You are here

Should we believe the latest CDC vaccine study?

The CDC recently released a study titled "Historical Comparisons of Morbidity and Mortality for Vaccine-Preventable Diseases in the United States", just published in JAMA and reported on in the New York Times.  It alleges to demonstrate that vaccines were responsible for a preponderance of the decline in deaths for most "vaccine-preventable" diseases. 

You can almost hear the collective sigh of relief.  But is what they are reporting true?

In order to try and begin to understand whether or not the CDC fairly made their case, some of the data concerning measles and pertussis will be examined and discussed. 

1)  When comparing pre-vaccine cases/deaths to post-vaccine cases/deaths, the years chosen appear to be those just prior to introduction of the vaccine.  Were all years for which data exist considered, however, dramatic declines prior to introduction of vaccine would be in evidence.  For instance, the estimated annual average pre-vaccine cases and deaths for measles were based on the years 1953-1962.  By then, the total number of deaths (see below) and death rate per 100,000, however, had already dropped significantly.  Ignoring the earlier years falsely attributes most of the decline in deaths to vaccination.

Click here (1,2)  for the CDC data upon which the death rates and incidence graphs are based.

2)  Although the pertussis vaccine was introduced before the measles vaccine and hence an earlier pre-vaccine period for pertussis, 1934-1943, was chosen, death incidence (see below) and death rates still had declined significantly by the end of the "pre-vaccine" period.  Ignoring the earlier years in this case also falsely attributes most of the decline in deaths to vaccination. 

Click here (1,2)  for the CDC data upon which the death rates and incidence graphs are based.

*Note that these declines are even more dramatic when one considers that the population was growing at the same time deaths from these two diseases were declining.

3)  For reasons that are unclear, they chose the year 1914 as the beginning date for pertussis vaccine, even though the vaccine was not deemed particularly effective or widely used in those early days, and even though by their own admission, 1934-1943 was "pre-vaccine".  Was the purpose in starting with that date an effort to make it seem that any early declines could be attributed to vaccination, when, perhaps they should not be?

4)  Cases and deaths reported as the "most recent post-vaccine reported numbers", came from different years - 2006 for the "cases", 2004 for the "deaths".  This is like comparing apples and oranges and is not meaningful.  It would have been better to use the 2004 cases, if they only had the 2004 deaths, than to use different years.

An interesting example of how cleverly a graph can be devised can be seen in the MMWR. Upon examination of their pertussis incidence and deaths graph, a number of important things become apparent.  First, the CDC inappropriately used a logarithmic scale to show the declines.  A logarithmic scale shows "rate of " change rather than absolute change.  This particular graph shows 90% declines.

The graph also indicates that the vaccine's "minimum potency (was) established" around 1950 and that the "standard unit" was established around 1955.

Now a 90% drop from 10 is more significant in people terms than a 90% drop from 0.1.  And, although it is not easily discerned from the graph, most of the big drops in deaths occurred prior to standardization of the pertussis vaccine.  In fact there were two approximately 90% drops by the time the standard unit had been established.

Had the y-axis, which shows the incidence, plotted absolute numbers rather than 90% rates of decline, as it does in this graph, the huge declines prior to standardization would have been clear as a bell. 

When the y-axis is plotted properly, as it is in the "pertussis deaths by year" graph above, it will show that significant declines occurred even before the 1934-1943 CDC-designated "pre-vaccine" period.  But unless standardization added little to nothing to pertussis vaccine effectiveness, "pre-vaccine" should extend to around 1955, in which case the drop in deaths prior to vaccination would be seen to be even more dramatic.  (Click here to see what the death rate would have looked like had the data been properly and meaningfully graphed.) 

Besides, given that the CDC has data going back as far as 1922, why didn't they call 1922-1943 "pre-vaccine" rather than start in 1934?  Is it because the drop in deaths prior to vaccination would have been shown to be even more striking?

All this lends credence to the notion that the CDC, in the recent JAMA article, seems to be continuing its pattern of ignoring the huge declines in deaths for these diseases prior to vaccination.

They have a long history of a) exaggerating the benefits of vaccines while at the same time ignoring and/or dismissing their long-term risks and b) exaggerating the long-term consequences of diseases.  At least insofar as pertussis and measles are concerned, this most recent article appears to be merely more of the same.  For those seeking increased confidence in at least those particular vaccines, this study does not appear to be the place to find it.

Sandy Gottstein

Date: 11-15-2007

Among the many Scandals columns relevant to this discussion are the following:  Scandals: Shoot First, Don't Ask Questions Later; Scandals: Measles In The Vaccination Age:  Is It Now Deadlier? and Scandals: More confusing disease stats.

"Eternal vigilance is the price of liberty." - Wendell Phillips (1811-1884), paraphrasing John Philpot Curran (1808)