Reframing the Side Effect: Walter Brown, MD, on Placebo and Nocebo

Patients or clinicians may frame medication side effects as evidence that a treatment is working, a strategy that can improve tolerability and adherence even though the relationship between side effects and efficacy is often unreliable. This approach intersects with the nocebo phenomenon, in which negative expectations about a drug increase the likelihood that patients will experience and report adverse effects, driven in part by symptom amplification and misattribution of unrelated complaints to treatment.¹ Beyond side-effect management, evidence from open-label placebo trials, in which patients are told they are receiving an inactive treatment, suggests that placebo effects can be harnessed transparently and durably, with response rates remaining stable for weeks after treatment initiation.² In this conversation with Walter Brown, MD, and Joe Goldberg, MD, experts discuss practical strategies for framing side effects, tailoring disclosure to individual patients, and leveraging positive expectation and classical conditioning to optimize both placebo response and medication dosing in everyday psychiatric practice.

See the first part of this clinical conversation here.

Joe Goldberg, MD: So, in reframing side effects for patients, it is fair to say that some people might also think, "Well, I'm glad I'm getting a side effect, because that must mean that the stuff is going to work. In fact, if I don't get a side effect, maybe you gave me a lame treatment."

Walter Brown, MD: Absolutely. And that is another sort of way to positively frame side effects. You say the fact that you are getting diarrhea the first week, or that your vision is a little blurred, means that the drug is working. Sometimes that is true, sometimes not, but the message that it means that the drug is working makes those side effects more tolerable for patients. It is not that they do not get the side effect, but they tolerate it better and they are less likely to want to suddenly stop the drug.

Goldberg: Patients on ketamine, and they are disappointed if they do not dissociate, because they figure, well, it is probably not going to work, right?

Brown: Barsky actually identified 2 psychological mechanisms behind the nocebo effect. One was amplification of symptoms, that if you know that the COVID vaccine is likely to give you a headache, you are more likely to pay attention to a minor headache than if you do not. That is amplification. And misattribution: that most of us, particularly as they get into my age category, have aches and pains all the time, other kinds of nonspecific effects, that if you are told that there are certain side effects of your drug, you will attribute them to the treatment you are getting even if they have nothing to do with it.

Goldberg: You get into trouble with people with somatic symptom disorder, or highly somaticizing patients—what about that population?

Brown: Yeah, I mean Barsky has talked about that also. Another way of approaching side effects is to tailor what you tell patients to the context. So if somebody is a big somaticizer or has had a lot of side effects in the past, that should tailor how you approach the information you give them. You do not want to tell them about every possible side effect, you do not want to make it scary. So you need to take into account that sort of thing.

Goldberg: Absolutely. Now, we have been making better placebos over the years, haven't we? The placebo effect in depression has kind of been in the 30 to 40% range, but the studies that go back since the advent of the SSRIs and beyond, seem to hint that while we are making better drug effects, we are also making better placebo effects, and the difference may not be as salient.

But my question is, how can clinicians make the best possible placebo imaginable in their clinical work with patients? What are they not realizing? I sort of hinted at some of this about how you come across and expectations. I think also just imparting a sense that "I care" in a genuine kind of way will amplify the placebo. But any advice about how our audience can deliver better placebo effects?

Brown: I think you want to offer the treatment with a reasonable degree of optimism and positive spin. So say, "I'm giving you this medicine, you should feel significantly better within a week or so." It is a lot better, it creates a more positive expectation than if you say, "Try this, it might work," or to hand somebody in primary care, "This is a powerful painkiller" — that is a very strong message, rather than saying, "Try this, so people get better with it." So I think simple sort of language issues come into play here. And if you wanted to be transparent with patients, I suppose you point out, look, even if something were a placebo, that is not no treatment. I always love Mauricio Fava's work on this, the open trial of placebo in depression, where you consent the patient and you tell them, "I'm going to give you something that doesn't have any active medicine in it, but people often find it helpful. I want you to take it every day. You're going to meet with me on a regular basis. We're going to talk about how you're feeling." And then by week 4 he found a significant difference compared to wait list, I think, was his comparison.

