The Art of Deprescribing: A Framework for Ending Medication with Joseph Goldberg, MD

The Journal of Clinical Psychiatry
https://www.psychiatrist.com/jcp/
Publisher of peer-reviewed research discussed in this episode.

Dr. Joseph Goldberg – LinkedIn
https://www.linkedin.com/in/joseph-goldberg-922b5317/

American Society of Clinical Psychopharmacology (ASCP)
https://www.ascpp.org/
The organization that convened the 45-member international deprescribing task force discussed in this episode.

ASCP Deprescribing Consensus Statement — JAMA Network Open
https://jamanetwork.com/journals/jamanetworkopen/fullarticle/2845497
The parent consensus paper on general principles of deprescribing psychotropic medications.

Dr. Joseph F. Goldberg is clinical professor of psychiatry at the Icahn School of Medicine at Mount Sinai and deputy editor-in-chief of the Journal of Clinical Psychiatry. He also serves as associate editor of CNS Spectrums. His career has focused on the treatment and long-term outcomes of mood disorders, resulting in more than 240 peer-reviewed publications and several books, including Practical Psychopharmacology with Stephen Stahl. He recently completed his term as president of the American Society of Clinical Psychopharmacology.

Ben Everett, PhD, is the creator and host of The JCP Podcast, a series that brings together leading voices in psychiatry to explore the latest research and its clinical implications. Everett earned his PhD in Biochemistry with an emphasis in Neuroscience from the University of Tennessee Health Science Center. Over a two-decade career spanning academia, publishing, and the pharmaceutical industry, he has helped launch more than a dozen new treatments across psychiatry, neurology, and cardiometabolic medicine. His current work focuses on translating complex scientific advances into accessible, evidence-based insights that inform clinical practice and foster meaningful dialogue among mental health professionals.

This transcript has been auto-generated and may contain errors. Please refer to the original audio recording for full accuracy.

00:00 – Introduction and Defining Deprescribing

Dr. Ben Everett: Hello and welcome to the JCP podcast. I’m your host, Dr. Ben Everett. In each episode, we sit down with leading clinicians, researchers, and educators to explore the science shaping mental health care today with a focus on insights that matter most in clinical practice. Today’s conversation focuses on one of the most common, yet least standardized decisions in psychiatry, knowing when and how to discontinue a medication. My guest has spent the past year leading a 45-member international task force to bring greater clarity to that process. Joining me today is Joe Goldberg, clinical professor of psychiatry at the Icahn School of Medicine at Mount Sinai, deputy editor-in-chief of the Journal of Clinical Psychiatry, and associate editor of CNS Spectrums. His career is focused on treatment and long-term outcomes of mood disorders, resulting in more than 240 peer-reviewed publications and several influential books, including Practical Psychopharmacology with Stephen Stahl and Clinical Reasoning and Decision Making in Psychiatry, which was published in 2024. Goldberg recently completed his term as president of the American Society of Clinical Psychopharmacology, where he led the initiative we’ll be discussing today. He was also one of the earliest podcast guests, so it’s especially fitting to welcome him back and to talk about what I believe is one of the most practical consensus statements ASCP has produced in recent years, a framework for when and how to deprescribe psychotropic medications. The paper was published in JAMA Network Open earlier this year. Joe, welcome back.

Dr. Joseph Goldberg: Thanks, Ben. It’s a pleasure to be here and to talk about such an important topic as this.

Dr. Ben Everett: Last time we talked about your leadership at ASCP and the big picture future for psychopharmacology. Today, I really want to get in the clinical weeds with you on something that I think every prescriber encounters, but few have systematic guidance on. Before we dive into this concept of deprescribing, how do you define it, and what does it need in its own word, or why does it need its own word? Why not just say, “Hey, we’re gonna stop this medication”?

Dr. Joseph Goldberg: That’s essentially what we’re doing, although we’re not just stopping a medicine the same way we don’t just start a medicine. We have to do it with planning and forethought, and there needs to be a reason and a rationale and a safety plan in place. And back to the rationale is why do we want to stop a medicine in the first place?

Is it because it’s not working, and so why would you continue to take a medicine that’s not working? Or is it because there are side effects that outweigh the benefits? Or is it because there’s a different medicine that might work better? Or is it because it’s the wrong choice? I’m treating you for depression when you’re really manic, or I’m treating you for anxiety when you’re really psychotic, or because you’re misusing the medicine that I prescribed and taking it in ways that I can’t safely monitor.

Or the medicine’s outlived its usefulness. You’ve had a single episode of depression. You’ve been treated to remission. A sustained period of wellness, like a year, has gone by, and it’s worth asking, do you still need to be on it? We’ve been stopping medicines forever. We just haven’t necessarily thought so much about the way in which we do it, the way I was just describing, or this question of determining endpoints in psychopharmacology. I think the first thing I’d like to say is when we use the term deprescribing, we’re really talking about how do you define and identify a pharmacological endpoint for one of the reasons that I just mentioned. And this is important because it really needs to be non-arbitrary. It’s really a careful reappraisal and reassessment of the treatment in total. I often like to think of medicines like employees on a payroll, and from time to time, it’s a good thing to do a performance review of everybody on the payroll and ask what’s your job, and are you doing it well, and is your job still there, or is that niche no longer present? And if you’re not doing your job well, why is that? Do you need a PIP, or do you need an assistant, or do we let you go? It’s very empowering, I think, for both patients and clinicians to together periodically do these kinds of reviews, performance reviews on a medicine regimen so that everyone is aligned as to what we’re using and why. And we use the term deprescribing with some caution. Historically, the word grew out of the geriatric medicine literature to describe older adults who were on lots of medicines, often non-psychotropic medicines, and sometimes without a lot of attention to the drug-drug interactions or the cumulative effects or the adverse potential for things like falls or cognitive side effects.

