Jo Carlowe Hello, welcome to the Papers Podcast series for the Association for Child and Adolescent Mental Health, or ACAMH for short. I’m Jo Carlowe, a Freelance Journalist with a specialism in psychology. Today, I’m interviewing Dr Robert Penfold, Senior Investigator at the Kaiser Permanente Washington Health Research Institute, Seattle, USA, first author of the paper, “Safer and Targeted Use of Antipsychotics in Youth an Embedded, Pragmatic Randomised Trial,” published in the JCPP. Rob, thanks ever so much for joining me. Can you start with an introduction about who you are and what you do? Dr Robert Penfold Sure. So, I’m a Mental Health Services Researcher. I primarily focus on child and adolescent mental health. I’m a member of something called the “Mental Health Services Research Network,” in the United States, which is an affiliation of health systems across the US that collaborate on research projects. And so part of the work we’re going to talk about today is based in that Mental Health Research Network. Our group studies ways to improve access and quality of mental health care. My goal is to improve mental health outcomes for young people, and we do that in a variety of ways, including, risk stratification, making machine learning models to predict people who are at risk of poor outcomes, designing and conducting clinical trials to test new interventions to improve outcomes. Jo Carlowe Hmmm, excellent. Thank you ever so much. I’m aware there is concern about the overprescribing of antipsychotic medications to young people. So, before we go into the detail of your paper, can you set the scene for us? What is the culture around prescribing to young people, and are the concerns about overprescribing justified? Dr Robert Penfold You know, in the early 2000s, when I was a, you know, dewy eyed Postdoctoral Fellow, antipsychotic medications were, sort of, just – the second generation of antipsychotics, in particular, which were thought to have milder side effects, were coming on the market and being prescribed. And, you know, I talked to Psychiatrists who are prescribing these medications, they really thought that they would be miracle drugs… Jo Carlowe Right. Dr Robert Penfold …based on the clinical trial evidence. And so they’re, you know, starting in 2000 or so, and up to 2015, rapidly prescribing. And, you know, the drug that was meant to – and designed to treat psychotic desires, or later, you know, mania associated with bipolar disorder, came to be prescribed for a wide variety of things other than that. And in fact, even today, antipsychotic medications are primarily used for poor behaviour, at least in the child and adolescent population, poor behaviour and depressive symptoms. And so, prescribing for those milder conditions, or maybe more readily thought of not serious mental illness, raises concerns about whether the risk versus the benefit for using those medications balances out. And part of the concern is that the medications are associated with rapid weight gain and more - and serious weight gain, so one standard deviation compared to the growth chart, which is a significant amount of weight, and also we have sleepiness. And so, the things that teens care a lot about, you know, their, you know, body and self-image and… Jo Carlowe Hmmm, absolutely. Dr Robert Penfold …you know, being alert and able to interact with friends in sports and school are affected by that. And then, even more seriously, those same things can lead to incident diabetes. And so there really are serious side effects, and so people were thinking to themselves, wow, is the risk-benefit really balancing out here? Jo Carlowe Has it improved since that, sort of, onset of when it was thought to be the miracle cure? Dr Robert Penfold Hmmm, good question. So, it has, in fact - at least in the US, the US data that I’ve looked at, the prevalence of prescribing antipsychotics to young people who do not have a psychotic disorder or bipolar with mania has levelled out. So, we seem to have seen a levelling out, at least in the way of the prescriber. Jo Carlowe Hmmm hmm, that’s good to hear. Dr Robert Penfold Hmmm. Jo Carlowe Let’s turn now to your JCPP paper, this is, “Safer and Targeted Use of Antipsychotics in Youth an Embedded, Pragmatic Randomised Trial.” Can you give us a brief overview of the paper? What did you look at, and why? Dr Robert Penfold Maybe I could decrypt the title a little bit, “Embedded, Pragmatic Randomised Trial.” So, as Researchers who work at places like Kaiser Permanente or HealthPartners, or Henry Ford in the United States, it’s a lot more like the Canadian or British systems, where everybody who’s enrolled in that system is covered by the same medical record, has the same kind of access to care, and has a set of harmonised policies that improve access to care. And so, the other component of that embeddedness is that we work with the Clinicians who are actually delivering care and the people who are making insurance policy and, you know, the people who are scheduling care, to design programmes that fit into the existing infrastructure, in a way that’s often not done in university-based settings. You know, in university-based settings, you often have somebody who comes up with an idea, they recruit some undergrads and some graduate students, they test