Posted: November 5, 2011
This page is categorized as:    link to Pandemic and Other Infectious Diseases index
Hover here for
Article SummaryOn October 25, 2011, a team led by Michael Osterholm of the University of Minnesota published a meta-analysis of prior research on the effectiveness of the flu vaccine, showing it to be less effective than public health officials and experts have usually claimed. In the resulting media coverage, many in public health said the Osterholm paper wasn’t really surprising and denied that flu vaccine effectiveness has been routinely oversold. So Jody Lanard and I made a case that it was still being oversold, focusing particularly on two very recent updates on the CDC website, and emailed it to Robert Roos of CIDRAP News. Bob interviewed public health professionals about what we said and put together a November 4 story called “Flu vaccine efficacy: Time to revise public messages?” There’s no question mark in the title we’re giving our email: “Overselling Flu Vaccine Effectiveness Risks Undermining Public Health Credibility.”

Overselling Flu Vaccine Effectiveness Risks Undermining Public Health Credibility


Robert Roos’s CIDRAP News article is also online.

The Osterholm et al. article in Lancet Infectious Diseases provides ample evidence that the flu vaccine is less effective than we all wish – and less effective than most patients and some physicians have believed.

A point the paper doesn’t address is the distinction between asserting that the flu vaccine is less effective than flu experts previously thought and asserting that the flu vaccine is less effective than flu experts previously acknowledged.

Here’s the question: Have the flu/infectious diseases/vaccination/public health establishments known the comparatively low effectiveness of the flu vaccine for some time, but intentionally downplayed that information? Or were these establishments genuinely unaware of this reality, and surprised by what the Osterholm et al. meta-analysis revealed?

It’s pretty clear to us that the former is closer to the truth than the latter.

Over the years, we have documented a number of flu vaccine exaggerations on The Peter M. Sandman Risk Communication Website and elsewhere. (See for example “Convincing Health Care Workers to Get a Flu Shot … Without the Hype,” which we wrote in January 2009.) But let us add in this email one very recent example – not from some expert’s brief, hurried comment to a reporter, but from the CDC’s carefully reviewed website:

On October 12, 2011 the CDC updated its “Flu Vaccine Effectiveness: Questions and Answers for Health Professionals” website page. Perhaps in anticipation of the Lancet I.D. paper, the update estimates flu vaccine efficacy in reducing lab-confirmed influenza in healthy adults under 65 (in a year with a good match) at 50–70%. Its previous estimate had been 70–90%. (The Q&A for health professionals used the 70–90% figure as recently as June 23, 2011; we don’t have proof that it used it right up until October 12, 2011.) The three RCT studies cited for the new estimate were published in 2009, 2009, and early 2010 – so the “new” information has been available at least since then.

The October 12 update does not mention that the 50–70% estimate is a departure from what the CDC has been saying for years. No CDC news release accompanied the update and heralded this significant change. (We don’t know what effort if any the CDC has made to alert local and state health departments to its downgraded assessment of the vaccine’s efficacy.)

The 70–90% estimate also appeared in an evergreen flu vax Q&A for the general public (as opposed to the one directed at health professionals) from as early as January 11, 2007.

To its credit – partial credit only for this! – when the CDC updated this page on February 25, 2011, it removed the 70–90% estimate. Instead of offering a new estimate, the February 25 revision said: “CDC is currently reviewing recently published studies on VE [vaccine effectiveness] to update existing estimates.” (See http://web.archive.org/web/20110623070217/http://www.cdc.gov/flu/about/qa/vaccineeffect.htm.)

And then on October 13, 2011, a day after updating the Q&A for health professionals, the CDC updated this evergreen Q&A for non-professionals. The new one is at http://www.cdc.gov/flu/about/qa/vaccineeffect.htm. Unlike the previous day’s update for professionals, the update for the general public does not use the 50–70% figure. It avoids specifying a numerical estimate of flu vaccine effectiveness or efficacy in the overall public, in healthy adults, or in the elderly.

The October 13 update for general readers does give a numerical estimate for the effectiveness of the Live Attenuated Influenza Vaccine in children 2 and older – the age group for whom the live vaccine is most effective. It also mentions overall lower effectiveness data than usually acknowledged for the 2010–2011 season – “60% for all age groups combined” (from “unpublished CDC data”).

The update for the general public also discusses one of the three studies cited in the previous day’s update for health professionals – the 2009 Monto et al. study, which showed that in the 2007-2008 season inactivated flu vaccine prevented lab-confirmed influenza in 70% of healthy adults aged 18–49. Ignoring the study’s inapplicability to adults over 49, or to anyone in a high-risk group, the October 13 update says simply (and misleadingly):

A randomized study (by Monto et al link is to a PDF file) looking at the 2007–2008 influenza season found trivalent inactivated vaccine (flu shot) protected 7 out of 10 people from influenza illness.

It is also worth noting that the mean age of Monto et al.’s study population was 23.3 years (see Table 1 of the study) – a cohort very much skewed toward the younger end of the 18–49 range.

Finally, we wonder why the October 13 update for general readers states explicitly that “Studies show that LAIV works about as well as the flu shot.” The Monto et al. study cited in that very paragraph provides data showing differential benefits between the LAIV and the TIV in various age groups, and the Osterholm et al. meta-analysis found no adequate RCT data on the LAIV for children over 7 or healthy adults under 50. Other evidence apparently suggests that the LAIV and the TIV have different efficacy profiles in different age groups.

