skip navigation nih record
Vol. LXI, No. 14
July 10, 2009
cover

next story



Epidemiology Studies Prone to Buffeting by Politics, Uncertainty

On the front page...

Because novels and essays are often better than scientific studies at expressing nuanced truths, Johns Hopkins epidemiologist Dr. Leon Gordis’s delivery of the 15th annual Robert S. Gordon, Jr. Lecture on June 3 sometimes seemed more like a liberal arts seminar than a scientific presentation. After all, how many times have you heard Hungarian novelist Sandor Marai quoted in Masur Auditorium?

Continued...


  Epidemiologist Dr. Leon Gordis (r) accepts the Gordon Lecture plaque from Dr. Barry Kramer, NIH associate director for disease prevention.  
  Epidemiologist Dr. Leon Gordis (r) accepts the Gordon Lecture plaque from Dr. Barry Kramer, NIH associate director for disease prevention.  

But that’s not to slight a science that has the power to trace the worldwide spread of flu virus to a pig farm in Mexico, or an outbreak of salmonella-tainted peanut products to a specific producer. Rather, it’s because the humanities have something to say—and Gordis’s 32 years of teaching epidemiology to more than 17,000 students have a certain authority—when orderly facts bump up against messy humans.

In a talk titled “Epidemiology at the Interface of Science, Policy and Politics,” Gordis used two prominent screening studies—one involving mammography and the other involving spiral CT to detect lung cancer—to show that for all our public bluster about data driving public health practice, “truth” can be up for grabs when it A) fails to align with entrenched political power and B) has already been purchased by an unseen hand.

In the first example, Gordis examined the 22- year debate about the value of mammograms for women ages 40-49. In 1987, the National Cancer Institute recommended that screening commence for women at 40. Six years later, experts could not agree on the value of routine screening for women in this age group. Gordis himself chaired a 1997 NIH Consensus Development Conference which found that the data did not warrant a universal recommendation, leaving it up to the discretion of a woman and her physician. Before the conference even ended, controversy erupted as the NCI director and a past NIH director took immediate exception to the findings.

“What we see here is the effect of political considerations,” said Gordis. Even today, the issue “simmers and boils,” he said. “There are still major disagreements.”

It’s not always that easy to issue a smackdown to suspicious science. Uncertainty is a fact of life in science, and is indeed “the reason so many of you go to work upstairs in the labs—it’s actually an integral part of science.”

The second example owes much to the congenital skepticism of the seasoned novel-reader (Gordis took a degree in Hebrew literature before earning his M.D.). A paper based on a 2006 study, the International Early Lung Cancer Action Program, claimed that spiral CT scanning can detect lung cancer in time to cure it. Four flaws, however, tainted its conclusions— Gordis said the study was not randomized, had no control group, failed to address the harms of screening itself and, perhaps most fatally, had been financed by a tobacco company. Quoting former NCI oncologist Dr. Otis Brawley, Gordis noted, “If you’re using blood money, you need to tell people.”

It’s not always that easy to issue a smackdown to suspicious science, Gordis warned. Uncertainty is a fact of life in science and is indeed “the reason so many of you go to work upstairs in the labs—it’s actually an integral part of science. But it makes policy based on epidemiology very hard. Scientific truth is often transient.”

More literary allusions followed, especially with respect to being too sure of one’s conclusions. Gordis finished with advice for the upcoming generation of epidemiologists: “I think we are resting too much on our laurels. We need to re-examine what we are doing. Populations are made up of people and their families. All those numbers represent human beings, who are often stressed and ill. There is a danger of treating people as abstractions.”

Today’s young trainees, he observed, are apt to do timid, lukewarm, incremental work. “I think there is a great fear of being found wrong,” he lamented. “Youth is a time to be daring.”

In an era when a doctoral candidate is more likely to download a PDF of a paper in his or her field than to browse freely among a variety of offerings, Gordis recalls the old days when a trip to the library yielded “a breadth of vistas” as he flipped through dozens of articles in fields as diverse as economics, sociology, psychology and health administration. High-tech downloads may be fast and accurate, but they are narrow, he warned.

During a brief Q&A, Gordis offered a frank response to a query about whether Gardasil, the much-touted HPV vaccine, should be made mandatory by certain states: “There is no evidence that Gardasil prevents cervical cancer. We really don’t know if it’s going to work. It seems to be effective for [viral types] 16 and 18, but there is some concern that other types may come in to displace those. I don’t think the last word is in yet. But the drug company’s behavior in pushing this vaccine is inexcusable. I see no reason why it should be mandatory.” NIHRecord Icon

back to top of page