In the early 2000s, some authors started writing about how can we simplify complicated regimens in older adults. But then the terminology of deprescribing extended to other areas of medicine too. In neurology, migraine specialists talk about duplication of treatments, like you don’t need to be on two triptans, or if you’re on two anticonvulsants that have anti-migraine properties, are their mechanisms redundant or contradictory? In gastroenterology, if you have reflux, you don’t need to be on a proton pump inhibitor forever. They haven’t been shown to be prophylactically useful. In gynecology, after a certain age, hormone replacement therapy may no longer be that useful, and the risks may outweigh the benefits. It ain’t new for us to invoke this terminology, but ASCP deliberately opted to retain the word deprescribing in part to make it on par with our colleagues elsewhere in medicine.

And then just to define the terminology a little more clearly, the term deprescribing has been taken or used by some circles or communities that speak to the idea that psychiatric medicines per se are inappropriate. The anti-psychiatry factions that are out there and invoke this term of deprescribing to mean not the way I’ve defined it as inappropriate because of the pharmacological reasons, but just because medicines are not good things that people should be taking, and so let’s deprescribe. And that’s very much what we don’t mean at all. We’d like to reclaim the term and bring it back into the fold with the rest of medicine to mean the very deliberate, thoughtful, and purposeful identification of an endpoint and a determination as to whether or not it’s wisest to continue, discontinue, or replace.

Dr. Ben Everett: Excellent. That was that you covered a lot very seamlessly there. Yeah, no, that was really good. I’m just taken back by, how flawlessly you really covered a lot of different areas in medicine and that really resonates with me. I’m sure it resonates with the audience now. But I think it begs the question what made this the right time to take this on through ASCP?

07:00 – Why Now: ASCP’s Task Force and the Political Landscape

Dr. Joseph Goldberg: Funny you ask. A few years ago, as we were mapping out the direction that we wanted to go in with ASCP, when I assumed the presidency I thought it’d be a good thing for ASCP to make its mark in the field by developing some task forces that would issue authoritative expert commentaries on a variety of issues.

Deprescribing was one. I’ll talk about how that arose in a second. But we also did a task force on the disparities between ketamine use in real-world clinical settings compared to what the evidence base shows, and we’re just finishing that up at the moment. We have a task force on reproductive psychiatry and women’s health, and some of the ways in which the literature doesn’t give us as much information about, say, what to do with a woman with a mood disorder who wants to become pregnant or is pregnant, and should you or should you not tinker with the pharmacology.

We have a task force that’s also looked at polypharmacy. We just presented some data on that at the ASCP meeting in Miami last month. We have a task force on pharmacogenetics. We try to identify areas in psychopharmacology where the literature falls short and practitioners need some assistance or navigation through the disparity really between what do we actually do with patients, what does the literature tell us or not tell us, and how to fill in those missing pieces. And deprescribing suggested itself at a time when a few things were coming into the forefront in our field. One was some popular media coverage that was leaning a little toward that use of the term deprescribing that I mentioned before, where it was more the anti-psychopharmacology sentiment, that medicines are just inherently, things to be distrusted.

And there were people that were writing things in the lay press about the difficulties they were having coming off of medicines and how challenging it can sometimes be if a medicine is stopped, without a thoughtful plan. You could get some discontinuation symptoms or rebound effects.

And so there was some dialogue occurring in the lay literature around that, as well as people in the patient community that were saying, I think I’m better off without medicines than with them. Maybe people out there should start questioning whether or not medicines are really such a great thing. There was that backdrop, and we felt almost a professional obligation to lend some expertise on these important questions as to when medicines might be inappropriate. Heck, if you’re getting treated for something you don’t have, that’s inappropriate. If there’s a better choice for you than what you’re on, perpetuating it is inappropriate. If you’re taking medicines that have bad interactions with each other or side effects that outweigh benefits, that’s inappropriate. So was the spirit of our starting this task force and doing these studies. And then by sheer coincidence, the national landscape moved in such a direction that the Department of Health and Human Services began to articulate concerns about the lack of endpoints in psychopharmacology and questions about whether or not medicines, particularly for depression, are not being carefully monitored and thoughtfully implemented especially in kids.

That was a concern that was raised, within the administration over a year ago, in a statement that was made, I think, in February of twenty twenty-five. And then more recently, the Department of Health and Human Services has raised the question of whether antidepressants are over-prescribed, whether or not psychiatry as a field has fallen short of articulating guidelines for when and how an antidepressant ought to be stopped and if it should be stopped. There was a sort of a coincidence in timing. All of our papers came out right at about the same time that HHS offered this announcement. And I won’t say we’re strange bedfellows, but I will say that there is an articulation coming from HHS at the federal level about concerns of antidepressant safety and their use in situations where they may not be appropriate.

But ASCP’s concern has been to not overshoot the mark by saying antidepressants are over-prescribed, but rather, are they being prescribed for reasons other than they’re known to work? Antidepressants don’t treat adjustment disorders. Antidepressants don’t treat normal grief. Antidepressants don’t treat mania. Antidepressants have a particular safety profile. We’ve wanted to capitalize on the trove of information that’s come from our consensus statements to support an agenda that let’s just get everybody off antidepressants because not good, but rather to prompt thinking on the part of both patients and practitioners. Is a medicine indicated? If so, are its benefits clear? Do those benefits outweigh side effects? Is it being taken properly? Does your condition evolve over time? And last but not least, on the whole depression is undertreated in the United States and worldwide for lots of reasons. Less than one in five people with major depression gets even minimally adequate treatment, and so it’s conceivable two things are true at once. Antidepressants might get overprescribed for conditions other than major depression or anxiety disorders, when other things might be more appropriate. At the same time, they’re probably greatly under-prescribed and under-dosed and not necessarily followed carefully, maybe in part a lot of antidepressants are not prescribed by mental health professionals, but by our colleagues elsewhere.

