Placebo Response in Antidepressant Treatment
Jim Peacey, MD
PAL Consultant
Clinical trials of antidepressants in children and adolescents are often burdened by very high placebo response rates making it very difficult to show any additional benefit from the drug being studied. Is most of what we see when youth get better after starting an antidepressant due to placebo effect? If so, is that a problem? If the depression is getting better, shouldn't we want to make the most of whatever can contribute to that improvement?
It's useful to distinguish between placebo response and the placebo effect. The term "placebo response" describes improvement seen in the placebo arm of a study and may be due to factors in addition to classic placebo effect. Placebo effect is usually attributed to belief that a treatment will be effective. We know that there are biological changes associated with placebo effect, and that improvement is not just imaginary. Placebo effect can account for a portion of placebo response. Belief in a treatment may be enhanced by confidence communicated by the one recommending it or testimonials from others. There are other cognitive and behavioral elements of medication treatment that may also be important such as the daily practice of taking medication with the intention of feeling better. Patients may be paying attention to possible changes or signals that they may be getting better, and we can encourage that process by asking about changes that may signal improvement. Regular follow-up appointments can produce an additional psychotherapy effect: the benefit of regular interaction with someone who is not only expressing expectation of improvement but also concern about their life and an interest in wanting to help as well as supporting changes in behavior such as better sleep hygiene, exercise, and more involvement in activities.
There are still other reasons that patients may be getting better in the months following the start of an antidepressant that may have nothing to do with treatment response at all. The most typical course of an episode of untreated depression is eventual recovery with average illness duration of a year. When antidepressant trials and titration can take a few months, spontaneous recovery will account for a portion of the response we see. We often start antidepressants when people are at a particularly low point in mood, so subsequent improvement might be expected just due to "regression to the mean" regardless of treatment decisions. And there are depressive episodes that will be, in retrospect, better described as adjustment disorders that improve spontaneously when stressors improve.
Patients and clinician should be glad to see improvement regardless of the best explanation for it. But there are elements that contribute to placebo response that are less salutary. There may be observer bias in parents and clinicians with overinvestment in a treatment and grasping for signs that it may be working even when there is no actual improvement so that an ineffective treatment is prolonged. And patients may try to please adults who are very invested in a treatment by minimizing symptoms that are still there; rating scales may not always give us an accurate picture of symptoms. We can develop opinions from "clinical experience" that can mislead us to either become overly confident in treatments that we know have a specific benefit in a minority of patients ("Number Needed to Treat" for antidepressants is 4 at best) or think that we are doing something wrong when response doesn't happen. A more realistic view of the effectiveness of our treatments does make it easier to understand certain phenomena. The observation of antidepressant "poop out" where a treatment that had been working well stops working or doesn't work when it is restarted at a later time may not be that surprising if much of the improvement we see in clinical practice was never due to the pharmacologic properties of the drug in the first place.
What we can tell patients and families is that more than half of patients have meaningful improvement when treated with one of our first-line SSRIs, and that half of those that don't will respond to an alternative SSRI or SNRI. We can enhance basic psychotherapy effects by having regular follow-up with patients to inquire about their mood and life circumstances, encourage healthy habits of sleep, exercise, and social activity, and troubleshoot any problems with residual symptoms or side effects. We can encourage engagement in evidence-based psychotherapies that not only increase the likelihood of improvement but can improve the durability of response. And we should be trying to take people off their antidepressants after a year or so of recovery with supports in place and stressors managed. We will discover some who do have a more persistent disorder with a specific life-changing response to medication treatment that will need to be continued. And we will spare many others who never did have a specific drug response from the burden and risk (though minimal!) of prolonged drug treatment.
|