Wednesday, September 26, 2007

Haunted Health Care: the Scope of Ghost Management

An important new article in PLoS Medicine expanded thinking about the involvement of pharmaceutical companies (and possibly other health care corporations) in the shaping of clinical research. [Sismondo S. Ghost management: how much of the medical literature is shaped behind the scenes by industry? PLoS Med 4(9): e286 doi:10.1371/journal.pmed.0040286]

Sismondo defined ghost management of medical research and publishing:

when pharmaceutical companies and their agents control or shape multiple steps in the research, analysis, writing, and publication of articles. Such articles are 'ghostly' because signs of their actual production are largely invisible—academic authors whose names appear at the tops of ghost-managed articles give corporate research a veneer of independence and credibility. They are 'managed' because those companies shape the eventual message conveyed by the article or by a suite of articles.

Sismondo thus went beyond the issue of ghost-writing, the contribution of unnamed authors who may be influenced, or hired, by pharmaceutical companies or other organizations with vested interests. (See recent post here.) He also went beyond the issue of how pharmaceutical companies and other research sponsors may attempt to influence the design and implementation of research studies, the analysis of their results, and/or how the results are reported. (See this post.)

As Sismondo noted,

It has been repeatedly and firmly established that pharmaceutical company funding strongly biases published results in favor of the company's products. Ghost management amplifies that bias, because when one set of commercial interests exerts influence at multiple stages of research, writing, and publication, it will shape the resulting article. In turn, bias affects medical opinion and practice, and ultimately, patients.

Sismondo also attempted to estimate the prevalence of ghost management. This would have been predicted to be a difficult task, because,

it is not in the interests of writers, authors, or sponsors and their agents to reveal ghost management processes; hence a number of the published accounts of ghost management have stemmed from legal proceedings and investigative journalism.

So he set out to review existing evidence and find some new evidence.

First he noted the article by Healey and Cattall, based on documents produced during legal proceedings, that explained how Pfizer Inc ghost managed publications of multiple articles on the drug sertraline [Zoloft], producing 85 manuscripts, and accounting for an important part of the literature on this one drug. (See previous post here.)

An important contribution by Sismondo was what he called his "supply-side analysis."

A survey in 2001 identified 182 MECCs in the United States, up from 153 in 1998. A number specialize in producing, placing, and tracking journal articles, known in the trade as 'publication planning' or 'strategic communication planning.' While these firms hide details of their work—from potential critics and competitors—they also energetically promote themselves and their services.

I spent six hours searching web pages for MECCs offering publication planning or similar or overlapping services to the pharmaceutical industry, and found 23 (list available from the author). This is not an estimate of the number of such firms, but indicates how common they are. There may be many more firms providing publication planning, including some not uncovered in this search, and some not advertising these services on the Internet.

In a primer on publication planning, the director of one MECC defines the activity as: 'gaining product adoption and usage through the systematic, planned dissemination of key messages and data to appropriate target audiences at the optimum time using the most effective communication channels'.

Complete Healthcare Communications (CHC) claims on its banner that it 'has honed the systems and skills needed to develop the intellectual heart of pharmaceutical marketing—the publication plan. The result for your product? A continuum of awareness, interest, and prescriber confidence.'

CHC includes among its clients Pfizer, Sanofi-Aventis, Ortho Biotech, Wyeth, Schering-Plough, Shire, AstraZeneca, and other pharmaceutical companies.

Other agencies offer very similar services. As described in an article by three of its managers, the Medical Knowledge Group starts publication planning with a phase of exploring 'key messages' and 'author/journal options' before designing any publications to incorporate those messages

Another MECC, Envision Pharma, says that 'data generated from clinical trials programs are the most powerful marketing tools available to a pharmaceutical company.' Envision will work from early on in the process to ensure 'consistent message dissemination,' will plan and track the 'data dissemination plan,' and will produce 'scientifically accurate, commercially focused abstracts, posters, and primary and secondary publications'

Several of the publication planning firms identified are owned by major publishing houses. For example, Excerpta Medica is 'an Elsevier business' and writes that its 'relationship with Elsevier allows… access to editors and editorial boards who provide professional advice and deep opinion leader networks.'