And so the monitoring and the surveillance and tracking the evolution of a psychiatric condition may not be as carefully done as one might wish. Another lengthy response to all kinds of factors that led to our interest in wanting to formalize our studies in deprescribing.

Dr. Ben Everett: All right, so it seems like it was both a very deliberate decision to go about starting this task force, but at the same time, it’s really the timing could not be better, because y’all could just now be starting, what would be a two-year process. And so to have this in the mainstream right now is really good. The task force used an iterative Delphi methodology. You had forty-five international experts. With that comes a lot of experience, but also a lot of opinions. Even though, that being said, y’all managed to reach consensus on eighty-eight percent of final statements, which is really good for a Delphi process, I think. What surprised you most about where the group agreed, maybe where you didn’t reach a hundred percent consensus or any other thoughts on the process in and of itself?

13:30 – Delphi Methodology: Consensus and Divergence

Dr. Joseph Goldberg: Yeah. I think the most interesting things, if one was to read the study in JAMA Network Open, the parent study about general principles, and then there was a second paper in British Journal of Psychiatry that came out a month later just on mood disorders, and there was a third paper that was on the deprescribing of stimulants in adult ADHD that came out just about a month ago in European Neuropsychopharmacology.

And we’re now finishing up our antipsychotics deprescribing paper that hopefully we’ll be sending out shortly. It was a kind of a series of things. But I think what we found most interesting were the areas of disagreement or failing to reach consensus and reasons for that. I’ll give just a few examples that I think speak to the complexity.

One overarching issue and a challenge with consensus statements is oftentimes in medicine and in psychopharmacology in particular, things can be so nuanced and individual specific and context-driven that it gets hard to make a general rule to say, “Here’s a patient. Should you stop their medicine?”

And I think every one of us who filled out these Delphi statements was thinking it depends. I might do this in this situation. In that situation, I might do something else. Depends on patient preference, depends on risks and alternatives. I think it’s very hard to formulate a consensus opinion statement for something so nuanced as psychiatric treatment because there’s so many variables.

I was surprised that we got consensus agreement on so many.

15:30 – The Lithium Overdose Dilemma: When Stopping Defies a Rule

Dr. Joseph Goldberg: But here are some examples of areas where there was divergence. One that I thought was interesting was the statement that we’d ask if a patient with a mood disorder overdoses on lithium, and gets hospitalized from a suicide attempt, should you then deprescribe the lithium and say, “We’re not gonna give it back to you,” on the premise that look, you just made a suicide attempt, ergo, it wasn’t helping, it? And it’d be dangerous to give you that. And we actually had an enormous amount of diversity of opinion on this question.

Some respondents said what I just said, which was, yeah, of course stop the medicine. Look, it could happen again. It wasn’t working. Another view was to say how do you know it wasn’t working?

The patient made a suicide attempt, but over the course of their lifetime, for all we know, lithium was actually reducing their impulsivity and proclivity toward making multiple suicide attempts. So the patient made an attempt. Maybe if they’d never been on lithium, they would make ten attempts, or they would have completed the act before, or an attempt may have been less lethal. Context depends. A third had to do with what are the alternatives? If you have something better and safer, sure, consider that. But if you don’t, then it would be a bit concrete in your thinking to automatically and reflexively say we need to stop this medicine. And many of us have had patients where this scenario has occurred, and they go in the hospital, and the inpatient team reflexively says you can’t be on this anymore because look what happens.

That was one instance where it seemed like deprescribing was defying a general rule. I think another where we ran into some was the balance of, in particular, metabolic risks with efficacy, and in particular, with agents that are known to have a somewhat higher cardiometabolic liability.

That is a number of—not all, but a number of—the atypical antipsychotics when being used for non-psychotic disorders like adjunctive treatment in major depressive disorder, where there are several that are FDA-approved, or even in maintenance treatment of bipolar disorder, where some have been approved and some have not for maintenance treatment. But the question of when do you rock the boat, when do you say, “You’re better now, but at some point, do I declare the adjunctive atypical antipsychotic probably has gone as far as it needs to go, and so we’re gonna jettison it like a retrorocket and just leave you on the mood stabilizer,” particularly if you’re having metabolic side effects, significant weight gain, or for that matter involuntary movements, and we have to think about the risk for things like tardive dyskinesia over time. We don’t have endpoints established from randomized discontinuation studies to teach us continue the adjunctive antipsychotic for X amount of time and then stop it. There was a lot of trepidation among some of our respondents about saying do I really want to rock the boat if the patient’s doing well?

If they’ve had a really rocky course of illness previously, I’m not sure I want to automatically stop the medicine. Maybe that’s a candidate for more aggressive antidote strategies. I could use metformin or GLP-1 agonist or phentermine or anything you like to try to counteract, say, metabolic side effects, or if someone’s developing signs of a movement disorder or tardive dyskinesia, at what point do you say let’s retain the medicine because we think the benefits are likely to be substantial, so we’ll just add a VMAT2 inhibitor. All of these, I think, there was unanimity about shared decision-making with patients and to make the patient an active partner in the process. So we don’t just say, “I think you should continue,” or, “I think you should deprescribe,” but to make sure the patient has knowledge and agency to be able to express a preference and share in the decision-making process.

