Journalists Researching Drugs Should Be Wary of Statistical Traps

By Chris Adams

For Dr. Peter D. Kramer, a prominent psychiatrist and author, the mission of his most recent book was basic: “It seemed important to say the simple truth that antidepressants work,” he said.

In recent years, his understanding of antidepressants – how well they work, for whom they work – has come under scrutiny by some researchers, as well as the public, who claim that the widely used medications are little more than “placebos with side effects.”

Kramer, a practicing psychiatrist and emeritus professor of psychiatry and human behavior at Brown University, is best known for “Listening to Prozac,” a bestseller from 1993 that explored the impact of antidepressants. His latest book is “Ordinarily Well,” out in 2016 to positive reviews. (His views are summarized here, in an essay in The Wall Street Journal.)

The book is a deep dive on the statistical underpinnings of research on antidepressants, as well as the field of medicine in general. In a session with National Press Foundation fellows, Kramer detailed the pitfalls in researching and reporting on psychiatric drugs.

One of the key points was a discussion on meta-analyses, the research technique in which several studies are combined to discern the impact a medication might have. While the technique is widely used, it’s also improperly used, often due to the decision on which studies to include – or to exclude. (Among the meta-analyses he critiques in his book and his talk is this 2010 study from the Journal of the American Medical Association on the impact antidepressants have on mild depression.)

This program is funded by The Janssen Pharmaceutical Companies. NPF is solely responsible for the content.

More Presentations
Help Make Good Journalists Better
Donate to the National Press Foundation to help us keep journalists informed on the issues that matter most.
DONATE ANY AMOUNT