Here's The Calculation The CDC Used To Determine That This Year's Flu Is The Worst In A Long Time

nurse, hospital, doctor

Photo: Wikimedia Commons

The centre for Disease Control is the flu-monitoring powerhouse of the federal government, with an army of statisticians working around the clock during flu season to stave off a pandemic. In order to monitor the impact that microscopic virus has on a population of 300 million, the CDC has built a vast network of more than 2,700 outpatient health care providers.

The U.S. Outpatient Influenza-like Illness Surveillance Network (ILINet) reports statistical information about more than 30 million patient visits every year.

Here’s how it works. Every week, around 1,800 of the 2,700 ILINet participants report data to the CDC about the per cent of patients seen with influenza-like illness symptoms.

Anytime someone comes in with a 100 degree fever and a cough or sore throat without a known cause, that’s one person marked with influenza like illness symptoms. 

The national baseline is 2.2 per cent of patients reporting influenza-like illness symptoms every week. The baseline is derived by calculating the mean percentage of patient visits for flu-like symptoms during non-influenza weeks for the preceding three seasons and adding two standard deviations. 

Regional baselines are also calculated. These range from a baseline as low as 1.4 per cent in the upper northwest region to a baseline as high as 3.5 per cent in the southwest region.

The way the CDC determines flu severity depends on the difference between the reported percentage of influenza-like symptoms in a given week compared to the national or regional baseline. 

Here’s where the maths comes in.

standard deviation normal distributionTo the right, 4-6 standard deviations is a moderate flu risk. Higher than 6 standard deviations is high.

Photo: Wolfram Alpha

Every week, the CDC calculates the number of standard deviations away from the mean the reported percentages for each region are.So for the southwest region, the CDC calculates severity based on how many standard deviations higher the reported percentage is from the 3.5 per cent baseline. 

The 10 flu “activity levels” correspond to the number of standard deviations above or below the mean the current week compared to the baseline.  

Here’s how they translate standard deviations to English:

  • Values lower than 2 standard deviations above the mean are classified as “minimal”
  • Values that are between 2 and 4 standard deviations above the mean are “low”
  • Values that are between 4 and 6 standard deviations above the mean are “moderate”
  • Values that are between 6 and 9 standard deviations above the mean are “high”
standard deviation1.645 standard deviations above a mean of 0 on a normal distribution

Photo: Wolfram ALpha

Now, here’s where mortality comes into play. The CDC also receives data from 122 cities reporting the total number of death certificates processed and the number of those that listed pneumonia and influenza as a contributing or underlying cause. 

These number of deaths are compared to a seasonal baseline and also an “epidemic threshold value” calculated for each week. 

The baseline comes from a regression of data from the previous five years. 

The “epidemic threshold” is an increase of 1.645 standard deviations above that baseline, when the proportion of deaths from the flu are significantly higher than expected at that time in the absence of widespread flu mortality. 

We just passed the epidemic threshold for 2013. This sort of thing is exactly why if someone hasn’t recieved a flu shot, they need to get one now. The epidemic threshold for this week was 7.2 per cent of deaths being the result of the flu. We’re at 7.3 per cent. 

Read more about the CDC methods here. 


Business Insider Emails & Alerts

Site highlights each day to your inbox.

Follow Business Insider Australia on Facebook, Twitter, LinkedIn, and Instagram.

Tagged In

flu math politics-us