures, they told us 36K was the best they had.

Didn't seem quite right that it never changed year after year.

Now it turns out the 36K was calculated way back in 1999, when flu deaths hit a peak.

A new analysis of the past 31 flu seasons, put out by the CDC this afternoon, shows that 36K might have been a third too high. It's more like 23,607 per year, on average. But it all depends on how you slice the data.

The new average is an average of the past 31 seasons. It goes all the way back to 1976. But if you were to look at just the past 10 seasons, the average per year is higher -- 32,743.

Or you can look at this way: In some years as few as 3,349 have died (back in 1986-7).  But the highest annual average toll was 48,614, just seven seasons ago (2003-4).

The reason for this wide variation: Deaths peak when the H3N2 strain of influenza A dominates. When it's H1N1 or influenza B, the toll is quite a bit lower.  Scientists can guess which strain may predominate in a given year, but it's only a guess. That's why the flu vaccine doesn't work in some years.  But it usually does.

It may be that the average death toll will decline somewhat as immunization rates go up. The CDC is now recommending that everyone older than 6 months get a flu shot each year.  (It used to be just children and seniors.)

The CDC wants reporters to drop the 36,000 annual estimate (at last) and instead say that the death toll ranges from 3,300 to 49,000.

So what number should NPR use?  We try to be as precise as possible, but the wide range doesn't seem quite right.  After all, it's been 25 years since as few as 3,300 people died from the flu – most years since the death toll been anywhere from 4 to 16 times higher.  So we won't be reporting the lower number very often, unless we're reviewing history.

We're still mulling over exactly what we'll say in our reporting – we definitely say "tens of thousands die from flu, on average, each year." But whether we get more specific than that is still under discussion.  Your ideas are welcome.