the idea works, and then they’re shocked to find out that it doesn’t work when they try to do it in the real world. Jo Carlowe Right. Dr Robert Penfold And then we have a whole field of implementation science to fix that problem. Our group tries to start with the end user and say, “Okay, what are the things that would hold you back from doing this thing, and can we design something that will work?” So, that’s the embeddedness. The pragmatic part is that we had very few exclusion criteria for who we would allow to participate in the study. It’s often the case that there are very strict conditions for who can participate, in order to, sort of, weed out the things that would interfere with the programme or medication or whatever else, working. So, we had very few exclusion criteria. The randomised part is interesting, and I know we’re going to talk a little bit more about the design later, but… Jo Carlowe Hmmm hmm. Dr Robert Penfold …we invite over four health systems that participated in the study, so that was Kaiser Permanente in the state of Washington, Kaiser Permanente in Oregon and Kaiser Permanente in Colorado, and a hospital system in Ohio called the Nationwide Children’s Hospital. Each of them is a, you know, an integrated delivery system that both provides care and has insurance products. We randomised more than 7,000 Doctors, Advanced Nurse Practitioners and other prescribers in those health systems before we started the study at all. And so, it’s a cluster randomised trial insofar as when a Physician or other Clinician prescribed an antipsychotic medication, that triggered enrolment of the young person in the trial, as well. That’s a, kind of – an innovative design. We thought that we would not be able to stop the initial decision to prescribe the medication, but that we might be able to shorten the amount of time that the person took the antipsychotic medication if they had better access to the first-line treatments, psychotherapy, for example, or other lower risk medications. So, our two complementary outcomes were the proportion of young people who were still taking an antipsychotic medication six months after their initial visit, and then the total number of days that they used medication over that first six months. And we thought, if we give these people access to psychotherapy and we do some consultation with a professional Child and Adolescent Psychiatrist, we might be able to shorten the time that they took antipsychotics. Jo Carlowe Thank you. So, my understanding is, the goal of the Safer Use of Antipsychotics in Youth Pragmatic Trial was to test a more comprehensive approach to the reduction of antipsychotic medications to young people. The approach has four components to it, I understand, so, can you describe each of these? Dr Robert Penfold So, we designed clinical decision support embedded within the medical record, as the, sort of, first entrée into the study. So, in the Epic electronic health record, there’s something called a “best practice alert.” And when the Clinician opened a dialogue box to order antipsychotic medication for a person in the clinical encounter, it triggered this best practice alert, which is basically a dialogue box that said, “Are you sure you want to do this?” And there were four options then presented in their – in that best practice alert for the Clinician. One was, “Do you want to phone a friend?” And so, we had a Child and Adolescent Psychiatrist waiting to give them consultation on what they might want to do. We said, “Do you want to expedite access to psychotherapy for this young person?” In other words, we had a study – a group of study Psychologists ready to provide psychother – appropriate psychotherapy for the young person. We also had, “Do you want to refer this person to a Mental Health Navigator?” The Navigators were licensed Social Workers who would help the family get connected with care. So, I’m less familiar with the situation in the UK, but in the United States there aren’t enough mental health trained Clinicians to do the work, and so it’s very difficult often for, you know, moms and dads to find appropriate care in a timely way for their kids. So, we had these Social Workers in the role of, “Okay, let me find you a place to get care,” and help the family navigate that. So, the four – just to recap then, the four components were decision support in the record, expert advice from a Child and Adolescent Psychiatrist, navigation to help people get connected with care, and expedited access to psychotherapy. Jo Carlowe Hmmm hmm. Dr Robert Penfold It seems very quaint now, but, you know, back when we were designing this study, we said, “Well, what if we make these available, these – this expedited access available via video visits?” This was pretty novel ten years ago, and, you know, of course, after the pandemic became quite routine. Jo Carlowe Absolutely. Dr Robert Penfold Kaiser Permanente delivers more than 80% of its mental health now via televisit – you know, video visits. So, those are the four components. And let me just say, two of those things, the expert consultation and the expedited access to psychotherapy were designed after consulting a national group of experts on, “What are the barriers to first-line or preferred treatments that may result in an antipsychotic being prescribed?” And