How full is the glass?

Instead of speculating about our differing methods as the source of our differing conclusions, Turner and Rosenthal instead focused on our adoption of UK’s National Institute for Health and Clinical Excellence (NICE) criteria for clinical significance. NICE adopted Cohen’s definition of a “medium” effect size (0.5) for clinical significance.^{1 2} Cohen intended that a medium effect size represent an effect of a size likely to be visible to the naked eye of a careful observer (p 156). By this interpretation, clinical significance of antidepressants is reached when a careful observer can see that medicated patients have noticeably lower depression than patients taking placebo. In this light, the NICE criteria seem sensible indeed.

A more appropriate use of Turner and Rosenthal’s “d-juice” metaphor (where d-juice is the level of antidepressant efficacy) to explain our findings is that the amount of d-juice in the trials was pitifully small or even non-existent when people with less severe depression were examined, but larger when very severely depressed individuals were examined. Yet note that effect sizes (in their d form) can in theory take values as large as infinity.³ Thus, even the medium effect size observed for trials that sampled extremely depressed individuals could be viewed as inconsequential.

In considering treatment for depressed patients, should clinicians be content that antidepressants attain the clinical significance criterion only for the most severely depressed patients? Even following drug (or placebo) therapy, the average patient can still be characterized as depressed (see reference 5 for the logic on this inference⁴). Consequently, it seems unsafe to assume that antidepressants will fully succeed for the average severely depressed patient. In short, this trickle of “d-juice” is likely to leave a depressed patient thirsty and looking for alternative means to fill the proverbial glass.