You may have seen the recent news stories this past week that the seasonal influenza vaccine is working better than the original predictions. New estimates report that the overall effectiveness is 39% and in children it rises to 59%. This is fantastic news, but why the change in the estimates?
First, our estimates of both what viral strains to include in each year’s vaccine and the estimates of its effectiveness are based off of data in the southern hemisphere, such as Australia. This may seem obvious, but I will just say it -the seasons there are reversed compared to us. Thus, while we are sweltering in the heat of the summer, they are fighting off the cold and influenza season. In addition to understanding where the initial estimates come from, it’s important to understand how they are calculated and what they mean.
The flu effectiveness rate is calculated from a modified case control study called a test negative design. Essentially, the health departments look at people seeking care for “influenza like illnesses” or ILI, at certain centers. Influenza like illnesses are anything that looks like it may be the flu with symptoms of fever, cough, congestion etc. They then look at all those people that came to the centers with influenza like illness that were tested for the flu and look at who was positive, who was negative, and who was vaccinated and who wasn’t. The effectiveness is calculated by comparing the proportion of people with laboratory confirmed influenza in patients who have been vaccinated compared to those who were positive and unvaccinated. Article detailing test negative design for calculating vaccine effectiveness can be found here.
In addition, the single estimates of 39% or 59% are point estimates. A point estimate is the best estimate because the the entire population cannot be observed. A point estimate always has confidence intervals which is the interval that the true number actually lies within. If the confidence interval is wide, then the point estimate may not be that accurate. If the confidence interval is narrow, then we can trust the point estimate more.
One important thing to understand about this calculation and the early estimates we receive from the southern hemisphere is that Australia has different recommendations for influenza vaccination that the United States. Australia only recommends vaccination for people who are older than 65 or chronically ill. These people will have a less robust immune response to any vaccine, including the influenza vaccine. Thus, the number of people who are vaccinated, but will still contract the flu, will be higher in the Australian population because they don’t mount as good of a response to the vaccine. This is also why our children are more protected than the general population in the United States right now. They have better immune responses to vaccines than older patients or people who are chronically ill.
Why would other countries only recommend vaccines for people that don’t mount a good response to the vaccine? First, multiple studies show that the flu vaccine is better at preventing serious complications and death from the flu, than actually preventing “catching the virus”. Thus, even though you may get still contract influenza after the vaccine, you are less likely to get pneumonia, sepsis, or die. People who are older or chronically ill are more likely to have this happen, so it makes sense to try to protect them. Second (and my opinion), the recommendations in countries with universal health care likely take cost into account and weigh this against the risk of serious complications from the flu and concentrate on vaccinating only the people at highest risk. So, if we ever switch to a universal health system would our recommendations change? I don’t know but, it’s likely that it would be scrutinized heavily.
Here is a graph of the last several years influenza vaccine effectiveness rate. Remember, this is only calculating the proportion of people who were vaccinated and avoided contracting the flu. This does not tell us those people who were vaccinated, got the flu, but avoided death.
So, the flu vaccine can prevent infection in a proportion of vaccinated individuals and can prevent death in those people who received the vaccine but still contract the flu. If you want to read the multitude of primary articles detailing this with scientific rigor, they can be found here: Good flu articles.
The above graph shows where we stand right now in the flu season, it’s not over yet, but it looks like it has peaked. Thanks to Dr. Rodriguez for explaining the calculations to me several weeks ago.