But there were no black-and-white absolutes, but there was diversity around some of those kinds of questions. And I think, what’s most interesting to read about in these papers is the areas such as those where there is diversity of opinion and the reasons for it because it highlights the complexity of what we’re doing rather than saying cookbook style do X.

Dr. Ben Everett: Yeah, I think that’s really good. Look, I wanna come back to how you started that response with you have a patient who attempts suicide by overdosing on a medication, in your example, lithium. We can kinda look at that at the reciprocal side of that as well in terms of deprescribing. That would be maybe you have a schizophrenic patient who’s on risperidone and, they end up in the hospital due to a psychotic break, and somebody might just say obviously, the risperidone isn’t working anymore because they had a psychotic break. But we actually need to ask some questions like, “Hey, is the patient actually being adherent to their medication?” Tell me about how y’all explore this idea of adherence, medication and what y’all look through with that.

20:00 – Adherence, Pseudotreatment Resistance, and Schizophrenia

Dr. Joseph Goldberg: Yeah, very important points there, Ben. In our JAMA Network Open paper in our general principles, I think in table one we did talk about as a precursor or prerequisite before whipping out your deprescribing pad is to assure things like was there an adequate trial, an adequate duration, and were there any mitigating factors that might have mucked things up?

Risperidone, for instance, what if the patient was on a potent 2D6 inducer like carbamazepine? They never got to first base in terms of an adequate dose, or if they’re a poor metabolizer, and so you have more unpredictability in their ability to convert risperidone as a prodrug to paliperidone, the active metabolite.

We wanted to prompt thinking about: pay attention to what you’re doing, and as you’re deciding if a medicine is or isn’t working, is that because of poor adherence? Is that because of a pharmacokinetic or a drug-drug interaction or conversion of prodrug to active metabolite issue? And also patient preference.

Given that in serious mental illnesses, up to eighty percent of people don’t take their medicines as prescribed, and oftentimes that could be quite substantial. It does run the risk of erroneously concluding, this drug didn’t work for you. I guess the phrasing would be it failed to work for you rather than it actually didn’t have the adequate trial.

One of the concepts in our body of works on this is to give great thought to adherence and to not forget about the role of long-acting injectable antipsychotics, particularly to rule out pseudotreatment resistance when in doubt. And in the body of studies that we did on deprescribing strictly in schizophrenia, two or three main points came up.

One was this very point, to be very aware of and attuned to the possibility of pseudotreatment resistance. And some of our panelists felt the only way to do that is with a long-acting injectable. Before you make strong conclusions about shifting to something else, you’d like to have some confidence.

And that also may mean, interviewing a collateral historian or doing pill counts. How long have you known the patient? If this is a new patient versus someone I’ve known for years and years. The thoughtful practitioner, I think, has to factor all those things in. A second issue with deprescribing here is if there really truly is lack of response to one or two atypical antipsychotics. In schizophrenia, in particular, deprescribing really cannot mean stop the drug and do nothing. That’s like saying to someone with diabetes, “Stop your diabetes regimen and do nothing,” or someone with hypertension or COPD. Deprescribing, particularly in schizophrenia, can’t say this strongly enough really means discontinuation with replacement.

There’s a couple of instances where I think that’s especially important. One is the non-adherence piece, where an LAI could help to assure pseudotreatment resistance. Second is the all too common scenario of someone who’s on two or more atypical antipsychotics, and they’re still symptomatic. And here we also had a fair amount of diversity in our respondents. Some felt that there are times where two atypical antipsychotics might be synergistic, in particular, clozapine with a drug like aripiprazole or maybe another partial agonist where there are some data. But all too often, patients may wind up on two non-clozapine atypical antipsychotics, and all they’re getting is additive side effects.

The issue there is if you’re not seeing clear benefit with two or more atypical antipsychotics, especially if there’s duplication of mechanisms, which one should you deprescribe, and which one should you optimize? And for that matter, you should probably optimize a first drug before contemplating the addition of a second antipsychotic.

And then the third final point around this was if someone is on one or two or more adequate trials of an atypical antipsychotic and not benefiting, there was actually some consensus saying, assuming you’ve ruled out poor adherence, and you’ve gone down the LAI route, and the patient is still symptomatic, strong consideration should be given to deprescribing what the patient’s taking and consider a clozapine trial.

And our task force seemed to have a strong sentiment that, just a sentiment, but I think pointed out, clozapine is widely still underutilized in treatment-resistant schizophrenia. Something like one percent of American mental health prescribers prescribe clozapine. And now that the REMS stipulation has been lifted, we still have to monitor the hematologic safety, but it’s not quite the same as it was previously.

And I think there was a real sense that we’re underutilizing clozapine and overutilizing polyatypical antipsychotics, and that’s an instance where deprescribing and replacing becomes important.

Dr. Ben Everett: Yeah, I think that’s spot on. All right, so let’s kind of segue into clinical nuance, and some of your answers have gotten to some of this nuance before. But this is really an area I think generates the most uncertainty in practice, where you have complex polypharmacy patients. They’re on multiple agents for years. We understand that this patient population often tends to be heavier weight anyway, so they’re at higher risk for metabolic syndrome, diabetes, heart disease. And as patients age, that comes with a lot more medications. Specifically, the task force looked at this and how did y’all, come up with through these types of complex polypharmacy patients?

