Let’s play with numbers, and your head, just a little bit

From this site:

“New research brought to us by The Lancet shows some startling news regarding the true effectiveness of the flu vaccination. The study involved a control group of 13,095 adults who were not vaccinated. The group were watched to see if they caught the influenza virus, but 97 percent of them did not. Only 2.7 percent, or 357 people, of the non-vaccinated group ended up catching the virus. Another group of adults whom were vaccinated with a trivalent inactivated influenza vaccine ended up with 1.2 percent of them not catching the flu. The difference between the two outcomes is 1.5 people out of 100 which shows that the flu vaccine only prevents the flu in 1.5 out of every 100 adults injected with the flu vaccine.”
 
Emphasis so totally not mine.

But the person then explains their own misunderstanding:

“While the media runs around “spreading the rumor” that flu shots are 60 percent effective, one would assume that 60 out of 100 people receive the flu based on those claims. The problem with this claim is that it’s wrong. Anyone who takes a crash course in college statistics knows how to skew data. Methods for exaggerating data range from manipulating the graph to using complex statistical algorithms to eventually reach the desired conclusion. In this case, the 60 percent effectiveness claim births from an ongoing equation which transforms the numbers properly. First, 2.73% is taken for the people who got the flu in the control group. That number is then divided into 1.18% which stands for the percentage of people who got the flu in the treatment group. The answer comes out  to be 0.43. You are now able to say that 0.43 is 43% of 2.73 (control group people who got the flu) and make the claim 57% are protected by the flu vaccine.”

Well, yes, that’s how it works. This is what we epidemiologists call a “case-control” study, and it is very robust in terms of determining whether or not things happen by chance. In short, that reduction in influenza was not by chance, and it was significant, and it was by more than a half. But let me explain it differently.

The problem with looking at things in terms of percentages is that you lose sight of the magnitude of what you are looking at. If I tell you that 1% of the population of the United States has an ailment, you might think that’s not worth it to try and find a cure or prevention for it. But that 1% translates into roughly 3 million people. That’s a large city!

In the study cited by this particular anti-vaxer, you had a reduction in cases from 2.73% to 1.18%. Using our example of the US population, this would translate into helping 4.65 million people avoid the flu. In this person’s mind, going from 8.19 million to 4.65 million is meaningless. This person seems to be playing mind games to try and convince you that this is not a significant finding. It is, very much. It just seems small because, again, you’re looking at percentages.

So don’t just look at the percentages. Look a the whole picture. Even a reduction in disease burden of 1% or 2% is huge when it comes to saving lives and maintaining productivity.


Advertisements

One thought on “Let’s play with numbers, and your head, just a little bit

  1. But look at the wailing and teeth-gnashing by the same people going on about the 1.1% ASD prevalence in the U.S.; they seem to grasp what a small proportion of a population means. But only when it is convenient apparently.

Comments are closed.