On Monday, February 12 2007, the U.S. Center for Disease Control and Prevention released its Annual Summary of Vital Statistics. Inside the report, suicide rates for children — and especially teens — appeared to have rose dramatically over a one-year time period, 2003 to 2004.
And like moths drawn to a flame, every news outlet in the U.S. dutifully reports the supposed link between this increase in suicide rates to a decrease in antidepressant prescriptions in 2004.
The problem is — like most mainstream medical reporting done on tight deadlines with virtually no time to fact check and get the story right — every news outlet was dead wrong.
But it’s not like we could only blame the news outlets (such as ABC News, whose story was typical reporting on this issue).
Experts in the field, including researchers and university professors, were quick to make pronouncements about the cause of this increase. The report is simply an annual tabulation of data — it is not a study and does not propose any hypotheses to test. Yet researchers, who supposedly should know better, instantly created a causal relationship between two variables where only a correlational relationship exists.
What did these experts blame on the rise in suicide in children?
An FDA label on certain antidepressant medications.
For instance, Dr. Charles Nemeroff, chairman of the department of psychiatry and behavioral sciences at the Emory University School of Medicine, was quoted by ABC News as saying, “I have no doubt that there is such a relationship.” That is a bold, unsupported statement.
The ABC News article continues…
“This is very disturbing news,” said Dr. David Fassler, clinical professor of psychiatry at the University of Vermont College of Medicine. “The adolescent suicide rate has been declining steadily since the early 1990s.
“The sudden increase in the adolescent suicide rate,” Fassler continued, “corresponds to the significant and precipitous decrease in the use of SSRI antidepressants in this age group.”
I can see why these researchers feel this way, because the data seem to support a decline in SSRI prescriptions to teens and children in 2004.
But, as others are making clear, the data and the datatrends about childhood prescriptions is not at all clear for 2004.
Correlation does not equal causation. Without a specific study examining causative factors in this increase, one simply cannot make factually-based claims. The rise in the child suicide rate could’ve been caused by a multitude of equally plausible other factors.
It also looks like researchers are quick to blame the FDA making the “wrong” decision with regard to the black box label. The FDA website has a detailed page about the controversy.
As any scientist will tell you, though, the data don’t support this conclusion. Without an empirical study that examines this specific hypothesis, all we have is a possible correlational relationship. We shouldn’t make public policy decisions — or research or treatment decisions — based upon such a thin relationship.