25:30 – Complex Polypharmacy and Medication Performance Reviews

Dr. Joseph Goldberg: The hard way. Our next effort in this task force that Dr. John Rush from University of Texas is leading the way with, is to come up with a somewhat protocolized assessment, the performance review as I was alluding to earlier, looking at the totality of a patient’s regimen on some periodic basis, not ad hoc, not just when they get hospitalized, not just because oops, they were off their meds for a week, so that occasions a kind of a de facto serendipitous, “Let’s review what you’re on and see if you should restart or not,” but more like in the business model of a deliberate performance evaluation of everything. That has to include the non-psychotropic medications. It does require that the clinician reviewer be aware of the drug interactions with the non-psychotropic medications.

Those could be pharmacokinetic. “Ooh, I didn’t realize you’re on a 2D6 inhibitor and I’m giving you venlafaxine,” or, “I’m giving you risperidone or duloxetine or atomoxetine.” But also the pharmacodynamic issues. “My, oh my, I noticed that you’re on a really potent alpha-1 blocker for your hypertension, and I’m also giving you iloperidone. And you’re on trazodone. That’s a real powder keg of alpha-1 blockade. You’re lucky if you can get vertical.” Some either deprescribing or rethinking of that regimen makes sense. “You’re on lithium and, you’re on a low sodium diet, so your lithium level’s gonna go up,” or, “You’re on a thiazide diuretic,” or there’s all kinds of potential for drug-drug interactions that we don’t always think about.

Also, if we stop one medicine that happens to be an inhibitor or inducer of a substrate of another medication, what’s the implication there? If I stop your valproate and you’re on lamotrigine, now I’ve just slashed your lamotrigine level in half. Am I gonna compensate for that? If you’re on lithium and you’re going on a non-steroidal, what do I wanna do there?

In some ways, this is a bit beyond just deprescribing, if you like. It gets into the very mindful curation of an overall regimen. In people with serious mental illnesses, extensive polypharmacy isn’t rare. In fact, the paper we just finished on complex polypharmacy made the point that the sheer number of medicines you’re on isn’t necessarily a reflection of quality of treatment.

We look at a patient who comes to us, oh, they’re on umpteen medicines, how sloppy that is. And sometimes that’s true. But it also could be just a marker for chronicity and severity and even desperation. You’re trying to help alleviate suffering, and the patient may accrue a number of medicines, and sometimes an impartial outside consultant can look them over and say, “Here are some redundancies. Here are some conflicts. Here are some ways to simplify the regimen. Here are some unnecessary side effects,” which can be very helpful, not with the goal of just trying to slim down the number of pills. That’s often a laudable goal when you can do it, but also to make sure that every member of the team is earning its keep and is properly being utilized, and that means taking into account both the non-psychotropic and the psychotropic meds.

Dr. Ben Everett: So you brought up valproate a minute ago, and I want to get into valproate a little bit because I know you and Dr. Marlene Freeman have different opinions about this and but both very thoughtful differing opinions, and I think this gets to a lot of the nuance in clinical practice. And this was also a strong opinion piece in this paper. Specifically with respect to valproate and women of childbearing potential, how do you think about that? And tell our listeners the conclusion y’all were able to come to here.

29:30 – Valproate and Women of Reproductive Potential

Dr. Joseph Goldberg: Sure. I think our Delphi statement was something on the order of valproate should be deprescribed in any girl or woman of childbearing potential who’s taking it, period. And the thinking behind that was that about half of pregnancies are unplanned, and valproate is a known teratogen in terms of neural tube defects and central nervous system development, and potentially also postnatal complications and neurodevelopmental problems.

Most reproductive psychiatrists such as my dear friend and colleague and boss, Dr. Freeman, the journal, fairly strongly, I think, advances the position just pretend it doesn’t exist. Take it off the table entirely. Our task force fell short of consensus on that declaration. Numerically, there was a majority of respondents who agreed, “Yeah, you should try to not use valproate, and if a woman who could get pregnant is taking it, you should probably try to get rid of it.” But we fell short of consensus. We define consensus as at least seventy-five percent of us agreed on something.

And the reasons why some people didn’t go that far with it was back to this issue of nuance. What do you do in a scenario of, let’s say, someone, a woman with, very significant either mood symptoms or impulse control symptoms where other things have not worked except for valproate, or where a woman expresses the preference to continue it because she’s found it helpful, but pinky promises to not get pregnant. And it does take the patient preference out of the loop or sidesteps the dialogue that might occur. In a perfect world, we should have no side effects. We certainly shouldn’t have teratogenic effects. And if we have a better alternative, I think this is true with everything in our pharmacopeia, when you have a better alternative, you should use it. But if someone is stable on a regimen and other things have not proven to be helpful, it’s a real gamble to deprescribe something and cross your fingers that whatever you replace it with is gonna work at least as well, if not better. That’s a dialogue to have.

While I personally don’t disagree with Dr. Freeman in the perfect world, valproate shouldn’t be used in women of reproductive years, there are scenarios where it becomes more individualized, and this requires oftentimes, consultation with someone with expertise, second opinion and, all-important patient awareness.

I can think of a patient or two I’ve known over the years back before we knew much about the teratogenicity with valproate who was told by their practitioner if you take valproate and there’s a nervous system problem, then just get an abortion. Yikes. That’s not the way we want to think about these things.

But at the same time, I think the flip side of that is look if you’re really benefiting from this but there is a known risk, how do you want to think about undertaking that risk? Whether it’s the risk of teratogenicity with valproate or whether for that matter it’s the risk of metabolic dysregulation with an atypical antipsychotic. Everything has its risks. My own view, and I think the reason why we fell short on consensus was because there are nuanced situations you can’t just make a declaration. You have to say, “In the perfect world, this is what we’d want to do,” but it’s not a perfect world. And I think my response in the journal was something like living in a pharmacologically imperfect world or something like that. You just have to set all the information out, make the patient as aware as possible of what the risks and benefits are, and then help them make the best decision for themself.

