[ RadSafe ] Irreproducibility

Thu Jul 27 08:48:34 CDT 2006

In the Boston Globe today is an Opinion article claiming an estimate of 
over 50% for the irreproducibility of scientific literature:

http://www.boston.com/news/globe/editorial_opinion/oped/articles/2006/07/27/
science_and_shams/

DAVID A. SHAYWITZ
Science and shams

By David A. Shaywitz  |  July 27, 2006

THE NATIONAL preoccupation with university researchers who collaborate 
with drug companies has now blossomed into a full-fledged witch hunt. 
Before we burn these heathen scientists at the stake, however, we might 
want to step back and examine our underlying assumptions.

The worry about academics working with industry stems from the fear that 
drug company money will bias the conclusions of otherwise trustworthy 
university scientists. Yet this argument is ultimately a dangerous 
sham -- a sham, because it derives from a naive preconception about the 
purity of academic scientists, and dangerous, because by discouraging 
collaborations between researchers from university and industry, we will 
slow the pace of medical progress.

The quaint notion that university researchers operate on a higher moral 
plane than business people might be charming if the implications weren't 
so profound; attention-getting results can lead directly to employment, 
promotion, and prizes, all leading to personal gain. Not surprisingly, 
some of the most notorious perpetrators of research misconduct have been 
academic investigators with no company relationships.

While fraud itself is fortunately quite rare in science, you still 
shouldn't believe everything you read. It turns out that the scientific 
literature is characterized by a shockingly high rate of 
irreproducibility, often estimated to exceed 50 percent. A review of 
this subject last year in the journal PLOS-Medicine by Dr. John 
Ioannidis became an instant cult classic.

Among the factors associated with irreproducibility includes a category 
Ioannidis terms ``financial and other interests and prejudices," since 
``prejudice may not necessarily have financial roots." Such nonfinancial 
prejudice may include a bias toward a particular scientific theory, a 
bias against competing researchers or perspectives, and a desire to 
generate publications for career advancement. Unfortunately, the 
newfound obsession with financial conflicts of interest obscures these 
other, often more compelling pressures on university researchers.

Ultimately, the myopic focus on financial conflicts is likely to 
discourage relationships between university researchers and drug 
companies -- a bad idea, since these associations offer enormous 
potential for medical advancement.

While almost all of the basic research in the nation occurs at 
universities, translating these discoveries into clinical application 
generally requires the contribution of industry. In fact, the initial 
scientific insight, though essential, is typically only the first step 
of an exceptionally long, difficult, and characteristically unsuccessful 
process, since only a vanishingly small number of leads ultimately 
evolve into clinically significant products.

The commercial development of a new drug or device is frequently 
catalyzed by the contributions of academic experts at many steps along 
the way.

Basic scientists contribute detailed knowledge of the biological pathway 
the new agent will target, for example, while clinical researchers 
provide important information about the medical condition the new 
product will treat. Not surprisingly, the same university researchers 
sought out by industry can also provide valuable insight to scientific 
colleagues, in the form of review articles, and to government 
regulators, through participation on FDA advisory committees. But both 
of these activities have alarmed critics, who maintain that researchers 
with industry involvement are indelibly tainted and no longer worthy of 
our trust.

To be sure, drug companies have brought some of this skepticism upon 
themselves. Consider Nexium, a product created not to address an unmet 
medical need, but purely to offset the fact that the company's previous 
(and fundamentally identical) heartburn drug, Prilosec, was about to go 
off-patent.

At the same time, I've also seen patients on death's door recover 
because they were given a powerful new antibiotic, or because their 
condition was accurately diagnosed by a technologically advanced 
machine. I've also had the remarkable experience of restoring the vision 
of a patient who had nearly gone blind because a fast-growing pituitary 
tumor had compressed his optic nerves; treated with the appropriate 
drug, the tumor receded within days. In each of these cases, patients 
benefited from the extensive efforts of a drug or a medical device 
company, and from the pivotal role played by the academic physicians who 
worked with the company to accelerate the development of these important 
products.

Coming up with novel treatments for dreadful diseases is a difficult 
enough task under the best of circumstances; let's not complicate this 
mission further by misjudging either the intrinsic virtue of academic 
scientists or the alleged venality of drug companies; instead, let us 
learn how to foster productive, transparent collaborations so that the 
promise of modern science can be realized in the treatment of our 
patients.