The Potential of Digital Therapeutics and Management Strategies for Major Depressive Disorder

Digital therapeutics (DTx) may help address unmet needs in the management of major depressive disorder (MDD). Before considering DTx for patients with depression, HCPs may find it helpful to first understand what they are and what they involve. In this webinar, Dr Craig Chepke will review considerations in the use of DTx including nuances in the regulatory pathway, clinical trial design, interpretation of the data, and some current evidence for DTx in the management of MDD.

Dr Tarolyn Carlton: Hello and welcome to the program. My name is Tarolyn Carlton, and I am the Senior Director for Global Medical Strategy, Digital Medicine & Digital Therapeutics at Otsuka Pharmaceutical. Thank you so much for taking time out of your busy schedules today to join us for the webcast entitled The Potential of Digital Therapeutics and Management Strategies for Major Depressive Disorder. So as I mentioned, I'm from Otsuka. This presentation is brought to you by Otsuka Pharmaceutical Development and Commercialization Incorporated, and the presenter is a paid consultant of Otsuka. 

So let me introduce you to our presenter for today's webinar. Dr Craig Chepke is a board-certified psychiatrist and a distinguished fellow of the American Psychiatric Association. He attended NYU School of Medicine and completed his residency training at Duke University. Dr Chepke is medical director of Excel Psychiatric Associates in Huntersville, North Carolina, as well as adjunct associate professor of Psychiatry for Atrium Health. Dr Chepke’s research and clinical interests are in serious mental illness, movement disorders, ADHD, and sleep medicine. As part of an interdisciplinary treatment team, in his practice he employs a person-centered care model to tailor treatments to each individual's needs, integrating traditional pharmacotherapy with psychotherapeutic and physical health and wellness interventions.

So, let's look at the objectives for today. So really the first objective is to look at considerations in the use of digital therapeutics—and you'll see the acronym DTx used across the presentation, so that means digital therapeutics—including the nuances in the regulatory pathway, the clinical trial design, interpretation of the data, and some current evidence for DTx and the treatment of depression. Next, we're also going to describe Rejoyn, which is a recent FDA-cleared prescription digital therapeutic for the treatment of major depressive disorder symptoms as an adjunct to clinician-managed outpatient care for patients who are 22 years and older and who are on an antidepressant medication. And then lastly, we're going to discuss the pivotal Mirai trial that's supporting the recent FDA clearance of Rejoyn. So I'm going to turn it over to Dr Chepke.

Dr Craig Chepke: Thank you very much for that introduction, Tarolyn, and thank you everyone for logging in. I think we've got a really exciting topic today. Digital therapeutics are a really hot, up and coming area in our field in psychiatry, and I think there's a lot of promise and I'm looking forward to being able to discuss this with you. And we have a lot of material, I'm going to try to get through it all as quickly as I can, because I do want us to have a good robust Q&A session. So like Tarolyn said, please submit your questions anytime you think of them throughout the entire presentation. Don't wait till the end; put them in, and that way they'll already be locked and loaded for us by the time we get to the Q&A. 

So let's start off with a few of the considerations in the use of digital therapeutics in the treatment of depression. We know that MDD has just voluminous unmet needs, and there's no one area that is going to be filled with that unmet need that is going to take care of it all. So pharmaceuticals, psychotherapy, digital therapeutics, there's so much unmet need in MDD, we need lots of new interventions, and DTx is hopefully going to be one that is going to be used increasingly more and more as time goes on. 

So what are we talking about with DTx specifically? Let's drill down a little bit. So DTx delivers software-based interventions. Alright, makes sense so far. But here's a key point. The software intervention has to treat or alleviate a disease, disorder, condition, or injury. So it actually has to be actively doing something with the person acting as the agent, so to speak. And another important aspect is that if it is a digital therapeutic, it has to be supported by clinical data and should be guided also by real-world evidence by the Digital Therapeutics Alliance definitions.

And there can be both prescription DTx and nonprescription. So you can't necessarily tell the two apart. It could be, DTx could be something that does require or does not require a prescription, but very importantly, it is subject to the FDA's oversight. So it is regulated by the FDA. 

Dr Craig Chepke: And a couple more highlights here. Some digital therapeutics can be used by themselves, and then some are adjunct to, or alongside of at least, medications or other devices or therapies. So it could be one of a multitude of different things. That's what makes, it can be a little bit difficult for some clinicians to figure out, well is this or isn't it a digital therapeutic, because it’s with this and that or it's by itself. We'll go through some examples in just a little bit.

And the thing that is really exciting about it is it may increase access. In many parts of the country, we have very few mental health clinicians no matter what their discipline, whether it's as a prescriber, as a psychotherapist, case manager, whatever it might be. And we need to desperately increase the amount of services we're providing, because the demand has skyrocketed in not just years but decades, and the pace of the supply of clinicians has not met the charge there. So we need something to help fill that gap, because people are going untreated every day, or undertreated, which can be just as bad. 

So a couple examples of digital therapeutics. They're down at the bottom. There are digital therapeutics out there for a lot of things that we're treating on a regular basis. Substance use disorders, sleep disturbance, one in particular for a nightmare disorder. And then there's even a video game-based treatment for ADHD. So I said it could be difficult to determine is it a digital therapeutic or not. So I'm going to bring Tarolyn back up on screen and we're going to play a little game show. So is it or is it not a digital therapeutic? Welcome back, Tarolyn.

