Flu Vaccine’s Declining Efficacy – the Past 21 Years
November 23, 2015

How effective is the flu vaccine?

Usually not very good – with a trend towards declining efficacy. The Fitzpatrick team analyzed the overall influenza vaccine effectiveness (VE) published over the past 21 years (see graph below).

The whole determination of effectiveness is nebulous, so rather than have us err on our interpretation, we pulled the “Overall Vaccine Effectiveness” stats right from the CDC and/or other credible sources. In seasons where we were unable to find “U.S. Overall Vaccine Effectiveness”, we applied data from Scotland as surrogates.

Using a 7-point grouped analysis, the r2 = 0.39, showing correlation. While this is not a perfect 1.0 linear regression correlation, it shows a trend. Compounding the declining efficacy is the huge surge in vaccinations. The graph below illustrates how vaccination levels have tripled over the past 2 decades.

How well the flu vaccine works (or its ability to prevent flu illness) can range widely from season to season and by geography. Vaccine effectiveness can also vary depending on who is being vaccinated. The following factors play an important role in determining the likelihood that flu vaccine will protect a person from flu illness:

a. A person’s age
b. A person’s health
c. The similarity or “match” between the flu viruses and the flu vaccine
d. The severity of the season
e. Antigenic drift between the egg-based vaccine and the WHO-recommended strain
f. Other miscellaneous factors (i.e., Even in good-match years, the benefits of vaccination will vary across the population, depending on characteristics of the person being vaccinated and even, potentially, which vaccine was used. Vaccine effectiveness can vary significantly by age cohort and health status.)

There are a lot of researchers (hundreds? thousands?) at the World Health Organization, CDC, DHHS and state health offices, who are responsible for gathering and publishing continuous influenza surveillance and vaccination information. There is no shortage of literature accessible from a variety of government-sponsored sites that share important, up-to-date information on prevailing strains. The influenza data they gather and display is valuable, but why are researchers ignoring the 800-lb gorilla in the room? That seasonal influenza vaccine effectiveness is following a dangerous downward trend.

The perception that the influenza vaccine is highly effective in preventing seasonal flu is perhaps half true. Looking at the coefficient of determination plotting out 21 years, the r2 value = 0.267.

Year Reference Overall Match:
1994-95: 62%
1995-96: 66%
1996-97: 90%
1997-98: 50%
1998-99: 86%
1999-00: 54%
2000-01: 77%
2001-02: 68%b2002-03: 49%b2003-04: 44%b2004-05: 10%c2005-06: 21%e2006-07: 52%e2007-08: 37%e2008-09: 10%e2009-10: 56%e2010-11: 60%e2011-12: 47%e2012-13: 49%e2013-14: 51%e2014-15: 23%eNOTE: Except for the CDC’s Overall VE data for 2005-2015, there is no single repository for identifying prior year’s efficacy; the authors are including the most appropriate references.

a2009-10 the vaccine effectiveness could not be estimated; pandemic year, Monovalent vaccine was created.
bData from Simpson, et al. Scotland data. This is not U.S. data, however the vaccines are the same northern hemisphere trivalent.
cManufacturing problems resulted in a loss of nearly half of the US vaccine supply for the 2004-2005 season
dOverall vaccine effectiveness data was not published, however the match by strain was.
e2005 – 2015 also had a single reference for Overall Vaccine Effectiveness (VE):

We would be remiss in suggesting that mismatched strains were the sole cause of poor vaccine effectiveness. In some years, the field isolates were very well-matched against the WHO-recommended strain, as what happened in 2012-2013. In a paper published by Skowronski, et al, in that season, the egg-based strain was antigenically different fro the WHO strain (in the vaccine); there were three mutations which must have diminished effectiveness. So when the flu vaccine does not perfectly match the CDC-recommended strain, the problem would seem to be antigenic drift.

In summary, comparing the randomness of flu vaccine effectiveness versus other health treatments, I can’t think of many good examples where the doctor would be happy saying:

• Sorry, Kevin – there is only a 23% chance that this pain drug will work while we are performing your hernia operation.

• Sarah, you can get vaccinated; the chances of it being effective ranges from 10% – 90% – but then again, you might not get the flu anyway.

• Sorry Miss, there is only a 50% chance that this oral conceptive will work.

If the WHO and FDA’s vaccine advisory committee is going to continue to use their existing, antiquated flu vaccine strain development system, perhaps we need to lower our expectations for flu vaccine effectiveness. Or, we could embrace JIT (just in time) recombinant technology and not wait 6 months to let the influenza strain(s) mutate. JIT manufacturing is a common standard across a wide variety of industries; one tenet of which is to lower risk. For flu vaccine selection, this would mean lowering the risk of a poor match. Policy-makers should demand improved and more predictable vaccine efficacy.

Kevin Fitzpatrick, MBA is president of Fitzpatrick Translational Science www.fitzpatricktranslationalscience.com and has written or co-designed >300 evidence-based health tools.

Brock Fitzpatrick is a sophomore at the University of Wisconsin, College of Letters and Science.

Competing interests: The authors have declared that no competing interests exist.
Copyright © 2013 - 2019 Fitzpatrick Translational Science | All Rights Reserved