A new, important study published in the prestigious Archives of General Psychiatry found that antidepressants decrease the risk of suicidal thoughts and behavior in adults and have no effect on the risk in children.
This is big news, since in 2004 the FDA slapped a black box warning on antidepressants, cautioning that they could cause suicidal tendencies in people under 18. In 2007, the agency extended that warning to young adults under age 25.
I’ve read the study and news accounts about it, including PsychCentral’s, but I’m still left with a lingering question. Perhaps some astute readers who know more about statistics than I can weigh in.
Based on the studies findings, can we conclude that there is really no association between antidepressant use and the risk of suicidal thoughts and behaviors in kids? Or did the kids who grew more suicidal while taking antidepressants and the kids who got less suicidal taking the medications just cancel each other out?
Perhaps this is a thick-headed question, and if so, my apologies (I’m on day four of a very stubborn migraine). But I’m thinking of a recent study, also in the Archives of General Psychiatry, that used a statistical technique called “growth-mixture modeling” to track how individual patients did over time. This method allows researchers to see how individual patients fare – to avoid them getting lost in the shuffle if some improve and some worsen. It found that when patients who have a placebo response are taken out of the equation, many people have a strong positive response to antidepressants.
This most recent study on antidepressants and suicide also followed individual patients over time, which is part of what made it novel. But it’s not clear to me if the researchers used growth-mixture modeling, analyzing those who grew suicidal or more suicidal separately from those who didn’t.
It’s possible they did make these distinctions, and I’m just not seeing the information in the study. What I do see is the odds ratio, or the calculations of how likely those who began taking antidepressants were to experience suicidal thoughts and behaviors versus those who weren’t taking medication.
If anyone with a better grasp of statistics than I do wants to share his or her thoughts, I’d be most grateful. The original article is available for free online here:
In the meantime, I’d also love to hear people’s thoughts on the study’s larger conclusions. Were you surprised to hear that antidepressants appear to decrease the risk of suicide in adults, but not in children?
Also, were you at all puzzled that the study found antidepressants did help to decrease depressive symptoms in children, though it did not decrease the risk of suicidal impulses? This struck me as curious, especially since the research found suicidal impulses were related to the severity of depression, including in children.
The researchers speculated that some other force, maybe not related to antidepressants or depression, might be contributing to suicidal impulses. What do you think that could be?