those were, "I can’t get access to a Psychiatrist," and, "I can’t get my patient fast enough or, you know, easy access to psychotherapy." So, we were trying to make it easy to do the right thing. Jo Carlowe I want to turn to the findings very shortly, but is there anything more you want to tell us about the methodology before we move to… Dr Robert Penfold Hmmm hmm. Jo Carlowe …the findings? Dr Robert Penfold So, we used something called a “modified Zelen design,” which is actually a very old study design, but not often used, and I have to say, you know, in initial reviews of the paper, people were quite critical of it. And the rationale behind the Zelen design is that if you only recruit people to your study that are willing to change their behaviour, you will get exactly the wrong answer in your design. And so we pre-randomised the Clinicians to intervention versus control. By the way, the control group got some of the decision support but they didn’t obviously get any of the expert advice or psychotherapy - and offered these things, and offered them to the patients of course, as well. But we retained all of the people who turned us down in the study. Jo Carlowe Oh. Dr Robert Penfold And we received approval to do that from our Research Ethics Board, or Institutional Review Board, whichever way you want to call it, to do that, because the people who turn you down are the people who you are trying to help or change their behaviour. Jo Carlowe Right, hmmm hmm. Dr Robert Penfold And if they turn you down right from the beginning, you failed. So, we retained all those people, as I say, so that’s a, kind of, an interesting thing. But interesting from an ethical perspective too that we retained people, especially in the control arm, for whom we had no contact at all, patients we had no conta – in the control arm. We retained them and used their information from standard medical records and administrative claims to measure their outcomes without their consent. But we justified that in the sense of, you can’t answer this question without retaining those people. Jo Carlowe Yeah, hmmm. Dr Robert Penfold And so, that’s the, sort of, ethical justification. So, I think that’s a, sort of, an interesting thing, and it’s a design that we’ve employed many times, because we’re trying to change people who don’t necessarily want to change or consent to change. Jo Carlowe Very interesting. Let’s turn to the findings. What key findings from the paper would you like to highlight? Dr Robert Penfold Kudos to JCPP for taking the paper, because it was a negative study. We did not change the total exposure to antipsychotics or the percentage of youth in the arm, but there was a promising result. So, the, you know, the odds ratio for being still on an antipsychotic at six months compared to the control group was .75, but it was not statistically significant, and there was a reason for that. We – when we designed the study, we powered it thinking that 80 or 90% of young people who were prescribed an antipsychotic would still be taking that medication at six months, and so, we designed our recruitment around that comparison, right? Jo Carlowe Hmmm hmm. Dr Robert Penfold We thought, we can probably cut in half the people who are in the intervention arm, which we did by the way, it turned out that only half of the people in the control arm, with whom we had no contact, were still taking their antipsychotic at six months, which was not expected. Jo Carlowe Hmmm. Dr Robert Penfold And so, even though it went, you know, sort of, in the right direction, that assumption about our power turned out not to be true, so we were slightly underpowered for that outcome. However, we did conduct what are called “heterogeneity of treatment effect analyses,” and so we looked at whether groups, in particular boys versus girls, versus those for whom we did not have recorded sex, for people of colour versus white, non-Hispanic folks, and other groups, whether in those subgroups of people there was an effect or not. And it turned out that for young people of colour in the intervention arm, randomised to the intervention arm, and young people of colour randomised to the control arm, there was a significant difference, and in fact, we reduced by about a month the amount of exposure or use of antipsychotic medications. So, that was a, kind of, interesting result. Secondary outcome, we did, in fact, increase the use of appropriate psychotherapy in the intervention arm compared to the control group, not by a lot, but by a little, so that was a positive outcome. Jo Carlowe Can I go into more detail about some of those findings? So, you… Dr Robert Penfold Hmmm hmm. Jo Carlowe …talked about the fact that the intervention may be effective at reducing unnecessary antipsychotic prescribing for youth of colour. What do you make of that finding? How do you explain it? Dr Robert Penfold So, we need to be careful. The result is true. We randomised youth of colour in both arms and compared them via that mechanism. So, we, at least theoretically, reduced the likelihood that these are related to other reasons associated with those young people, right? Socioeconomic status, comorbid conditions, all of these kinds of things, ought to be, at least theoretically, balanced vis-à-vis randomisation. One of my mentors famously