Dr. Ben Everett: Yeah, it’s spot on, man. I couldn’t agree more because you’ve got to take the patient preference into it and, like you said, if it’s the only thing that’s helped, I think there’s a lot to that. But yeah, certainly the idea of if you get pregnant on it, you just have an abortion, that’s…

Dr. Joseph Goldberg: No.

Dr. Ben Everett: Not on video now. Yeah, our listeners can’t see the face I made when you said that, which was the way you intended it. All

Dr. Joseph Goldberg: Yeah.

Dr. Ben Everett: Let’s move on to the psychological dimension. The paper really devoted a good bit of space to this, and that’s the idea. You alluded to it earlier. Patients have an unconscious relationship to their medications. They certainly have a conscious relationship as well. Attachment theory, medication as transitional objects. Is that framework useful to a busy prescriber, or do you think that’s purely an academic thought?

34:00 – The Psychology of Stopping: Medication Attachment and Placebo

Dr. Joseph Goldberg: Oh, I think it is among the most fundamental for several reasons. First of all, in the world of depression, as many of our listeners I’m sure know, a good thirty to forty percent of response is technically placebo response to the inactive ingredients of the drug. There’s expectancy biases. There’s numerous psychological levels around what the taking of a medicine means.

It’s a tangible statement that I perceive you’re suffering, and so I am gonna give you something to try to ease your suffering. That’s a very doctorly, old-time notion. And to say, “I’m gonna take away this tangible thing that’s meant to alleviate your suffering,” could mean, “I think you’re doing pretty well and you don’t need this anymore.” Or it could also mean, “I’m not sure you’re really suffering as much as you think you are. I don’t think you need this. I think you’re stronger than you think you are. Maybe I’m questioning the validity of your suffering. Maybe I’m gonna push you out of the nest and tell you you can fly on your own wings and you don’t believe me. Maybe you think you’re gonna collapse under your own weight.”

I had a patient once in a clinical trial who was doing spectacularly well, was having tons of side effects that we were treating. It was a depression maintenance trial. At the end of the study we broke the blind and yeah, it turned out he was on placebo that whole year in which he was euthymic and having all kinds of sexual side effects and sedation side effects and weight gain. And I had to tell this fella, A, you’re not depressed, B, you don’t need a psychiatric medicine because you weren’t on one, C, you weren’t having side effects from an inert substance, and D, you don’t need mental health treatment. You’re well. He got really upset about that. It was a kind of a deprescribing.

And while he intellectually knew and understood that it was not subterfuge, he knew he could be on a placebo, he was astounded to learn that’s what he was on. Anyway, he got really upset after that and ended up getting referred for psychotherapy, and he felt like a manipulation even though intellectually he knew that it wasn’t. There’s an awful lot of meaning attached to medicines, and I think the busy practitioner who just reflexively in a brief visit says, “All right, your PHQ is still in the teens. Let’s raise your dose. Come back next time,” is missing an awful lot, particularly if medicines are not yielding as great a benefit as we want. The effect sizes of a lot of psychotropic medicines, as you well know, Ben, are fairly modest, and so we actually are trying to capitalize on the placebo effect. Researchers try to minimize it. But if I can get a little more improvement from you because of the ways in which I know that you think about medicines and what they can do, or the ways in which I might say, I really think a psychotherapy is a better augmentation than something else, we’re appreciating the complexities of what goes on for the patient rather than just treating, say, depression like strep throat and switching your amoxicillin to Augmentin.

Dr. Ben Everett: Yeah. Thanks for that. All right, moving on to pharmacokinetics. There’s an important point in the paper about pharmacokinetics and certain drugs have longer half-lives even if you’re not using the LAI version. The idea that drugs like fluoxetine, aripiprazole, cariprazine, you can essentially auto-taper. Or they essentially, the pharmacokinetics allow these drugs to be auto-tapered on their own. Explain that maybe a little bit more, go into some weeds on what y’all mean by that.

37:30 – Long Half-Life Drugs and Auto-Tapering

Dr. Joseph Goldberg: Yeah. I don’t think we got consensus on that one either, if I remember correctly. Even though what you just stated is pharmacokinetically correct. Look, take a drug like cariprazine, which has its two active metabolites, the second of which, didesmethyl cariprazine, has a half-life of one to three weeks. It takes five half-lives for terminal elimination of the drug. That means that if I was taking cariprazine today on June 8th and I stopped it cold, it is still measurable in my system by Labor Day, which is two and a half months from now. And same, fluoxetine has norfluoxetine. It’s got a one-week half-life, times five, is five weeks. On one level you could say, if I have a drug with a fairly long terminal elimination half-life, an abrupt cessation is not so likely to cause rebound effects or discontinuation phenomena as would be the case with a long half-life drug. And so in principle at least you should be on pretty good ground.

If my patient’s on one of those long half-life drugs and missed a day or two and contacts me and says, “Oh, my gosh, Dr. Goldberg, I’m feeling terrible,” I’m half thinking, “That doesn’t make pharmacokinetic sense. You’re still at steady state.” But I’m not gonna say that to the patient. I’m gonna say, “All right, let’s make you as comfortable as possible.” But I think part of why we didn’t get consensus is a lot of our respondents said that’s true in principle, but in reality you never know. Just to be safe. And here’s another example of maybe some disagreement among the experts. Some felt no one will ever fault you for wanting to go slow and do a gradual discontinuation or gradual cross-taper, as opposed to saying, “This will auto-taper, and as long as you’re not introducing something that’s conflictual, feel free to just stop the long half-life drug and bring in, the new agent as long as it’s not conflictual.” I think there was just a feeling or a sentiment almost subjectively that you never know if things might go awry. There could be other factors like, let’s suppose it was an anticholinergic drug and you start to get rebound cholinergic effects. Or if it was an alpha-1 blocking drug with a long half-life, and you stop it, could you get rebound hypertension? And I’d say if it’s a long half-life, that’s not so likely. But some of the experts said yeah. In principle, that’s an academically valid point, but in reality no one will ever dispute slow tapers per se.

