statcounter

Saturday, November 14, 2009

Can we trust the CDC?

CDC, Atlanta Campus

CDC, Atlanta Campus

In a press conference this week representatives of the Centers for Disease Control revealed estimates of the extent of the H1N1 Influenza pandemic that increased by several orders of magnitude from previous estimates. For instance the number of pediatric cases reported by the CDC on November 6th was 129, but on November 12th, the date of this news conference, it was reported to be 540! Other surprising numbers revealed included the following:

• CDC estimates that between 14 million and 34 million cases of 2009 H1N1 occurred between April and October 17, 2009.
• CDC estimates that between about 63,000 and 153,000 2009 H1N1-related hospitalizations occurred between April and October 17, 2009.
• CDC estimates that between about 2,500 and 6,000 2009 H1N1-related deaths occurred between April and October 17, 2009.

I currently help staff an urgent care facility that sees nearly 100 patients per day. We have seen very large numbers of people with complaints that could be due to influenza, yet have documented only about five to ten per week and that only began about the first of October. We have hospitalized no patients and our parent hospital that serves a population of about 150,000 has reported one death due to H1N1 and that was in a child with severe underlying immunological disease.

I know that anecdotal reports are not always representative of a situation, but still the new CDC estimates seemed so far off from what many of us front-line docs are actually seeing that I had to investigate this further.

I read the entire transcript of the news conference as well as the accompanying press release to try to get a better idea of what the CDC is now reporting. Amongst a lot of uncipherable medico-babble (e.g., “Emerging infection programming network is one key source for our estimate. We also are using data from other symptoms like aggregate state reporting of laboratory hospitalizations and death.”) I think I determined the fundamental problem with these latest figures.

The CDC has adopted, for the first time in epidemiological methodology, a technique that extrapolates numbers from very limited and dubious samplings. Reports were compiled 64 counties from ten states. How many counties are there in our fifty states, total? I didn’t feel it necessary to calculate an exact number, but Indiana has 92 and South Dakota has 66. Can we conclude that we are working with a very small sample?

Next, researches at the CDC used the “Emerging infection programming network” that compiles data and “develop[s] new methods for gathering epidemiological and clinical information.” Some of those new techniques are a bit over-reaching. For instance they have determined that for every case of “flu” reported (and not necessarily proven by testing) there are 79 cases that are unreported. Similarly, the number of hospitalizations due to flu is “corrected” by a factor of 2.7. In those cases in which an influenza test was done and found to be negative researchers at the CDC have determined that up to 30% were actually H1N1 and should be counted. Additionally, deaths from what could be a complication of influenza, such as pneumococcal pneumonia, are assumed to be related to the H1N1 virus. (Pneumococcal pneumonia is the most common type of pneumonia and claims tens of thousands of lives every year.)

In short, I am very skeptical of the latest estimates from the CDC. I fear this organization that was once scrupulously devoted to science and working only for the best interest of the American people has been corrupted by politics and now serves as a propaganda arm of the entrenched bureaucracy. Yet another example of how the government can screw-up most anything it touches.