CDC Proves Flu Vaccine Ineffective

Hurmanetar

Hurmanetar

New
As discussed in this thread...
http://www.skeptiko-forum.com/threa...never-made-headlines-and-what-that-mean.2688/

We've been battling some stupidity in high places regarding the flu vaccine.

This led me today to look at some of the CDC's published information on the flu vaccine and I came across this CDC presentation regarding the effectiveness of last year's (2014-2015) flu vaccine. It was a bad year for the flu vaccine since the CDC only claimed 23% vaccine effectiveness, but after looking at the numbers I think that even this paltry result is an outright fraudulent claim.
http://www.cdc.gov/vaccines/acip/meetings/downloads/slides-2015-06/flu-02-flannery.pdf
http://www.cdc.gov/flu/fluvaxview/coverage-1415estimates.htm#age-group-adults

Some interesting things I noted:
  1. The vaccination rate of the enrollees with flu-like symptoms (ILI) was 53% which was higher than the national average vaccination rate of 47.1%.
  2. The vaccination rate of the group with confirmed flu was 49% which was higher than the national average of 47.1%.
  3. The vaccination rate of the enrollees with ILI of 18+ years was 59% which is higher than the national average for this age group of 43.6%.
  4. The vaccination rate of the group with confirmed flu and 18+ age was 57% which is higher than the national average for this age group of 43.6%.
  5. The vaccination rate of the enrollees with ILI and 18-49 age was 44% which is higher than the national average for this age group of 33.5%.
  6. The vaccination rate of the group with confirmed flu and 18-49 age was 43% which is higher than the national average for this age group of 33.5%
  7. Since vaccination rates among the enrollees are higher than the national averages, the CDC's own data suggests that the vaccine has negative effectiveness.
  8. Over 80% of the flu viruses tested were antigenically different than the vaccine.
  9. 378 enrollees seem to have disappeared without explanation.
  10. This is not a peer-reviewed published study.
  11. A note is repeatedly displayed: "Adjustment for study site, age, sex, race/Hispanic ethnicity, self-rated general health, days from illness onset to enrollment, and calendar time (2-week intervals)," but no explanation is ever given for how or on what basis these mysterious adjustments were made.
  12. It is not clear how "adjusted OR" is calculated in their effectiveness formula: VE = (1 – adjusted OR) x 100%
  13. Therefore it is not clear how the 6% higher average rate of vaccination in those with flu-like disease vs. the confirmed flu is transmogrified into the claimed 23% effectiveness.
So does anyone with more knowledge of statistical magic than me know how the CDC transmogrifies this data into the claim of "23% vaccine effectiveness"?

Am I missing something or are the CDC's own numbers indicating negative effectiveness since vaccination rates are higher than the national averages in these groups of people with flu-like and flu-confirmed disease? I can maybe imagine that the kind of people who are likely to visit the doctor due to sickness are the kind of people who are more likely to get the vaccine, but I don't know that for sure.

Does anyone with more knowledge of scientific studies and statistics see other blatant flaws in this methodology?
 
Last edited:
Hurmanetar said:
As discussed in this thread...
http://www.skeptiko-forum.com/threa...never-made-headlines-and-what-that-mean.2688/

We've been battling some stupidity in high places regarding the flu vaccine.

This led me today to look at some of the CDC's published information on the flu vaccine and I came across this CDC presentation regarding the effectiveness of last year's (2014-2015) flu vaccine. It was a bad year for the flu vaccine since the CDC only claimed 23% vaccine effectiveness, but after looking at the numbers I think that even this paltry result is an outright fraudulent claim.
http://www.cdc.gov/vaccines/acip/meetings/downloads/slides-2015-06/flu-02-flannery.pdf
http://www.cdc.gov/flu/fluvaxview/coverage-1415estimates.htm#age-group-adults

