Marcia Angell summarizes the many negative influences that pharmaceutical-industry marketing now exerts over medicine. She appropriately focuses attention less on industry wrongdoing and more on how medicine itself acquiesces in a system that threatens practitioners’ professional commitment to patients’ well-being.

Patients should be able to trust their physicians and not have to wonder whether the physician recommends a treatment simply because she was taken out to dinner or had her way paid to a conference at a luxury resort by the company. But it is even more important that medical research, on which all physicians must rely in order to know what treatments to recommend, retains its essential integrity.

At stake are not just individual outcomes, but large public-health issues.

In the early autumn of 2009, H1N1 influenza, so-called swine flu, was threatening to cause major outbreaks in many countries, including the United States. In one way, the H1N1 was eerily like the 1918 flu pandemic that killed millions of people—early evidence suggested that young and otherwise healthy people might be especially vulnerable. An effective vaccine was being rushed into production but was not yet available in sufficient quantity. Standard antibiotics are no good for treating influenza, and no drug regimen has been discovered that reverses the severe lung inflammation that occurs in serious cases.

The medicine in which many placed their faith was oseltamivir, sold under the brand name Tamiflu. Tamiflu was thought to have two effects. If you came down with a garden-variety case of the flu, and started taking Tamiflu right away, the medicine slightly shortened the course of the disease. That was a good thing for the flu sufferer, but not a reason for public-health agencies to stockpile large supplies of the drug. The other effect was to provide some protection in cases severe enough to require hospitalization. It was believed that in those cases, Tamiflu might prevent the worst complications and aid the patient’s recovery. Governments spent millions of dollars to lay aside stores of the drug, which is manufactured by Roche.

Faith in Tamiflu was bolstered by a highly respected source of medical evidence, the Cochrane Collaboration. This international, noncommercial network of academic investigators conducts systematic reviews on the effectiveness of medical treatments, employing rigorous standards. In 2005 Tom Jefferson, a researcher with Cochrane, and his team conducted a review of Tamiflu and concluded that the evidence supported its use for prevention of complications from influenza.

But a Japanese investigator questioned the Cochrane review. He noted that Jefferson cited positive results contained in an earlier “meta-analysis” of Tamiflu data, but the original studies that made up that review were not publicly available. How could the Cochrane team be so sure that Tamiflu worked when they had not examined the original data?

The Cochrane reviewers accepted the criticism. First they approached the academic physicians who were said to have conducted the studies. The physicians said they did not have the actual study results and referred the team to Roche. (At least one of these investigators denied having conducted the study that was attributed to him.) Roche first refused to release any of the study data and then provided only a partial summary. Roche did eventually release data that came from a different set of studies, but contrary to the company’s assertion, those data did not show that Tamiflu substantially prevented flu complications. Meanwhile, the British Medical Journal (BMJ) partnered with investigative reporters at Britain’s Channel 4 News to get to the bottom of it all.

In the end Jefferson and his team issued a revised assessment of Tamiflu, which was published online by BMJ on December 8, 2009, along with an account of the BMJ/Channel 4 investigation of Roche’s withholding of data. The evidence simply did not exist to support Tamifulu’s usefulness in serious influenza cases. That meant that public-health authorities around the world had spent billions of dollars stockpiling a drug without (in hindsight) any scientific proof of its value.

Perhaps the data that Roche refused to provide would, if revealed, demonstrate the effectiveness of Tamiflu. But why then did the company refuse to reveal it? It is far more likely that Roche selectively concealed data that failed to support their drug marketing, and loudly publicized any remaining data that affirmed it.

In an editorial, the BMJ editors stressed that from a legal and regulatory point of view, Roche did nothing wrong. It broke no rules. It acted precisely as other drug companies act. That is why, the editors added, we need radical reforms to free medical research from the vice-grip of commercial marketing. We need to make sure that in the future, physicians and public health officials know the scientific facts when they decide whether to recommend a drug.

The Tamiflu case is hardly the first of its kind. In the late 1950s the U.S. Senate considered legislation to strengthen the Food and Drug Administration’s power to regulate drug firms. A star witness of the committee pressing for the new law was Dr. Dale Console. Formerly medical director at the drug company Squibb, Console had turned into a staunch industry critic.

A pro-industry senator tried to get Console to admit that it was reasonable that drug firms charged high prices for their products. Isn’t it true, the senator prodded, that companies must spend huge sums on researching new drugs, only to find that few work?

That’s true, Console instantly replied, as that’s the very essence of science. But, he went on, the problem with the drug companies is “that they market so many of their failures.”

Console’s comment from 50 years ago haunts us today. We’d like to think we have made major advances in pharmaceutical science and that the new drugs coming onto the market are highly effective. But then we have to look at the Tamiflu story and ask, until something changes for the better, how can we know for sure?