I'm busy working on my blog posts. Watch this space!
CDC recommendations: Making Absolute Declarations using Relative Risks/Benefits
February 19, 2015
When CDC director Tom Frieden stated unequivocally that neuraminidase inhibitors (NIs), such as Tamiflu, should be given to all elderly people with upper respiratory infection symptoms even before they have documented flu because NIs save lives, prevent hospitalizations, and curb disease spread in nursing homes, we were surprised. Having read the Cochrane report on NIs, it seemed clear to us that at best Tamiflu was ineffective, and at worst it could cause more harm than good.
We wrote to the CDC, expressing our concerns, and received a very lengthy response two weeks later. Essentially, the CDC responder stated that the CDC felt that Cochrane excluded too many studies from its analysis, and that a newer meta review by Joanna Dobson (which we later learned was sponsored by Roche) showed more promising results. But these results only reached significance in those with proven flu (not URI symptoms, as the CDC director stated should be treated), and the absolute numbers were small. 1/100 people who received Tamiflu avoided hospitalization in this study when compared to those not taking Tamiflu, and 4/100 avoided antibiotics. Meanwhile, almost 5/100 developed serious vomiting on Tamiflu compared to placebo. The number of people who derived benefit from NI's in this newer metaanalysis was far higher than the one done by Cochrane
When scientists at Cochrane re-examined this new data, they were not convinced of its conclusions, and in fact still questioned the inclusion of certain studies in the analysis that they believed were not reliable. They believed that, when the data was looked at carefully in the new study, only 3/1000 people taking Tamiflu avoided hospitalization even if one assumed that the analysis used only valid data. The CDC has remained unwavering in its advocacy for wide-spread use of NIs, despite the evident controversy. When we asked the CDC for their assessment of absolute risk and benefit of NIs in preventing hospitalization and death, they did not respond.
In fact, the CDC rarely offers absolute numbers to back its recommendations. Although the CDC did not make any recommendation regarding the high dose flu shot vs. regular dose shot, its website did cite a 24% reduction in flu with the higher dose shot among the elderly. That number, of course, is a meaningless relative benefit. In fact, when absolute numbers are looked at, only 5/1000 people taking high dose flu shots avoided influenza when compared to those who received the low dose shot, while 80-90/1000 developed serious side effects from both vaccines. These numbers were not mentioned by the CDC, despite their significance for doctors and patients. Simply stating that there is a 24% improvement is deceptive.
Prior to the Tamiflu controversy, the CDC recommended that people over 65 receive both Pneumovax and the newer (and more expensive) Prenvar 13 pneumonia vaccine. Their recommendation was based on a Pfizer-sponsored study called CAPiTA that showed a 45% reduction in pneumonia among elders who used Prenvar 13 vs. those who used placebo. Again, relaying data in terms of relative risk is meaningless. We asked the CDC what the absolute benefit of Prenvar 13 was compared to placebo, and why the CDC was recommending both vaccines when CAPiTA did not look at that endpoint. After many emails, we finally received an answer. The absolute benefit of Prenvar 13, by their calculation, was that 5 out of 10,000 people avoid getting pneumonia who receive the shot compared to those who get no shot. There is no good data on death, side effects, or hospital prevention. Nor is there any data on use of both vaccines, although the CDC suggested theoretical reasons for advocating both.
The CDC has enormous impact on determining how doctors, patients, health departments, and institutions act regarding use of immunizations and infectious treatments. By providing faulty data couched in misleading relative risk terminology, by using drug-sponsored studies to arrive at conclusions that are questioned by other very respected groups such as Cochrane, and by making recommendations (such as treating people with uri symptoms with Tamiflu, and recommending using two pneumonia vaccines in the elderly) not supported by any data, the CDC has lost a great deal of credibility. We have to count on our government institutions to provide accurate and sensible information conveyed in an absolute risk/benefit format, and to make declarative statements only when the data is both conclusive and clinically meaningful. It is easy to understand why so many Americans are skeptical about what the CDC is stating–even about measles vaccine–given their track record of providing misleading information. It is time we implore our government agencies to be more forthcoming when making such weighty policy declarations., and to present all date accurately.