Healthy Skepticism International News

April 2005

The new alchemy: Mixing doctors and journalists to spin gold

Prominent medical journalist, Jeanne Lenzer, has made available to Healthy Skepticism this important piece on the relationship between journalists and pharma.

Abstract

The chasm between raw research data and published analyses singing its praises is sometimes a yawning one. A number of researchers and writers have sounded the warning bell about how and why such distortions occur in both the medical literature and lay media.[1-7]

Shuchman and Wilkes warned “medical scientists alone cannot correct the deficiencies of medical news coverage, but neither can journalists…the underlying problem is an interactive dynamic to which both parties contribute.”[4]

We also need to acknowledge the role in this drama of the invisible third party: industry.

A recent example of inflated claims the media made for one drug, based on a study reported in The Lancet, illustrates the problem. The glowing reports reflect not only media vulnerabilities, but also the way the hidden hand of industry exploits those vulnerabilities, through prepublication news releases that amplify optimistic claims made by researchers in scientific journals, while minimizing or ignoring risks and failing to report scientific dissent and financial conflicts of interest.

The publicity

In March of 2004, the mainstream media erupted in enthusiastic coverage of an optimistic analysis touting the benefits of tPA for stroke. Much of the coverage was based, wholly or in part, on a glowing prepublication news release by MCS, a public relations (PR) firm that works for Genentech, the manufacturer of tPA.

The release was written in such a way that obscured the fact that it was written by a PR firm, especially since the attachments were written by the American Heart Association. In fact, the only tip-off in the release was that the contact person’s email address was xxx@mcspr – and the letters “pr” suggested to this writer that the origin of the news release was indeed a PR firm.

The results of the PR effort were predictable. HealthDayNews, a major news service that serves media giants such as Gannett, Cox Newspapers, and Discovery.com, posted an article entitled “Faster Stroke Treatment Saves Lives.”

However, contrary to the exuberant claim made in the HealthDay article that tPA “saves lives,” not one single study to date has shown that tPA saves lives in the setting of acute stroke – even when tPA is given within the currently recommended three-hour window. In fact, virtually all randomized controlled trials (RCTs) show quite the opposite: stroke patients given tPA are up to three-fold more likely to die than patients given placebo when treatment was initiated within 0-3 and 0-6 hours[8-11] and only one RCT found no difference in the death rate.[12]

How did the headlines come to diverge so profoundly from the raw data? To understand that, it is necessary to examine what is known as “checkbook science.”[13]

How it happened

At 9:56 A.M., the morning of March 4th, journalists received an electronic media packet from MCS, a public relations firm for Genentech, announcing the results of the to-be-published Lancet study along with the claim that rapid treatment in stroke centers reduces “death and disability.” The National Institutes of Health (NIH) followed with a news release at 6:30 p.m. on March 4th. The NIH release, “Early Treatment Confirmed as Key to Stroke Recovery,” quoted NIH Director Elias A. Zerhouni, M.D., as saying, “This scientific work is a good example of a cooperative effort between the Federal Government and the pharmaceutical industry.”

Joe Santangelo, editorial and press assistant at The Lancet, said his journal did not issue a prepublication news release and that any prepublication releases could only have come about through the cooperation of at least one of the study authors.

The HealthDay article was posted at 7:00 P.M. on March 4th, two days before the study itself was published in The Lancet on March 6th.[14] The news article cited The Lancet study as one of its sources. It did not mention the MCS news release.

The news article kicked off a flurry of articles, all titled, “Faster Stroke Treatment Saves Lives.” The news feed was picked up by a number of major outlets including Forbes and DrKoop.com. The article did not mention studies reporting increased death rates among patients treated with tPA, nor did it acknowledge the substantial scientific dissent regarding tPA for stroke[15-23] – dissent so intense that one author called the controversy “unprecedented” in his experience.[24] The article did not quote any experts representing any of the several national professional organizations that have issued statements cautioning against the use of tPA for stroke outside of clinical trials – even when used within the recommended 3-hour window.[19;23]

Readers were not informed that the study co-author, Gregory Albers, M.D., interviewed in the HealthDay article, had financial ties to Genentech, the manufacturer of tPA. Nor were they told that a number of the study authors and statisticians had financial ties to either Genentech or Genentech’s European marketing partner for tPA, Boehringer Ingelheim.[14]

Dr Albers is quoted in the HealthDay article saying, “(W)e found that the benefits of tPA extend beyond three hours, though there’s quite a bit of variability [in outcomes] once you get past the three-hour window.” Subsequent articles appeared with such titles as, “Stroke treatment ‘window’ stays open longer than previously thought.”

The study

The study reported in The Lancet was a pooled analysis of previously reported randomized controlled trials of tPA for stroke, including the National Institute of Neurological Disorders and Stroke (NINDS) trials. The authors reported that tPA “improved outcomes in [acute stroke] in previous trials” and that the pooled data “suggest that the benefit of rt-PA could extend beyond 3h…” with the mild caution that, “this potential might come with some risks.”