Some interesting things I noted:
  1. The vaccination rate of the enrollees (group of people with flu-like symptoms) was 53% which was higher than the national average vaccination rate of 47.1%.
  2. The vaccination rate of the group with confirmed flu was 49% which was higher than the national average of 47.1%.
  3. The vaccination rate of the enrollees (group of people with flu-like symptoms) of 18+ years was 59% which is higher than the national average for this age group of 43.6%.
  4. The vaccination rate of the group with confirmed flu and 18+ age was 57% which is higher than the national average for this age group of 43.6%.
  5. The vaccination rate of the enrollees (those with flu like symptoms) and 18-49 age was 44% which is higher than the national average for this age group of 33.5%.
  6. The vaccination rate of the group with confirmed flu and 18-49 age was 43% which is higher than the national average for this age group of 33.5%
  7. Since vaccination rates among the enrollees and those with confirmed flu are higher than the national averages, the CDC's own data as it is presented indicates that the vaccine has negative effectiveness.
  8. Over 80% of the flu viruses tested were antigenically different than the vaccine.
  9. 378 enrollees seem to have disappeared without explanation.
  10. This is not a peer-reviewed published study.
  11. A note is repeatedly displayed: "Adjustment for study site, age, sex, race/Hispanic ethnicity, self-rated general health, days from illness onset to enrollment, and calendar time (2-week intervals)," but no explanation is ever given for how or on what basis these mysterious adjustments were made.
  12. It is not clear how "adjusted OR" is calculated in their effectiveness formula: VE = (1 – adjusted OR) x 100%
  13. Therefore it is not clear how the 6% higher average rate of vaccination in those with flu-like disease vs. the confirmed flu is transmogrified into the claimed 23% effectiveness.
So does anyone with more knowledge of statistical magic than me know how the CDC transmogrifies this data into the claim of "23% vaccine effectiveness"?
Click to expand...

It does so by making direct comparisons instead of the indirect comparisons you made (no surprise - indirect comparisons tend not to be reliable or valid, so the results from indirect comparisons often differ from direct comparisons).

This is a case-control study, where the comparison is between the likelihood of vaccination in those who had influenza vs. the likelihood of vaccination in those who did not (amongst the at-risk population - in this case, people who were exposed and susceptible to respiratory viruses). It was found that those who had influenza were less likely (by 23%) to have been vaccinated. If the vaccine was ineffective, the likelihood would be the same. If the vaccine had "negative effectiveness", then those who had influenza would have been more likely to have been vaccinated.

Linda
 
fls said:
This is a case-control study, where the comparison is between the likelihood of vaccination in those who had influenza vs. the likelihood of vaccination in those who did not (amongst the at-risk population - in this case, people who were exposed and susceptible to respiratory viruses). It was found that those who had influenza were less likely (by 23%) to have been vaccinated. If the vaccine was ineffective, the likelihood would be the same. If the vaccine had "negative effectiveness", then those who had influenza would have been more likely to have been vaccinated.
Click to expand...

You don't see any flaw in this methodology??? Rub those brain cells together for a minute.

All this study shows is that those with the flu vaccine are slightly more likely to develop ILI than the Flu. It doesn't necessarily show that the vaccine reduces odds of getting the flu vs. not getting sick at all.

Vaccine Effectiveness (VE) is supposed to indicate the percent reduction in risk of getting the flu if you take the vaccine vs not taking the vaccine. (i.e. if your probability of getting the flu without the vaccine is 2% and the vaccine has a VE of 50% then taking the shot would leave you with a 1% chance of getting the flu.)

Suppose the vaccination rate of those with ILI was 96% and the vaccination rate of those with confirmed flu was 90%. I think it would be more obvious that the vaccine increases the chances of getting the flu and ILI because the vaccination rate is enormously higher than the general population, but under the current logic, the CDC would still say the vaccine is effective because the odds of being vaccinated with ILI are a little higher than the odds of being vaccinated with flu.
 
Last edited:
Hurmanetar said:
You don't see any flaw in this methodology??? Rub those brain cells together for a minute.

All this study shows is that those with the flu vaccine are slightly more likely to develop ILI than the Flu. It doesn't necessarily show that the vaccine reduces odds of getting the flu vs. not getting sick at all.

Vaccine Effectiveness (VE) is supposed to indicate the percent reduction in odds of getting the flu if you take the vaccine vs not taking the vaccine. (i.e. if your odds of getting the flu without the vaccine are 2% and the vaccine has a VE of 50% then taking the shot would leave you with a 1% chance of getting the flu.)

Suppose the vaccination rate of those with ILI was 96% and the vaccination rate of those with confirmed flu was 90%. I think it would be more obvious that the vaccine increases the odds of getting the flu and ILI because the vaccination rate is enormously higher than the general population, but under the current logic, the CDC would still say the vaccine is effective because the odds of being vaccinated with ILI are a little higher than the odds of being vaccinated with flu.
Click to expand...

I think the problem is that you are comparing apples and sardines. You can do this if you like, but you are going to come to different conclusions that someone who is comparing apples and apples sprayed with pesticides.