That said, “averaging” the various studies to reach an over-generalized conclusion that the LAIV and the TIV are equivalent does not constitute hyping flu vaccine efficacy, as the other points we have made here do.

Bottom lines from this example:

  • On October 12, with the Lancet I.D. paper about to be published, the CDC finally changed its flu vaccine effectiveness estimate for healthy adults under 65 from 70–90% to 50–70% in its Q&A for health professionals – years after the weight of the evidence suggested 70–90% was too high. It did nothing to underline the new number, and didn’t mention that it was a lot lower than the number it had been using for years.
  • On October 13, the CDC updated its Q&A for non-professionals (having dropped the 70–90% estimate from this page in February). Instead of using the 50–70% estimate, it used no estimate. It misleadingly described a study that showed 70% effectiveness in healthy adults 18–49 as applying to “7 out of 10 people” (presumably including the elderly and those in high-risk groups).

The underlying question, of course, is this: Insofar as public health experts and officials (at the CDC and elsewhere) have been knowingly exaggerating the effectiveness of the flu vaccine, how much trust should people have in official claims about the effectiveness and safety of other vaccines?

And the underlying tragedy is this: Insofar as the vast majority of vaccine claims (especially vaccine safety claims) are honest and trustworthy, then it is a huge self-inflicted wound for public health to have exaggerated the effectiveness of the flu vaccine.

Some of the news stories about the Osterholm et al. study (including one of yours) raises the concern that people who would otherwise have gotten vaccinated against the flu might decide differently once they learn that the flu vaccine works on average only 59% of the time (or 50–70%, if you prefer) on healthy adults under 65. That may turn out to be an issue, and it will need to be addressed with candid and humble communication.

But the more frightening issue, it seems to us, is whether people who learn that the flu vaccination establishment has been systematically hyping flu vaccine effectiveness might lose some of their confidence in the flu vaccine, in other vaccines, and in public health more generally.

We can understand everybody’s reluctance to address this latter concern. None of us wants to give aid and comfort to anti-vaccination activists and further undermine vaccine uptake. We certainly don’t. We are convinced, as are Osterholm and his coauthors, that the current flu vaccine is enormously preferable to none at all – as are all or nearly all the other vaccines in common use.

But a core principle of risk communication is that prior dishonesty must be contritely acknowledged, and that accountability mechanisms must be strengthened in response in order to help prevent recurrences.

In the news coverage of the Lancet I.D. paper, flu experts at the CDC and elsewhere have generally acknowledge the validity of the paper’s core conclusions. But most have taken the line that those conclusions are not news – not divergent from what the experts have believed in recent years and not divergent from what the experts have been saying in recent years. For example, the CDC’s list of “Influenza Key Points” regarding the paper, distributed to reporters and others, says: “The findings from this review are generally consistent with work on vaccine effects that has been presented and published by CDC and others in the last few years.”

This is far superior to claiming – as we feared might be claimed – that the paper is biased or inaccurate and presents an unfairly pessimistic picture of flu vaccine efficacy. But it is a far cry from conceding the demonstrable truth that the Osterholm et al. findings are surprising to many patients, local docs, and even local public health officials precisely because the U.S. public health establishment has (until now) generally avoided acknowledging the disappointing effectiveness data.

The repeated claim that the Osterholm et al. conclusions are “just about what we’ve been telling people” smacks of Orwell: “We have always been at war with Oceania.”

On October 25, for example, the CDC’s Joe Bresee, Chief of the Epidemiology and Prevention Branch of the Influenza Division, told Minnesota Public Radio that the CDC has been open with the public about the limitations of the flu vaccine and has not oversold it. (This is the reporter’s paraphrase, not a direct quotation.) On the same day, Dr. Bresee told you: “I think the authors have done a nice service in putting all the findings together in one place. But it isn’t surprising to us.”

But in addition to the CDC’s flu vaccine update for health professionals, which continued to claim 70 to 90% efficacy until October 2011, Dr. Bresee was on the list of top CDC influenza experts who signed onto an August 2010 CDC Morbidity and Mortality Weekly Report (MMWR) entitled “Recommendations of the Advisory Committee on Immunization Practices (ACIP), 2010.” The MMWR asserts:

When the vaccine and circulating viruses are antigenically similar, TIV prevents laboratory-confirmed influenza illness among approximately 70%–90% of healthy adults aged <65 years in randomized controlled trials (77,80,185–187).

This was in August 2010, remember, after the three studies the CDC now cites as support for its downgraded 50–70% estimate had been published.

Peter’s corporate clients are periodically caught having suppressed or played down some piece of technical information – and then they want to suppress or play down the information about the prior suppression. It doesn’t work for them. It may work a bit better for public health; both journalists and the public are more tolerant of public-spirited dishonesty than of self-interested dishonesty.

But in the long term, it won’t work for public health either – as the sadly growing ranks of vaccine abstainers demonstrates. You can’t build trust on such a shaky foundation.

Note: Peter Sandman is Deputy Editor of CIDRAP Business Source, and a member of the Expert Advisory Group of an ongoing foundation-funded study of flu vaccination issues and problems. Michael Osterholm heads both of these efforts. We didn’t say this in our email to Robert Roos because he already knew it, but other readers deserve to know it too, since it may be seen as a conflict of interest.

Copyright © 2011 by Peter M. Sandman and Jody Lanard

For more on infectious diseases risk communication:    link to Pandemic and Other Infectious Diseases index
      Comment or Ask      Read the comments
Contact information page:    Peter M. Sandman

Website design and management provided by SnowTao Editing Services.