The Prozac FDA studies—they date back 30 years and likely have the flaws that lead to an underestimate of efficacy—yield an effect size of 0.26. This figure is modest, but it attaches to a drug that in other trials helps a substantial proportion of patients. The new data analysis pulls some other antidepressants down to the Prozac figure. Celexa looks comparable, with an effect size of 0.24. And yet in a large study designed to mimic everyday uses of medications, three months of Celexa brought full remission to about one-third of chronically depressed subjects, with another 10 percent to 15 percent experiencing a substantial response. These were hard-to-treat patients, often with multiple accompanying disorders. Here's what an effect size of 0.24 in FDA studies may mean: In the first go-round, in the real world, almost half of previously treatment-resistant patients will show improvement.