All you can say about the comparison you are trying to make is that some people are at a higher risk of acquiring an acute respiratory illness than the general population, irrespective of vaccination status.

Linda
 
fls said:
I think the problem is that you are comparing apples and sardines. You can do this if you like, but you are going to come to different conclusions that someone who is comparing apples and apples sprayed with pesticides.
Click to expand...

This doesn't address any of my arguments.

All you can say about the comparison you are trying to make is that some people are at a higher risk of acquiring an acute respiratory illness than the general population...
Click to expand...

...and these people have a higher vaccination rate than the general population.
 
  • Like
Reactions: tim
Hurmanetar said:
This doesn't address any of my arguments.
Click to expand...

Correct. Because your arguments aren't valid. You can't draw any meaningful conclusions about vaccine efficacy by comparing the rates of vaccination and rates of flu or flu-like illness in higher risk populations with the general population.

...and these people have a higher vaccination rate than the general population.
Click to expand...

Correct. The vaccine is more strongly advised for people who are at higher risk, so it is to be expected that people proven to be at higher risk also have a higher vaccination rate.

Linda
 
fls said:
Correct. Because your arguments aren't valid. You can't draw any meaningful conclusions about vaccine efficacy by comparing the rates of vaccination and rates of flu or flu-like illness in higher risk populations with the general population.
Click to expand...

False. Because you've attempted to refute my arguments with rhetoric alone. You still haven't explained where the 23% number came from. You still haven't explained how slightly higher vaccination rate in ILI cases has any bearing on VE. I guess you accept the CDC's assumption that there is no way the vaccine could actually weaken the immune system and make people more susceptible to flu and ILI. How do you deal with my hypothetical above where vaccination rates are 96% and 90% respectively? Using the same logic the CDC would say the flu vaccine is effective. Clearly there is a serious flaw in this line of reasoning. I'm surprised you don't see it.

The vaccine is more strongly advised for people who are at higher risk, so it is to be expected that people proven to be at higher risk also have a higher vaccination rate.
Click to expand...

That's possible. It's also possible that taking the vaccine weakens your immune system making you more likely to get the flu or ILI. The old-fashioned double blind randomized controlled trial might actually tell us something, but this bullshit study by the CDC doesn't mean a thing.

Also, the largest disparity in vaccination rate (about 10% difference) was in the 18-49 age group which is the lowest risk group.
 
Hurmanetar said:
False. Because you've attempted to refute my arguments with rhetoric alone. You still haven't explained where the 23% number came from.
Click to expand...

I explained where it came from in my first post. The ratio of the odds of vaccination in the cases over the controls would have been 0.77 (in order for VE = (1-OR)x100% to equal "23%").

You still haven't explained how slightly higher vaccination rate in ILI cases has any bearing on VE.
Click to expand...

I did in my first post. The use of the case-control design is well-established. If the influenza vaccine is efficacious the VE will be positive, if it isn't efficacious, the VE will be about zero, and if there is negative efficacy, it will be negative.

I guess you accept the CDC's assumption that there is no way the vaccine could actually weaken the immune system and make people more susceptible to flu and ILI. How do you deal with my hypothetical above where vaccination rates are 96% and 90% respectively?
Click to expand...

That would represent a vaccine efficacy of 63%.

Using the same logic the CDC would say the flu vaccine is effective. Clearly there is a serious flaw in this line of reasoning. I'm surprised you don't see it.
Click to expand...

Are you sure you understand the methodology and the statistics? First you have to ask, why is there a difference in the vaccination rates in the first place, if the vaccination has no effect whatsoever on which group you might find yourself in?

That's possible. It's also possible that taking the vaccine weakens your immune system making you more likely to get the flu or ILI. The old-fashioned double blind randomized controlled trial might actually tell us something, but this bullshit study by the CDC doesn't mean a thing.
Click to expand...

Correct. If you are interested in comparing rates of ILI between vaccinated and unvaccinated groups, you need to make a different comparison. It can be done with a case-control study ("cases" will be ILI or flu and "controls" will be no ILI or flu) or a randomized-controlled trial.

Linda
 
fls said:
I explained where it came from in my first post. The ratio of the odds of vaccination in the cases over the controls would have been 0.77 (in order for VE = (1-OR)x100% to equal "23%").
Click to expand...

You said some words but you didn't explain anything.

What does OR stand for? Obviously they calculated OR to be .77, but how did they calculate that? I'm sure there's a formula, I'd just like to know what it is.

