Beware extrapolation from ecological dataBMJ 2010; 341 doi: http://dx.doi.org/10.1136/bmj.c6844 (Published 30 November 2010) Cite this as: BMJ 2010;341:c6844
- Tarek A Hammad, associate director of epidemiology1,
- Andrew D Mosholder, senior medical epidemiologist1
- 1Division of Epidemiology, Office of Surveillance and Epidemiology, Center for Drug Evaluation and Research, Food and Drug Administration, 10903 New Hampshire Avenue, Silver Spring, MD 20993, USA
Identification of an association between antidepressant drugs and suicidal thinking and behaviour in adolescents in 2004, and the regulatory warnings that followed, sparked a debate on the potential for discouraging appropriate use of antidepressant drugs,1 2 3 with the unintended consequence of eventually exposing more patients to the suicide risk of untreated depression.4 These concerns were strengthened by an upturn in 2004 in the US rate of adolescent suicide.4 Significant falls were also reported in antidepressant drug prescribing for paediatric patients after the regulatory actions.3 4 5
Six years later we present data from the Centers for Disease Control and Prevention on the overall trend in rates of suicide among adolescents showing that the rate of suicide decreased after the unexplained increase of 2004 (figure⇓). In 2007, the most recent year with available data, the rates were the lowest reported in 25 years.
These findings underscore the limitation of using ecological population based approaches and especially of relying on a single year’s data to draw strong conclusions and raise what may turn out to be premature concerns. Ecological data could not establish a cause and effect association, neither could such data determine whether the changes in prescribing were due to depression not treated with drugs or to prescription of fewer antidepressants for other conditions. Although being vigilant for unintended consequences of regulatory actions is important, ecological data may not be the best guide to whether drugs are being used appropriately in a given patient population.
Cite this as: BMJ 2010;341:c6844
No official support or endorsement by the US Food and Drug Administration is intended or should be inferred.
Competing interests: None declared.