Opinion Column

Fear of Flu: Shifting the Goalposts

If any public health message has alarmed Americans in recent weeks, it is the repeated claim in the media that healthy young people are dying of 2009 H1N1, or swine flu. A recent 60 Minutes report focused on the “unusual number” of otherwise robust children who’ve fallen ill and died from swine flu. USA Today’s headline on October 17 was, “Swine Flu Linked to 11 More Child Deaths.” The response has been predictable: frantic parents clogging waiting rooms in pediatricians’ offices, lining up for hours to get their kids vaccinated, and rushing to fill prescriptions for anti-viral drugs before the local pharmacy runs out.

While the death of a child is always a tragedy, is the panic justified? Let’s look at the statistics.

First up, while it’s happening earlier, is the 2009 H1N1 more dangerous than garden variety seasonal flu, for healthy kids compared to kids with underlying illnesses? According to a paper published last month in the New England Journal of Medicine, probably not. The authors found that 60 percent of children who were hospitalized with swine flu had underlying illnesses. In years past, only 31 to 43 percent of children who were hospitalized with seasonal flu had similar underlying illnesses – meaning that the rate at which 2009 H1N1 causes serious illness in healthy children is lower than with seasonal flu.

We found much the same picture when we looked at the actual reports of the first 36 pediatric flu deaths this year from around the country posted by the Centers for Disease Control and Prevention (CDC). The majority of those children (70 percent) had multiple, severe underlying conditions, such as congenital heart and lung disease, for instance, or cerebral palsy. And that’s not counting morbid obesity, which many flu experts believe should be considered a risk factor.

Certainly it appears as if there have been proportionately more deaths among young people than the elderly compared with seasonal flu years, but that’s largely because fewer old people are dying than usual. They may have some residual immunity from being exposed to related viruses in the past, but whatever the reason, the proportion of deaths among younger people is automatically larger this year.

According to the CDC’s most recent update, there have been 156 laboratory-confirmed 2009 H1N1 flu deaths in the U.S. among children and teens 17 years old and younger. Last week, the CDC extrapolated from those numbers to estimate about 540 pediatric deaths. (The estimate is derived by taking the number of laboratory-confirmed flu deaths and running it through a complicated formula and computer model.)

It sounds like a lot of kids dying, but we don’t really know how the death rate compares to previous years. During last year’s regular flu season, for example, there were 115 laboratory confirmed flu deaths reported among this same age group.

An additional complicating factor: It turns out that the number of reported pediatric flu deaths has been steadily rising each year since 2005 (and possibly earlier). During the 2005-2006 flu season, 46 pediatric lab-confirmed deaths were reported. In 2006-2007 the number rose to 78. By 2007-2008, there were 88 deaths, followed by 115 last year.

Although the number of this year’s deaths will probably rise in coming months, it’s important to compare apples to apples. To date, the number of laboratory-confirmed pediatric deaths is only marginally greater than last year’s – and a disproportionate number of those children have underlying illnesses.

No one knows for sure why flu deaths among children are going up, but the CDC told us that doctors are probably “doing a better job of reporting.” Better book-keeping, in other words, not more virulent viruses, is responsible for at least part of the apparent increase in pediatric deaths in recent years. Based on this trend alone, we would expect that the number of pediatric deaths would be higher this year than last.

What’s more, additional changes in book-keeping may exaggerate the total number of swine flu deaths reported among all age groups. This year, the CDC liberalized the definition of what it counts as a flu death for people over age 18. Until now, only deaths in which the patient had laboratory-confirmed influenza were used to estimate the annual mortality statistics for seasonal flu, which the CDC says causes about 36,000 deaths a year in the U.S., according to their mathematical modeling projections.

As of August 31, however, the CDC said that anyone 18 years or older who dies with a syndrome known as influenza-like illness, or ILI, can be included in the count. The CDC has no requirement for uniform reporting of flu deaths by the states. Twenty-eight states, according to the agency, report only laboratory-confirmed flu deaths, while the others are now reporting deaths that look to doctors like they might be the result of flu.

That decision could dramatically inflate the numbers of reported flu deaths this year. By the CDC’s weekly estimate, when doctors think a patient has the flu, they are wrong between 46 and 82 percent of the time. This isn’t a criticism of doctors; it is virtually impossible to distinguish flu from illnesses caused by germs like coronavirus and many others. However, it does mean that once the CDC extrapolates its estimates for the entire population, swine flu mortality could appear to be as much as an order of magnitude greater than seasonal flu, whether or not there is any real increase in deaths.

Why should any of this matter? In part, because statistics, when filtered by the popular media, can scare the daylights out of the public and feed pandemic worries that may not be justified. The numbers also matter to public policy. As reported deaths rise, the CDC tends to enlarge its recommendation for who should be vaccinated, and who should receive anti-viral drugs. For example, based on rising pediatric deaths, the CDC issued new recommendations in 2009 that all children from 6 months to 18 years should get the flu vaccine each year. Previously, the vaccine was recommended only for high-risk children.

That’s not to say that we all shouldn’t exercise caution this flu season, or that the U.S. government is deliberately fanning fear by changing its definition of what counts as a flu death. But at a time when we are worrying about rising health care costs, there’s an argument for basing multi-billion-dollar investments in vaccine campaigns and stockpiling antiviral drugs on solid numbers.