Dr. Ben Everett: Yeah. All right, the flip side of that is a discontinuation phenomenon. We can talk about brain zaps, dizziness from short half-life serotonergic drugs. Is there a commonly recommended strategy for substituting maybe something with a longer half-life like fluoxetine in that context?

40:30 – Discontinuation Phenomena, Brain Zaps, and Hyperbolic Tapering

Dr. Joseph Goldberg: Yeah. Another area where we didn’t get consensus. All these juicy gems are coming out. There’s this sort of lore, I think, in pharmacology as to what you just said. Somebody’s on a short half-life SSRI like escitalopram or sertraline, just switch them to a long half-life drug, and fluoxetine with its active metabolite is the classic example. And it was about 30 years ago, I think, that Alan Schatzberg published his paper in the Journal of Clinical Psychiatry, what’s come to be known as the Prozac Bridge, which is do exactly what you just said and just swap the patient directly over to 20 or 40 milligrams of fluoxetine, depending on what their dose of the other drug was. Give them a couple, three weeks, and then stop it, and it will auto-taper, voilà. And I know plenty of psychopharmacologists who advocate that and think of it as a heroic, or not even heroic, but just simple fix, clever hack. The reason we didn’t get consensus on that one, although there was a majority of respondents who felt that was a good idea, was there’s never been a study of that. And our panelists are such dyed-in-the-wool empiricist, science-minded people that people were saying, “Yeah, that sounds like a good idea, but till someone’s actually studied it, we just don’t know.” You can’t reflexively say, “Let’s swap to the long half-life.” You can say this is a viable option, but it’s got a rationale, but you can’t bank on it. Other things that were brought up on this issue of how to minimize discontinuation phenomena, which by the way, are not withdrawal phenomena. I know there’s some sentiment out there that antidepressants are addictive. They’re not. Or that they cause craving. They don’t. Or habituation. They don’t. But your body does become accustomed to the higher transmission of serotonin. If you take that away, then you can get specific kinds of symptoms like vertigo, dizziness, nervousness, in particular usually in the first couple of weeks, and then it dissipates, and it’s rarely, if ever, medically important.

But it’s uncomfortable, and we don’t want to make our patients uncomfortable. And relapse of depression can happen too, but that’s usually not a discontinuation phenomenon. That tends to come a little farther down the line. The working idea here is if a patient expresses great concern about this potential, and again, we’re back to expectancy bias and the psychological meaning of medicines or their deprescribing, slow it down. There is this notion of what’s come to be known as hyperbolic tapering as opposed to linear tapering. Linear tapering is so you’re on a hundred of sertraline. I’m gonna drop it to seventy-five for five half-lives, then fifty for five half-lives, then twenty for…, and then you stop it. That’s linear progression.

Hyperbolic argues the lower you get in the titration, the more there’s sensitivity of receptors to ligand binding, and so you actually have to slow it down at the very end, almost in a kind of an asymptotic way. I’ll drop your sertraline from one hundred to seventy-five to fifty and then to thirty-seven point five, and then maybe to twenty-five and then twenty-two point five and then twenty. And then I go to a compounding pharmacy, and I slow it down. We didn’t get consensus on that one either, but the sentiment that if there’s concern about rebound effects or discontinuation phenomena, you can certainly slow things down. And if a patient is really very concerned, it’s not unreasonable to invoke the notion of hyperbolic tapering. It’s a little bit clunky. You have to get a compounding pharmacy or get a liquid formulation, and now and then that works. Last thing I’ll say on this, and no one’s gonna like what I’m about to say, but there are certain ailments that we treat where somatic symptoms are part of the ailment, and people may be particularly vulnerable to somatic sensations.

I’ll say to patients, “There’s this part of your brain called the somatosensory cortex, which is like a microphone, and it picks up stuff.” And for some people, that’s tuned to eleven. And so little bodily sensations may amplify, and it’s not made up in your head, but it’s how your body perceives sensations. And to attribute to the medicine discontinuation protracted effect in someone who may have baseline symptoms that involve somatic features may be denying part of what the ailment is. I’m always very conscientious about the extent to which somebody with any psychiatric complaint has a lot of physical aspects that go with that because no matter what I taper them off of, they’re probably gonna have some amplification of, “Ooh, I can feel my ears popping, and I can feel my heart beating, and I think I’m having this and that.”

And so we want to validate the distress, but we don’t necessarily say, “Oh, that was the discontinuation effect.” We have to very delicately, I think, anticipate how to give people a soft landing.

Dr. Ben Everett: Yeah, that’s really good. All right, let’s transition into a couple of special populations, and then we’ll get ready to close this one out. But pregnancy is a special situation where, patients often will end up stopping medications. The data are very complex here, a lot of nuance here. Y’all were emphatic that relapse risk has to be weighted heavily. What’s the practical message for clinicians that are dealing with pregnant patients?

