That’s what the latest study published in the American Journal of Psychiatry says. In a nutshell, they initially selected their subjects based on whether they were depressed and were excluded based on a number of factors (one of which was previous failure to respond to prozac or adverse reactions to prozac). All of the participants were initially tried on prozac, with dosage increases in waranted. Those who didn’t respond to prozac were excluded from the continuation phase. Then they were randomly assigned to continue on prozac or placebo. So, the placebo group is abruptly taken off of prozac and put on sugar pill while prozac group experiences no such disruption. The authors thought this was okay because of prozac’s “long half-life.”

The study found that 42% of the prozac group relapsed within 6 months, whereas 69% of the placebo group relapsed. With an even stricter measure of relapse, only 22% of prozac group relapsed, whereas 48% of the placebo group relapsed.

This makes prozac sound pretty good for kids, does it not?? It does….

But there’s a catch. I’ll just focus on the “stricter” measure of relapse for now. This was based on scores on a scale that measures sympoms of depression the CDRS-R. Scores >= 40 over two sessions with a psychiatrist (separated by 2 weeks) were considered to be relapses, whereas scores under 40 were not considered to be relapses. How do we know whether 69% of the prozac group didn’t have scores of 39 (not relapsed), whereas 69% of the placebo group had scores of 41 (relapsed)?? We don’t know, because we don’t know any scores. We don’t know if the differences are truly clinically significant, because we don’t know what the differences in the actual scores were.

Also, the fact that the study was set up to only include people who respond to prozac, we know nothing about how this would work if applied to the general population of children who are depressed. In other words, the deck was stacked in prozac’s favor. The authors, in my opinion, just seemed to gloss over the possibility that the increased “relapse” in the placebo group could have been caused by the abrupt discontinuation of prozac.

I suppose we’ll have to wait another 10 years for Irving Kirsch to publish an analysis on whether antidepressants provide any clinically significant benefit over placebo. I, for one, am not convinced by this poorly designed and written study.