said to me years ago, “Uh, randomisation fails again.” So, it’s possible that there is a residuals confounding for these other factors that we didn’t measure or that we didn’t – that we don’t know about, so that’s a possibility. Another one, which is, you know, sort of, less desirable, is that the youth of colour in the study started off in a worse place before they got into the study, and so there was more room for movement amongst those young people of colour. And so, that’s something that we’re trying to investigate further. Jo Carlowe Looking at some of the other findings, so your results show what in the paper you call “light touches”… Dr Robert Penfold Hmmm hmm. Jo Carlowe …were not effective in changing antipsychotic prescribing behaviour. And in fact, in the paper, you state, this is a quote, “It appears that a decision was made by the time the Clinician opened the medication order dialogue box, and it was too late to affect the course of action.” Can you elaborate on that? What are its implications? Dr Robert Penfold You know, in the United States - to set the context, part of the rationale for this study was the positive finding from a study that we conducted earlier about the State of Washington and similar findings that happened in the State of Massachusetts, where state Medicaid agencies – and so, for listeners who aren’t familiar, states provide insurance to predominantly low-income and disabled folks through a programme called Medicaid. And we had Medicaid enrolled youth in the study, and we – and these are – this was a high risk population of young people, especially. But those programmes, and especially Washington State, introduced something called “prior authorisation,” and so when a Clinician in that scenario prescribes an antipsychotic medication, the initial prescription is allowed, but in order to continue more than three days, that prescribing Clinician is required to have a phone conversation with a Child and Adolescent Psychiatrist employed by the state to argue the clinical rationale for why they’re doing this. Jo Carlowe Hmmm hmm. Dr Robert Penfold So, they basically have to get permission before they – so, it’s a lot – a harder stop on that behaviour compared to what we were trying to do. I think one of the benefits of designing these light touches for us was talking to the prescribers at the four – in the four health systems about, “Hey, like, what could we do that would be helpful and be working with you, rather than at odds with you, in achieving this goal of reducing antipsychotic prescribing?” Jo Carlowe Hmmm. Dr Robert Penfold And so that light touch was meant to work with them and, you know, nudge their behaviour, rather than, you know, slapping their wrist. But it didn’t work. And so, we were left wondering, well, why not? And I think when we think about the workflow in clinic and how these Clinicians are working with families to help them, they’ve listened to the clinical scenario. They know in much more detail than the reviewing Psychiatrist would the scenario, the clinical scenario, the background of the family, what they’ve tried before and what we’re just trying to sniff from the medical records, they know the family, right? Jo Carlowe Hmmm hmm. Dr Robert Penfold And so, I think they’ve done the work in their mind that Doctors do, that prescribers do, and arrived at this decision. And if then when they go to make that order and they’ve made the decision on what to do, then to come and say, “Hey, we’re second guessing you,” I think that was not well received. The other piece, that’s a little bit humorous actually, is that two of our reviewing Psychiatrists in these health systems were quite junior and told hilarious stories of trying to have these conversations with their very senior Psychiatrist colleagues, “Hey, are you sure you want to do this? Like, I’m not really sure that you want to do this,” and then being, sort of, batted away, if I can use that – so, I think was a second reason. So, what should we do? I think we’re left with moving upstream in this process and I think that part of the reason that antipsychotic prescribing has levelled off in the last, you know, six or eight years is that professional organisations and medical schools have realised that this is probably not the best thing to do, and so people are receiving different training and looking for ways to do things better. Jo Carlowe Were there any other positives, or other changes in prescribing behaviour, that did result from the intervention, that you haven’t already mentioned? Dr Robert Penfold So, the short answer is no. A medical student who’s with the Kaiser Permanente School of Medicine in California, did an analysis of all the other medications that were being used by these young people, antidepressants, anti-anxiety medications, stimulants for attention, and there was no change, i.e., increase, in the use of those medications in the intervention group relative to the control group. There was some little glimmer there. It appeared that the girls in the intervention arm did have more use of antidepressant medications compared to girls in the control arm. So, there’s a little glimmer right there. I should say, there’s actually an interesting pattern with respect to boys and girls who were prescribed antipsychotics. Boys were overwhelming