45:30 – Pregnancy, Special Populations, and Relapse Risk

Dr. Joseph Goldberg: Sure. Mom’s mental health is the priority, and in the absence of known consequential adverse effects to fetal development or the potential for obstetrical gynecological complications like low birth weight or preterm delivery, it is rare to advocate deprescribing of psychotropic medications, with a few exceptions like valproate maybe. But for the most part we really don’t wanna tamper with mom’s mental health. An untreated mental health condition in a pregnancy by definition makes a high-risk pregnancy. And particularly in the case of depression, where I cannot think of a hazardous antidepressant in pregnancy. I can think of ones that have years and years of research in pregnancy and others that are newer, it is just simply wrong information to say, “Oh, you’re pregnant, you must now stop your SSRI.” That is just wrong information.

Our women’s reproductive psychiatry task force is calling this out in particular in the paper that we’re working on for that topic as well. I think there’s a few other rules of the road, though, in pregnancy, so you don’t wanna change things around much if someone’s pregnant. That means I’d rather not introduce multiple unknowns. If you’re on a particular medicine, during organogenesis in the first trimester, I’d rather not add more unknowns to the mix. Let’s keep you on what you’re on, try to get you through the first trimester before we start to mess around with things.

But everything is a risk-benefit, and unless there are known risks, it’s really gambling to stop something. One issue that came up is stimulants in ADHD, and here our reproductive psychiatry members were vocal in saying it depends. There are not particularly notorious teratogenic risks. There may be some small signal of cardiovascular development issues with some stimulants. However, if Mom has had three car accidents from texting when she’s off her stimulant or is accident-prone or is not able to function at her job and it’s in jeopardy, then those are very salient risks of not continuing the stimulant. Back to the dialogue, patient involvement, risk-benefit, and patient preference. I think our job as clinicians is to let the patients know, “Here are the risks, here are the benefits of untreated mental health symptoms. Here’s the extent to which what you’re taking is benefiting you.” If there are known risks, like we talked about with valproate, we might wanna, give due consideration. But if there are not known risks then to say, “I just wanna stop a medicine,” is potentially incurring some hazard. And if we are gonna deprescribe a medicine, we wanna do it safely and carefully and have a safety plan in place if someone was to relapse.

Dr. Ben Everett: Yeah, all right, two more questions. This paper, it really makes a pointed observation about the way psychiatrists are trained and residency programs provide almost no formal education around deprescribing. If you could add one thing to how psychiatrists are trained, what would it be?

49:30 – Training Gaps and the Future of Deprescribing Education

Dr. Joseph Goldberg: Oh boy. That’s a toughie. There’s so many ways.

Dr. Ben Everett: Maybe I’ll say with respect to this conversation.

Dr. Joseph Goldberg: Okay, thank you. I have to certainly speak to the amazing work that our colleague Matt Macaluso and Rick Shelton have done in assembling the new ASCP curriculum for psychopharmacology residents. It’s on our website, ascpp.org, and it’s just a marvel of up-to-date information about the best training. It’s got an interactive artificial intelligence component. We’ve got just a tremendous number of experts. I think that’s a really good place to get information. But on the topic of deprescribing, yeah, straw poll clinicians and most will say, “I was never taught how to discontinue a medicine,” formally speaking. And the FDA doesn’t require that drug companies demonstrate how do you discontinue something. Although, certain drugs, if you stop them too quickly, you can get rebound effects. Lithium, if you stop it sooner than two weeks, you’ll triple the odds of relapse in bipolar disorder. One has to be watchful for what is the proper way to taper off and then to monitor.

You don’t just say, “Okay, you’re done. Bye-bye.” If you’ve had multiple prior episodes, your risk for relapse is substantial. We are developing a module in our advanced course at ASCP on deprescribing that will include some of these features. And ASCP is working on this, as I say, protocol for how to deprescribe. That means to say, patient’s gonna get homework. They have to track their risk for relapse, particularly in the first three to six months after stopping something. That means, if you wanna deprescribe, we do it with informed consent. You need to know your chance of relapse based on your prior history, prior episodes, prodromal symptoms that might come up, what to look for, so that nobody gets ambushed by a relapse. Proper deprescribing is work. It is not just stop your meds. It’s doing it safely and carefully, and there has not been nearly as much formal education about that, and that’s one of our goals, is to make that better for the future.

Dr. Ben Everett: Yeah, I’ve been very impressed with the Edyou, and that’s E-D-Y-O-U program. I think it’s great. All right, final question for today. If a clinician that’s listening today, busy clinician, if they only take away one thing from today or from the consensus statement, what would it be? What should it be?

Dr. Joseph Goldberg: Think of medicine regimens as employees. On some periodic basis, do a review. Make sure both you and the patient know what everybody’s job is intended, and how good a job each component is doing, no medicine should have job security ad infinitum or definitely, “Oh, I see so many people in consultations, and I’m on this medicine. I don’t know why, and my last prescriber was reluctant to stop it, but I’m still symptomatic.” We want to be rather, I won’t say ruthless, but thoughtful about does each medicine in a regimen earn its keep and have its place and get its contract renewed or not? ‘Cause if not, there may be a better candidate for its position waiting in the wings to audition for its job.

Dr. Ben Everett: Excellent. Dr. Goldberg, been a great conversation as it was previously. I really applaud the work that you and the task force are doing, all these different areas of psychotropics. And it does something, at this point, it did something a little unusual for a consensus statement. It treats the clinical decision to stop a medication with the same rigor and intentionality that you’re really supposed to bring to starting a medication. And that framing alone, I think, is worth the price of admission, of reading the article. And I loved your analogy about the employee evaluation. That was really good as well. I think it’s something people can think through and it’s a quick reminder type of conversation.

Dr. Joseph Goldberg: Indeed.

Dr. Ben Everett: All right. This has been the JCP podcast. Insightful, evidence-based, human-centered.