Do Antidepressants Prevent Suicides?
Drugs like Prozac and Paxil have come under fire this year as two separate studies published in medical journals showed that patients taking antidepressants fared only slightly better than those taking placebos. Meanwhile, the British science magazine New Scientist reported in February that GlaxoSmithKline researchers may have tried to hide potential suicide risks associated with antidepressants it manufactures.
Ever since the F.D.A. forced drug companies to add suicide warning labels to happy-pill bottles in 2004, the effectiveness of the wonder drugs has been called into question. (The move came after evidence that children on antidepressants might exhibit a higher tendency towards hurting themselves.)
But, as always, there is another side to this debate. One way to measure the effectiveness of these drugs is to see their effects on suicide. That's the analysis performed in a recent study by Jens Ludwig of the University of Chicago, Dave Marcotte of the University of Maryland and Karen Norberg of Washington University.
Since suicide is relatively rare, samples sizes in clinical trials are too small to detect any significant impact. Most trials, thankfully, don't result in subjects ending their lives -- in part because drug companies don't allow people with a history of suicide attempts to enroll. The researchers say that to find an effect of 20 percent on suicide mortality, a clinical trial would need 1.9 million subjects. In comparison, the larger of the two studies above -- which actually pooled a number of different studies over many years -- had a sample size of about 10,000.
To combat this, they looked at the sales of one class of antidepressant -- selective serotonin reuptake inhibitor -- in 26 countries over 25 years. They focused on the differences in how quickly the use of these new antidepressants spread through each country. It turned out that that an increase in SSRIs sales of one pill per person actually reduced suicide by 5 percent:
"Our estimates suggest that at least on average SSRIs may be a very cost-effective means for saving lives. Commonly used SSRIs can currently be obtained in the United States for around $0.13 per pill. Our estimates thus imply that each additional $26,000 spent on SSRIs will avert one suicide completion, far below the cost per life saved from most other public health, regulatory, or other forms of government intervention."
Looking north of the border offers more concrete evidence of the potential effectiveness of antidepressants. Like the United States, Canada regulators warned about the dangers of antidepressants given to children. New research published this week has found that the incidence of suicide among kids and adolsecents increased 25 percent after years of decline as the use of antidepressants fell 14 percent. (The findings actually echo similar results in the U.S. following the F.D.A. warning.)
These results aren't likely to end the debate over suicide and antidepressants, so as an alternative, here's a third -- and admittedly way too morbid -- way to look at it: Research on smokers has shown that, on net, they're actually a benefit to society because they die sooner and use less resources. Could the same argument be made for those who want to commit suicide?
- Should the Fed Go Long?
- Dec 1 2008 4:38PM EST
- Bernanke's Speech
- Dec 1 2008 2:58PM EST
- Even Nobel Economists Can Be Intellectually Dishonest
- Nov 30 2008 9:36AM EST
- A 5-Point Plan for Getting Out of This
- Nov 28 2008 1:24PM EST
- Do Markets Filter Irrationality?
- Nov 26 2008 11:25PM EST
- Are Percentages Really That Hard?
- Nov 26 2008 10:07PM EST
- Chart of the Day
- Nov 25 2008 3:27PM EST
- Highlights of the Citi Bailout
- Nov 24 2008 12:29AM EST
- 24 Hours in the Stock Markets
- Nov 23 2008 6:44PM EST
- Bloomberg Not Shy About Buts
- Nov 22 2008 12:55AM EST
- FDIC Not Insuring Fed Funds
- Nov 21 2008 10:30PM EST
- Counterparty Risk and Potential Losses from OTC Derivatives
- Nov 20 2008 4:27PM EST
- Dining Democracy
- Nov 19 2008 6:44AM EST
- Recession Dating
- Nov 17 2008 11:21AM EST
- The Best and Worst Restaurants in Manhattan
- Nov 17 2008 7:45AM EST
Categories
Links
- Email me

- Geary Behaviour Centre

- NBER Working Papers

- Social Science Statistics Blog

- Decision Science News

- Freakonomics

- New York Federal Reserve Research

- Statistical Modeling, Causal Inference, and Social Science

- Marginal Revolution

- EconTalk

- MoneyScience

- VoxEU

- Journal of Interest

- Bluematter

- Economist's View

- Research Recap

- Social Science Research Network

- Institute for the Study of Labor

- EconPapers

- Real Time Economics

- Center for Economic Policy Research

- B.I.S. Working Papers

- C.B.O. Director's Blog

- Federal Reserve Working Papers

- Institute for the Study of Labor

- O.E.C.D. Factblog

- Philadelphia Fed Research

- St. Louis Fed Research

- Sabernomics

- Sabermetric Research

- Economic Principals

- Numbers Guy

- Econbrowser

- STATS Blog

- Jeff Frankel

- Junk Charts

- Predictably/Irrational

- Tim Harford

- TierneyLab

- Curious Capitalist

- DataPoints: The Dismal Scientist Blog