Goldberg: Right.

Walter Brown: There have been a lot of studies now, not only in psychiatry, as you point out, but also for pain syndromes and some other stuff, mostly in the last 2 decades. It is called open label placebo.

Goldberg: You have also written about endurance of the placebo effect, that there is a common perception that it is transient. And you published a paper saying, well, at least out to 12 weeks, something like 80% of people hold on to their response. Any tips you can share with our viewers and listeners about how to sustain that?

Brown: All of the data that we have is that the placebo response is not transient. It is not that simply people will feel a little better for a short period of time. It is quite persistent when it has been looked at over time in clinical trials. And in fact, everybody who has done a significant number of clinical trials which are placebo controlled has an anecdote where they broke the blind, the trial is over, and patients on placebo ask to continue on it. They say, "I tell them, 'But I just need to tell you what you got in this clinical trial was—'" "Doc, I don't care what it was. It made me feel better. I want to continue it."

Goldberg: Wow. So in terms of who is not a good candidate, I think you and others have written about severe cognitive symptoms, psychotic symptoms, suicidality, where again I think we temper our expectations, because for a lot of our audience I am sure, are in mental health, are seeing patients who have not gotten better with other things. How do I phrase this? Placebo should be started sooner than later. It is not the fourteenth thing you want to try, or, well, not as monotherapy anyhow, but judge when and how you expect to capitalize off its potential benefits in the trajectory of sequential treatments.

Brown: Yeah, we have not talked about this, but I think there are very few situations where one should actually prescribe placebo as the first or fundamental treatment. That is not to say that one should not try to enhance the placebo effect in all treatments, but to just give a placebo, I think there are some situations where I think that would be appropriate, where the condition is highly placebo responsive, where there is not a conventional treatment that is clearly going to work for this patient—I think those are instances where you might want to consider that.

But I think more relevant to the everyday practice of medicine and psychiatry is how to harness the placebo response in the context of ordinary treatment, not how to give a placebo. I think there are times when probably a placebo treatment would work. And by the way, speaking of suicide, I am sure you are aware there are a number of studies that have shown that when suicidal patients are discharged from the hospital, they are less likely to attempt suicide again if they get fairly infrequent phone calls just checking in on them, which suggests that even something like suicide or suicidality can be quite responsive to nonspecific supportive interventions, messages of hope.

Goldberg: To take it back to Aaron Beck, I mean, hopelessness at its worst point tends to be the enemy here. So anything, I suppose, when you are checking in, it is also saying, "I believe you can get better, I'm still here." It is not just about "I haven't abandoned you." It is that "I'm looking to see if this is working, because I believe it's working. At least I believe it has a chance to. As long as I have that hope, maybe I can, through osmosis, if nothing more, impart that onto the patient."

So, I do not think I have ever knowingly given a patient a placebo, but I confess I have given patients medicines that I know have smallish effect sizes. I am not going to name any specific medicines. I think most docs have done that, that they give a medicine they think, "Look, it's not going to work because of its inherent therapeutic value, but at least I'll get something like a placebo response out of it." I think that is done very frequently.

Brown: Yes, I would agree.

Goldberg: It sort of gets to the interesting way that clinician prescribers do their own risk-benefit analysis. And in the bipolar world, for instance, those of us who have written about lithium and talk about how it really works better when started sooner, but then some would say, "Well, but there's all these side effects with it, so I'm going to hold off on this heavyweight treatment and give you something milder." It may even be something that has not been shown to work in randomized trials. There are a few drugs that people I know do prescribe in bipolar illness that would not be the first, second, or third line option in a practice guideline because they really have not been vetted and shown to have either clear superiority to placebo or a meaningful magnitude of difference in terms of the effect size. But "I want to spare you side effects, so I'll give you something that I think is going to spare you side effects." It is a gesture in some ways.

