Thursday, June 11, 2009

Pandemic or Not?


On The New York Times:

The World Health Organization has backed itself into a corner on its Pandemic Influenza Phases system and the so-called swine flu. WHO knows it and is working on bureaucratic escape routes. But the real problem is what these alert systems mean for national governments that have detailed pandemic response plans — particularly the United States.

After decades of ignoring the influenza threat, the 2005 emergence of a virulent strain of H5N1 bird flu spawned global concern. The key outcome of that collective worry was a flurry of planning, at all levels of governance, from WHO in Geneva on down to municipal councils in Kansas and Japan. The plans vary, of course, but all start from roughly the same point: An announced pandemic.

There has been no pandemic announcement for H5N1 because the bird flu virus is extremely virulent, but not apparently contagious between human beings. The WHO Pandemic Influenza Phases system requires evidence of sustained transmission between human beings in two distinct parts of the planet at the same time. That standard has not been met for H5N1 bird flu: So far, so good.

But based on that same WHO threat scheme, we do now have a full-fledged pandemic of swine flu. The world should be at Pandemic Influenza Phase 6 right now.

However, the H1N1 virus has so far proven to be a wimpy virus. Its virulence seems to be well below that of routine seasonal flu, and its so-called RO (or reproductive number) is about 1.4, based on Mexican data. This statistic means each person who acquires the virus infects about 1.4 other people. That’s certainly not a terrifying level of contagion. In contrast, the dreaded 1918 influenza pandemic had an RO of around 3; on average, each infection was tripled. A constant three-fold expansion of an epidemic is a frightening prospect that utterly defies government planning. The concept of epidemic “containment” in the face of an RO of 3 is patently ridiculous.

But if the swine flu RO turns out to be 3, or even 4, do we care? Should the world be put on alert for a highly contagious — but very low severity — virus? WHO experts are mulling this problem, trying to figure out how to de-couple their system from geography, and link it more closely to severity and the level of contagion.

Nice idea, but not an easy feat, as New York City has discovered. When the first cases of swine flu emerged in mid-April among students at the St. Francis parochial school there was no immediate way to figure out either how virulent or contagious (RO) the virus was. An Australian group has studied the St. Francis school data and concluded the RO in that school was a very scary 3.45 — perhaps as high as 4.28. The H1N1 virus may be a mild organism, but if each individual case infects more than 4 others, mere statistical odds dictate that many people will suffer and die from the virus. To be blunt, if roughly four times more people can get infected with a mild flu, as compared to seasonal influenza, the virus could be 400 percent less virulent and, theoretically, kill just as many people as a tougher seasonal virus. It’s a crapshoot. But gambling is not a sound basis for policymaking.

The U.S. pandemic plan links government action to phases of viral activity. While the U.N. system has six threat levels, the Bush administration-designed approach grew out of the Hurricane Katrina experience, and, like our storm forecast system, has five categories. Each category reflects a different level of severity of the organism, based on an estimation of the percentage of people who die as a result of infection — a case fatality ratio. A Category 1 is theoretically declared for a pandemic organism with a case fatality ratio of less than 0.1 percent: Category 5, requiring urgent government action, is declared if more than 2 percent of those infected are killed by the virus.

This has proven ludicrous. In the case of bird flu, fatality rates are in the 63 percent ballpark — far above Category 5 levels. But the wimpy swine flu H1N1 doesn’t even appear to merit a Category 1 designation. The U.S. category system has no metric to discriminate human-to-human transmission rates.

This is not trivial. Federal guidance for such actions as mobilization of vaccines is based on these designations. And categories are based on death rates — numbers we still don’t have in validated form. Why does this matter? On a global scale these influenza phases and categories befuddle everything from vaccine manufacturing to negotiations on access to medicines.

We need rapid diagnostic tests to determine who has been infected. Much of our delay in understanding what category or phase to assign to swine flu has stemmed from the near impossibility of determining just how many Mexicans were infected with the H1N1 virus, a figure that now appears to have been far larger than initially guessed. Even in New York City, officials lack a toolkit that could easily discriminate genuine H1N1 infections from hay fever. We don’t know what our denominator is, so we cannot make a case fatality or morbidity ratio. (One death out of 100 New Yorkers infected would be horrific; 1 out of 100,000 would be a personal tragedy but hardly a public calamity.)

In the absence of such tools of precision, the wise course would be to merge the phase and category systems, putting the World Health Organization and national governments in harmony. For example, in the case of the H1N1 swine flu, WHO should declare that the world is now facing a Phase 6 pandemic influenza spread, caused by a Category 1 organism of low severity. An appropriate designation for the H5N1 avian flu: Phase pre-1, Category 5.

Governments should only set up action and threat assessment systems that are flexible, and useful. The Bush administration learned this in the years following 9/11, as the nation grew weary of color-alert systems for terrorist threat assessments. Nobody is well-served by repeating such mistakes in pandemic planning.