The way VE should be calculated:
VE = (ARU - ARV) / ARU * 100%
VE is Vaccine Effectiveness
ARU is Attack Rate (probability of getting infected) in Unvaccinated population
ARV is Attack Rate (probability of getting infected) in Vaccinated population

That would represent a vaccine efficacy of 63%.
Click to expand...

What formula did you use to calculate that?

So you would see no problem with this hypothetical claim of 63% VE if the vaccination rate in the sick people was 90%+ while the average vaccination rate is under 50%?

Are you sure you understand the methodology and the statistics? First you have to ask, why is there a difference in the vaccination rates in the first place, if the vaccination has no effect whatsoever on which group you might find yourself in?
Click to expand...

I don't fully understand the statistical formula used, that's why I'm asking for the formula.

Nevertheless, I'm certain that this methodology is flawed because it assumes the higher vaccination rate in the ILI group is not due to the vaccine itself. The numbers could very well indicate that the vaccine increases the chances of getting the flu and increases the chances of getting ILI even more. Our hypothetical situation above highlights the flaw in this methodology.
 
Last edited:
Hurmanetar said:
You said some words but you didn't explain anything.
Click to expand...

Sorry. I assumed some basic knowledge on your part.

What does OR stand for?
Click to expand...

Odds Ratio.

Obviously they calculated OR to be .77, but how did they calculate that?
Click to expand...

By using the standard formal for Odds Ratios.

I'm sure there's a formula, I'd just like to know what it is.
Click to expand...

(Odds of being vaccinated in the "case" group)/(Odds of being vaccinated in the "control" group)

\What formula did you use to calculate that?
Click to expand...

The Odds Ratio formula.

So you would see no problem with this hypothetical claim of 63% VE if the vaccination rate in the sick people was 90%+ while the average vaccination rate is under 50%?
Click to expand...

No, because I have the wherewithal to understand that no meaningful conclusions can be drawn from a comparison with the average vaccination rate.

Nevertheless, I'm certain that this methodology is flawed because it assumes the higher vaccination rate in the ILI group is not due to the vaccine itself. The numbers could very well indicate that the vaccine increases the chances of getting the flu and increases the chances of getting ILI even more. Our hypothetical situation above highlights the flaw in this methodology.
Click to expand...

So you don't know what a case-control study or odds ratios are, but you feel comfortable telling people who have the knowledge and experience to understand the data that their conclusions are fatally flawed? Is it possible that it doesn't make sense to you because you don't know enough, rather than it doesn't make sense because we're saying something stupid? I'm not trying to be mean or insulting, I'm just trying to understand the thought process here.

Linda
 
fls said:
Odds Ratio.

By using the standard formal for Odds Ratios.

(Odds of being vaccinated in the "case" group)/(Odds of being vaccinated in the "control" group)
Click to expand...

At last, a formula!

49/55 equals .89 not .77

90/96 equals .94 not .37

What am I doing wrong?

No, because I have the wherewithal to understand that no meaningful conclusions can be drawn from a comparison with the average vaccination rate.
Click to expand...

Well then you're an idiot.
 
fls said:
LOL. They don't have google where you live?

Linda
Click to expand...

If 49% of the case group was vaccinated, then the odds of being vaccinated in the case group are 49%, right? If 55% of the control group was vaccinated, then the odds of being vaccinated in the control group are 55%, right? Then according to your formula for OR, we would take 49/55 and get 89%, right? Tell me what I did wrong (if anything) if you actually know what you're talking about.
 
Hurmanetar said:
If 49% of the case group was vaccinated, then the odds of being vaccinated in the case group are 49%, right? If 55% of the control group was vaccinated, then the odds of being vaccinated in the control group are 55%, right? Then according to your formula for OR, we would take 49/55 and get 89%, right? Tell me what I did wrong (if anything) if you actually know what you're talking about.
Click to expand...

This blog post explains how the CDC calculates VE: http://haicontroversies.blogspot.ca/2013/01/whats-up-with-cdc-influenza-vaccine.html

Applying the formula to the stats in your link here's what I think the calculation is:

Vaccinated with Flu: 1097 (a)
Vaccinated no Flu: 3866 (b)
Unvaccinated with Flu: 1140 (c)
Unvaccinated no Flu: 3226 (d)

(VE) = (1-OR)*100 or (1-ad/bc)*100

OR=(1097*3226)/(3866*1140)=0.80

VE=(1-0.8)*100
VE=20% (which they adjusted to 23%)
 
You must log in or register to reply here.
Back
Top