In my mind, those interventions are okay when patients are not severely ill and when they are engaged in the treatment and you have your eye on the ball and you are paying attention to, well, if they do clinically worsen, I may need to step up the game. But with those small effect size drugs that may not have been as well established, I am sure banking on every ounce of placebo that I can muster to try to get something going.

Brown: Yep. It is good. The art and the science.

Goldberg: Well, this has been eye-opening, as so often is the case. Do you see anything about the future of placebo effects? I mean, the pharmaceutical industry works so hard to suppress them. We clinicians want to be cognizant of them. If we want to try to capitalize on them, especially whether you are an investigator or a clinician, is there a future for the placebo effect?

Brown: Well, one of the axes that I like to grind here, and we did not talk about this, but whereas expectation is by far the most studied feature of the placebo effect mechanisms, conditioning also plays a role. Classical conditioning, and you could have conditioned drug responses, and there have been a number of pretty good systematic studies which have shown that you can condition drug responses; that is, if somebody has been on an ADHD drug or an immunosuppressant, and then you substitute a placebo for some of the drug, you retain the effect as a result, because if you substitute placebo or some other feature, the placebo becomes a conditioned stimulus to the response.

Goldberg: That is a familiar concept to you, if I want to pull back on a dose, perhaps, too. So I'm coasting along at my 60 or 80 milligrams of fluoxetine, and if I drop that by 25%, I do not necessarily assume I'm suddenly going to get a 25% decline in efficacy.

Brown: According to the research that's been done, and it's not a huge amount, but if you drop the SSRI by 20 milligrams and substitute placebo, you don't lose the response, so the placebo has become a conditioned stimulus. And what this shows is that you could give people far lower doses of drugs if you use this conditioning model than you now use. And as a result, it would be less expensive and it would also save them side effects. But that has not really been tried in clinical practice. But it's certainly something our audience might want to think about.

Joe Goldberg: Not to rush out tomorrow and just drop everybody's dose, but if the occasion arises, if the issue of deprescribing comes up, or “I'd like to be on less," if you're starting from a place of wellness and the patient's got an established baseline now of doing well, it may not be so outlandish an idea to say, "Well, we can cautiously do this, and I've got my finger on the placebo button. However, I can impart that to you, whether it's tangibly or interpersonally, or just by otherwise conveying the safety net—we can do this, and we may be able to win, if we can spare you some side effects. I'm on your side, I'm here to help and sustain the benefits," which is really what we're all trying to do, right?

Brown: I mean, I would be happier about that approach if there was some more data to actually show that it worked. But I think that's been an understudied and underutilized approach.

Goldberg: It's going to be tough to do drug trials of placebo versus no treatment, right? We almost need those, don't we, just to really help tease out these active ingredients. I would love to see some studies looking at immunomodulator biomarkers of the placebo-responsive patients at different points in time, to some of the things that you were saying earlier on. So it's fascinating stuff. Dr Brown, I want to thank you so much for taking the time to join us today. Any final comments before we wrap up?

Brown: I think the idea that you should consider placebo, the placebo effect and placebo response, not a nuisance, which is what it has been thought of for a long time, but as an important part of all healing, and it should be enhanced. It's a third of the effect. We would be remiss to not appreciate and recognize that and use it.

Goldberg: Absolutely. Well, again, I want to thank you so much for joining us today. We hope you've found this helpful, and we will look forward to seeing you again real soon on the next episode of “Brain Trust: Conversations in Psychopharmacology.”

Dr Goldberg is a clinical professor of psychiatry at The Icahn School of Medicine at Mount Sinai in New York, NY and the immediate-past president of the American Society of Clinical Psychopharmacology.

Dr Brown is clinical professor emeritus of psychiatry and human behavior at Brown University.

References

1. Barsky AJ, Saintfort R, Rogers MP, et al. Nonspecific medication side effects and the nocebo phenomenon. JAMA. 2002;287(5):622-627.

2. Kaptchuk TJ, Friedlander E, Kelley JM, et al. Placebos without deception: a randomized controlled trial in irritable bowel syndrome. PLoS One. 2010;5(12):e15591.