Philip Dawdy takes a interesting look at a new study of the safety of placebo arms in clinical trials of antidepressants in teens. My own quick scan of the study [which Dawdy makes available as pdf download] suggests it’s full of great nuggets.
Its take-home: Placebo treatments produced remission rates of 48%, while the rate for active treatment was 59%. And, quite interestingly, the study concludes:
Patients who responded to placebo generally retained their response. Those who did not respond to placebo subsequently responded to active treatment at the same rates as those initiallyl assigned to active treatments. There were no differences between groups in rates of suicidal events, study retention, or symptom worsening.
In other words, placebo left options open, and avoided most side-effects, at the cost of reaching 5 of 10 patients rather than 6 of 10. (Effectiveness rates of both placebo and treatment vary among trials; these are a bit higher than rates in other meta-analyses I’ve read, but fairly on target.)
These findings are full of implications, some of which Dawdy lays out below. (Point of info: You’ll see different rates for ‘response’ and ‘remission’ here. Response means their symptoms eased at least a little. Remission means their symptoms lifted enough to classify them as no longer depressed.)
1. Placebo treatment didn’t hurt teens, researchers found, so the ethical question of “denying” treatment is addressed.
2. The teens who were in the initial 12-week placebo arm did fairly well. Of kids on placebo-only for 12 weeks, 35 percent responded to their “treatment” while 58 percent of the other groups did (a very common 23 percent effect size for active treatments including CBT).
3. At week 36, the response rate was 82 percent in the initial placebo group and 83 percent in the active treatment groups, which certainly argues for some level of sustained placebo effect (“Patients who responded to placebo generally retained their response,” the study authors note.) The remission rate was 48 percent in the initial placebo patients and 59 percent in the active treatment groups, so certainly not much of an argument can be made that offering kids a placebo limits their ability to respond to other treatments. (It also make me wonder, skeptically of course, about the various prevention paradigm claims made for anti-depressants.)
4. Although the authors don’t address this point, I think you can reasonably make the case that the placebo effect exists in both kids and adults undergoing various depression treatments (or taking a placebo) and it ought to become a legitimate object of research. Certainly, a study by Arif Khan last year showed an enormous and long-last placebo effect among adults and a recent meta-study in the AJP showed a large placebo effect in a slew of depression trials in kids and teens.
5. There were suicide attempts among initial placebo patients, mostly occurring when the patients either moved to active medication or psychotherapy or a combination. The authors note:
“It is important to note that the large majority of suicide attempts among placebo/open patients (80 percent; 12/15) occurred after the period during which they were on the placebo. Thus, depressed adolescents assigned to placebo condition are for the most part able to tolerate the waiting period without significant clinical distress.”
This may, as Dawdy notes, strengthen the argument for caution in using antidepressants for kids. This more interesting point:
I think you can reasonably make the case that the placebo effect exists in both kids and adults undergoing various depression treatments (or taking a placebo) and it ought to become a legitimate object of research. Certainly, a study by Arif Khan last year showed an enormous and long-last placebo effect among adults and a recent meta-study in the AJP showed a large placebo effect in a slew of depression trials in kids and teens.
I’m not sure how much money NIMH is getting of the $1.1 billion in stimulus money designed to sort out which medical treatments work best (I’m dubious of how much of this money will actually create jobs but that’s a separate issue), and it would make sense for NIMH to allocate some money to funding real world research on the placebo effect in depression treatment. There’s too large of a consistent effect out there to not want to understand how large it is and, if possible, what’s driving it (beyond psychological suggestion).
I’m not clear on whether Obama’s plans for comparative effectiveness include a way to assess placebo treatments — presumably a difficult task, but perhaps doable if enough doctors are indeed prescribing placebos. It would be interesting to see how this worked on large scales.