Dr. Albers blamed the media for distorting his study’s findings and contends that he and his co-authors never claimed that tPA saves lives. He told me in a phone interview that after the mainstream media articles appeared, some doctors did call him with concerns about the media claims and asking whether it was true that tPA could be given after three hours.

Although the media can be faulted for its kid-glove treatment of the study and its prepublication news releases, study authors who cooperate with their industry liaisons in the production of prepublication releases must also shoulder some blame. How did journalists (and many doctors) get the idea that tPA “saves lives”? The problem can be traced, at least in part, to the enthusiastic use of the composite endpoint, “death and disability,” by those with vested interests in the use of tPA.

Many doctors and journalists have assumed that the claim, “tPA reduces death and disability” could be interpreted to mean tPA reduces death and it reduces disability. However, it means no such thing. It only means that when the two endpoints are tallied together, if deaths remain the same and disability decreases even slightly, then the claim can be made that tPA “reduces death and disability.” Even if deaths increase, the claim can still be made that “death and disability is reduced” if overall disability among the survivors is reduced.

Using composite endpoints is a “brilliant marketing strategy,” according to Dr. David Brown, assistant professor of medicine at Albert Einstein College of Medicine, who says that as long as one of several endpoints is positive, companies can work in any other endpoint, (tPA grows hair; tPA reduces impotence) and they are still technically correct.

The use of composite endpoints is not well understood by most reporters, said Andrew Holtz, former president of the American Association of Health Care Journalists. “Many news reporters and wire [services] don’t know these tricks,” Holtz says.

This “trick” was readily exploited by the American Heart Association (AHA) which repeatedly claimed that tPA for stroke “saved lives” and promoted tPA for stroke to a Class I (definitely recommended) status following publication of the NINDs trial. The AHA only withdrew its claim that tPA “saves lives” after public scrutiny in which it was pointed out that no study has ever shown that tPA reduces mortality and that the AHA received $11 million from the manufacturer of tPA in the decade prior to its recommendation.[19;25]

If such a basic concept as composite endpoints can be so readily exploited in order to obfuscate rather than clarify, it should come as little surprise that more sophisticated analyses can lose readers – and, in the process, obscure the truth.

The authors of The Lancet article used a far more complex, and unexplained, statistical tool that appears to have conflated the results of several largely negative trials into one, great big positive trial: they introduced the use of a new “global statistic” in which they amalgamated patient scores from five different neurological assessment scales.

Dr. Jerome Hoffman, professor of emergency medicine at UCLA, the sole dissenting member of the American Heart Association expert panel that recommended tPA for stroke, is critical of The Lancet study’s use of a devised “global statistic,” saying the statistic has not been validated or even clearly defined in the article or any of its citations.

But Dr Albers defends the authors’ use of the global statistic saying that the studies from which the data were pooled used five different measurement scales and not all the scales were used in each of the studies analyzed. He added that the global statistic was just one of the endpoints the authors evaluated.

Dr Hoffman cautions, “Before you use the global statistic you have to decide if it’s valid or meaningful. Why did they pick it? How have they resolved issues of clinical versus statistical meaning?” Hoffman adds, “The Lancet study is a rehash of old information that applies a different set of statistics to a negative outcome to get a positive outcome. You can’t do an experiment under one set of conditions and choose one endpoint, or set of endpoints, and then just decide to reanalyze with a different set of endpoints. At best, what they are doing is raising a hypothesis. To imply anything else would be PR and shouldn’t be confused with science.”

What is to be done?

As one “promising new treatment” after another is promoted, only to be later abandoned as worthless, or worse, it is useful to examine how such therapies come to be embraced while warning signals are ignored. The rush of prepublication releases plays an important role in sidelining scientific dissent since the results of a study become “old news” within a matter of hours to days.

The gold was spun even before The Lancet study appeared in print – as the result of an intricate set of relationships among study authors, the drug manufacturer, a PR firm, and journalists who failed to ferret out scientific dissent or financial conflicts of interest.

Phil B. Fontanarosa, M.D., Executive Deputy Editor of the Journal of the American Medical Association, wrote that journal editors “generally discourage prepublication release so that the entire study, with full methods and results as well as proper caveats, can be appropriately evaluated and interpreted.”[26] Mr Santangelo told me in a phone interview that The Lancet discourages early reporting saying, “The reason our material is embargoed is so journalists can get it right – not get it first.”

If we are truly interested in science over advertising or alchemy, we will have to discourage researchers from promoting their work through prepublication releases – and journalists will have to stop acting as lapdogs for the scientific community and start acting as watchdogs.

View/Hide References

HS Int News index

Page views since 15 March 2010: 8810

Comments

Our members can see and make comments on this page.