diagnosed with externalising conditions, oppositional defiance disorder, conduct disorder, and given antidepressants, and there’s this, sort of, lore that this is a boy problem with respect to antipsychotic medications. Interestingly, in our study, we enrolled, randomised, about 50/50 boys and girls, and so that was a bit of a surprise. But maybe even more interesting, most of the girls had internalising disorders, like depression and anxiety. Jo Carlowe Hmmm. Dr Robert Penfold And that’s why they were getting prescribed antipsychotic medications. So, that’s, kind of, an interesting finding. Jo Carlowe Rob, is there anything else in the paper that you would like to highlight? Dr Robert Penfold Maybe not directly in the paper, but I think I want to emphasise that doing this, kind of, embedded research in health systems, where you’re working with the Clinicians who are providing care and the service line leaders who are directing that care and the insurance providers, you know, like, NHS and NICE guidelines and that kind of thing, when you actually work within the system to conduct the randomised trials and you enrol typical patients, i.e., you take this pragmatic approach, I think we’re much more likely to generate evidence that Clinicians are – will believe, and that will actually work in the general population. When we turn these programmes or new medications or new services, new psychotherapy, when we turn them loose, they’re more likely to work. So, I’m an advocate for that, kind of, approach. I think that in mental health and in psychiatry, for a long time we’ve been producing interventions and new kinds of approaches to psychotherapy that simply don’t work, and we wonder why people don’t participate more in psychotherapy. At least in our health system, the modal number of visits for a person initiating psychotherapy is one, and the next most frequent is two, and then they drop out. And so you’re left wondering, why don’t people want to buy what it is that we’re trying to give them? And I think it’s because we’re designing interventions that have not been developed embedded in research systems, and so people don’t want ‘em. Jo Carlowe Hmmm, and when you say “people,” you are specifically referring to “young people”? Dr Robert Penfold Yeah, young people, young people, and I think adults too. I think we’ve designed approaches to psychotherapy that are then – you know, that are usually tested in adults and then adapted for young people. Jo Carlowe Hmmm hmm. Dr Robert Penfold And they’re based on the wrong kind of evidence. Jo Carlowe What recommendations emerge from your research? Dr Robert Penfold You know, it’s an associated paper that we talk about in this main outcomes paper, but there are many guidelines around the world about how to prescribe, and how to prescribe for various kinds of mental health disorders. And there’s been a lot of work done about, "Why don’t Clinicians follow these guidelines?" And I think part of the reason is that those guidelines start with a condition and then make treatment recommendations. And we, kind of, turned that model on its head, actually, and we said, “Here are the symptoms and here’s the clinical scenario that you’re presented with. What kinds of treatments fit what you’re trying to do here?” And so, the – we have this companion paper about a symptoms-focused guide to prescribing guidelines, and we say, “Okay, under these scenarios, here’s what we think you should do, and here’s the first-line, and if that doesn’t work, here’s the second-line. And if that doesn’t work, well, maybe let’s try an antipsychotic medication, after you’ve tried all of those things.” And so, in that approach, it gives you some space to think about comorbid conditions, it gives you some space to think about the history, and is more flexible for the Clinicians. And so, we’re, sort of, hoping that that approach – and we’re doing some more work along this line, but we’re hoping that approach has some impact on changing prescribing behaviour. It’s a notoriously hard thing to do. Once a person gets trained and gets some experience, to try and change their prescribing practice later is very, very difficult, and so, we’re continuing research along those lines. Jo Carlowe Finally, Rob, what is your take-home message for our listeners? Dr Robert Penfold So, you know, we got a little bit unlucky with this trial. We launched it at a time when prescribing was levelling – prescribing antipsychotic medications, was levelling off. So, that’s a good – that’s good news, right? We’re getting better at targeting antipsychotic medications to the people who really need them and reducing the possibility of negative side effects. And I think that the professional organisations around the world and the medical schools and increased awareness from the media have created that, you know, momentum, and has resulted in, you know, alternative treatments, alternative medications, so that the, you know, young people can thrive. It’s a – I think it’s a good message overall. Jo Carlowe Rob, thank you ever so much. For more details on Dr Robert Penfold, please visit the ACAMH website, www.acamh.org, and don’t forget to follow us on your preferred streaming platform, let us know if you enjoy the podcast, with a rating or review, and do share with friends and colleagues.