Wolters Kluwer Health draws attention to its publisher Lippincott Williams & Wilkins, with 'nearly 275 periodicals and 1,500 books in more than 100 disciplines,' and to Ovid and its other medical information providers, emphasizing the links it can make between its different arms

Vertical integration is attractive in the industry as a whole: at least three of the world's largest advertising agencies own not only MECCs, but also CROs.

Ghost management of medical journal publications is clearly a substantial business, employing thousands of marketers, writers, and managers. It is large enough that the industry has established the International Publication Planning Association.

The conclusion is that

Given the amount of data that pharmaceutical companies control, the number of publication planning agencies that openly advertise on the Internet, the number of medical writers, the existence of two associations for publication planners, and meetings organized and reports written for them, we can conclude that ghost management is common.

Why is it so important?

Articles in medical journals have real effects upon physician prescribing behavior, which is why pharmaceutical companies invest so much in their publication. Journal articles are heavily used in detailing, to validate claims and rebut worries. Even independent of detailers, responsible physicians and medical researchers search the literature to gather evidence about the best treatments. Published scientific articles are the sources of medical information with the highest authority. Systematic reviews and meta-analyses almost all start with the published literature—so even fully independent reviews are influenced by ghostly activities.

It is clear that ghost management is a major part of the production of pseudoevidence-based medicine.

Sismondo's article is an important contribution to our knowledge of how research on human beings is manipulated into marketing and propaganda, if not outright disinformation. Such practices break the trust of research subjects who thought they were participating to advance science and medical care. Such practices mock physicians who think they are trying to make decisions based on clinical science. Such practices can harm patients who are subjected to tests and treatments based on hype and manipulation rather than unbiased data.

Sismondo ended with some suggestions for alleviating the problem, but noted,

There are no straightforward solutions, short of large changes to the nature of medical publishing and/or research, changes that would effectively sequester pharmaceutical company funding from research and publishing....

We have now said frequently that there are many reasons to think about whether organizations who have a vested interest having clinical research produce particular results should be banned from funding or other involvement in research on human beings. The probably pervasive practice of ghost management is another strong reason to consider this seemingly extreme step.

In the absence of such a drastic prohibition, all I could suggest is a much strengthened and more comprehensive disclosure policy. Many would now agree that all clinical trials should be registered. Such registration should include details of any involvement of pharmaceutical, biotechnology, device and similar companies in the design and implementation of the research, and the analysis and reporting of its results. Any involvement by contract research organizations, for-profit institutional review boards, medical education and communication companies, and the like should be disclosed in trial registration.

Furthermore, there should be full disclosure of all involvement by any of these organizations accompanying any publication or presentation of study results, including publications in peer-reviewed journals, abstract presentations, etc.

Furthermore, the penalties for concealing involvement of any of these organizations should be severe, all the way up to considering failing to disclose as potential criminal fraud.

Allowing ghost management to continue victimizes people who volunteer to participate in research, mocks physicians who try to base their clinical decisions on the best available evidence, and hurts patients.

Hat tip to the Medical Humanities Blog, and one of our corps of Health Care Renewal scouts.


Anonymous said...

Deplorable though it may be, once again the issue is money. Researchers go to the pharma companies to get funding to continue their research, because funding is difficult to obtain elsewhere. Thus they may perpetuate their own jobs, which depend on publish or perish. Is there any wonder at the inevitable result? I wish I had a good answer for the funding problem.
A recent article in the Wall St. Journal (I think that's where I saw it) about a neuroscientist who couldn't get funding for research indicating that progesterone may mitigate the sequelae of traumatic brain injury, because progesterone couldn't be patent-protected; is another illustration.

Anonymous said...

Comment on this from a ghost's perspective here.