Anti-Depressants Reduce Suicide

An increase in SSRI sales of one pill per person per year -- about a 12 percent increase over year 2000 sales levels -- is associated with a decrease in deaths from suicide of about 5 percent. Furthermore, now that SSRIs are off patent, spending an additional $20,000 on them in the United States could avert one death from suicide; that is a cost per life saved far below the cost of most other public health or regulatory government intervention.

Selective serotonin reuptake inhibitors (SSRIs) - a class of antidepressants -- are among the most widely prescribed medications in the world, but they are now at the center of an intense debate about whether antidepressant treatment may increase rather than decrease the risk of completed suicide. In Anti-Depressants and Suicide (NBER Working Paper No. 12906), co-authors Jens Ludwig, Dave Marcotte, and Karen Norberg find that SSRIs appear to save lives. In general, an increase in SSRI sales of one pill per person per year -- about a 12 percent increase over year 2000 sales levels -- is associated with a decrease in deaths from suicide of about 5 percent. Furthermore, now that SSRIs are off patent, spending an additional $20,000 on them in the United States could avert one death from suicide; that is a cost per life saved far below the cost of most other public health or regulatory government intervention.

Using a panel dataset from 26 countries over 25 years, the authors sort through the possibility that earlier studies may have overstated the benefits of SSRIs by attributing to them the effects of such things as improvements in mental health systems, or they may have understated the benefits by not taking into account any increases in the prevalence of mental health problems. The authors use the sales of four other drugs (as what economists term "instrumental variables") to control for the effects of systemic changes in the health system that are unrelated to the causal effects of the SSRIs themselves. It turns out that countries with higher rates of growth in new drugs, generally, have higher rates of growth in SSRI sales. Furthermore, despite clinical evidence that antidepressant use may increase the risk of non-lethal suicidal behavior in pediatric patients, the authors find that the protective effect of SSRI sales on suicide mortality is largest, in both proportional and absolute terms, for people aged 15-24.

These estimates of SSRI effects improve on those from randomized clinical trials (RCTs). Although RCTs are the "gold standard" for the study of treatment effectiveness in the medical literature, they have at least two important limitations in studying a rare but serious adverse event such as completed suicide: first, because even combined analyses of multiple RCTs have involved sample sizes that are too small to detect differences in rare outcomes like suicide; and second because RCTs may not adequately represent the average patient. People at high risk for suicide are often excluded from trials for ethical reasons, and the quality and intensity of care available to people in clinical studies may be unrepresentative of usual levels of community care.

-- Linda Gorman