Dr Tarolyn Carlton: Hi. Okay, so yeah, let's do, is it a digital therapeutic or not? So the first one: psychotherapy app, is it a digital therapeutic?

Dr Craig Chepke: Great question. So I will go back to what I said a minute ago that it has to be actively doing something. So I'm going to vote yes on this one because a psychotherapy app, it's actually going through and doing something, it's providing, yep, and there's the answer. If the app has been, great point here, an app would have to be authorized by a regulatory body. So if not, then it would not be. But assuming that it is, then it's the app doing the therapy. It is acting as the agent delivering the intervention, not a clinician. The clinician may prescribe it or recommend it if it's an over-the-counter one, a nonprescription that is, but it is the app doing it. So it's a digital therapeutic.

Dr Tarolyn Carlton: All right, one for one. Let's see, what about an AI virtual therapist that responds to the patient's mental health cues? Is it or isn't it?

Dr Craig Chepke: That one's going to be a yes too, because that's something similar. Don't let the AI buzzword throw you off. It is acting as the agent. It is responding and it is interacting with the patient, not us the clinician doing it. So that's going to be a digital therapeutic, but only if it's authorized by a regulatory body. Because that's who they're engaging with, the AI, not with the clinician live on the other end.

Dr Tarolyn Carlton: Okay. Alright, you’re two for two. Let's see. I'm going to throw in a hard one for you because a lot of people use sleep and activity trackers. Would that be considered a digital therapeutic?

Dr Craig Chepke: Great question. These are very widespread and have been for many years. I'm sure probably a lot of people on the call, and I know I have had this happen, that patients will come in with these, I won't use any of the names of them, but will come in with this little thing around the wrist and say, “Okay, I want to show you on my phone what it says about my sleep, what does this mean?” That happens fairly frequently. And so that is not a digital therapeutic, though, because it's just passively collecting information or the patient is entering it in there, but it is not reacting back out. Sure, it can display some data, but it's not providing an intervention to the person. So that's a no. Yep, that's the answer. And also, to reiterate, it doesn't treat or alleviate the disease, disorder, condition, or injury. So it tracks it, but it doesn't do anything to make it better. We can act on that data as clinicians and say, “Hey, I think you should do this with your sleep or that, and hopefully we'll see if your data gets better.” But that's me the clinician doing it, not the trackers. That's not a digital therapeutic.

Dr Tarolyn Carlton: Alright, let's see. Can you go four for four? Telehealth platform? Because a lot of times when we say the word digital, people think telehealth. So is telehealth considered a digital therapeutic?

Dr Craig Chepke: That one to me is an easy one. That's a no. And I can tell you from experience, my telehealth platform is not therapeutic in treating anything in my life, and that is exactly it. It's the clinician, not the software, delivering the intervention. It's just a conduit. It's like the highway. It's just there to carry our interventions across the internet.

Dr Tarolyn Carlton: Alright, well you're the big-time winner of Is it a digital therapeutic?

Dr Craig Chepke: Alright, well hopefully that was entertaining and elucidating to our audience out there that it can be complex. Some of these may not be immediately obvious or at least you might understand them in hindsight, but just if everyone was playing that game at home along with me, you might not have gotten all of these right. So that's something that we have to be mindful of. And let's take a look at how the regulation is of digital therapeutics. 

So these are categorized as actual medical devices by the FDA. That's what makes them subject to their purview. So there's 3 different classes, and each one depends on the amount of risk, more or less. So low risk, moderate risk, high risk. And then each one has a different pathway. Most likely pathway for a class 1, which is low risk, is something called a 510k, or it could be just exempt altogether.

Examples of things like that are manual toothbrush—also electric toothbrushes, by the way. Fun fact, both manual and electric toothbrushes, or handheld surgical instruments. Pretty simple straightforward things that are manipulated by the user directly and don't have a whole lot that can go wrong, really. Class 2, moderate risk, that is definitely a 510k, and it's something that is established and therefore substantially equivalent to a legally marketed predicate device. So if it's something that has already come out in the past and if there is something that is a prior art, and it substantially acts in the same way, then it can be authorized through that pathway. And the NightWare and EndeavorRx are some examples there. NightWare, that's the one that I mentioned that interrupts the nightmare disorder; EndeavorRx is for ADHD, the video game treatment. And then the class 3 high-risk ones, that's a pre-marketing authorization. So that you have to go through quite a bit of work with FDA in terms of that one to get that one authorized. And there has to be an extensive scientific review to make sure it's safe and efficacious, because we don't want things like pacemakers and deep brain stimulators going wrong. Anytime there's any surgery involved, it's going to be pretty highly regulated. 

Dr Craig Chepke: Now there were some exceptions to this during COVID-19, and I think that was a great idea and rightfully so, because there was unbelievable lack of access to healthcare of broad spectrum in the US. So things were exempted that the FDA did allow digital psychiatric health devices to be marketed prior to obtaining regulatory clearance. But that went away on November 7th.

Now I talked about prescription/nonprescription. So there's a couple of different ways that this can go about. So no authorization needed, that's what would be over the counter. So they can just go to any, the app store on their phone, whichever one they have, direct websites sometimes, without anything. They can find it themselves, access it themselves, and we may not necessarily be involved. However, just like over-the-counter medications currently, we may recommend that someone take melatonin. We don't write a script, we just say go and buy it. That's what an over-the-counter digital therapeutic would be like as well. Or there could be one that does require authorization by referral, so that it's licensed to the healthcare provider or somebody else, and then a referral is required for access. So they have to go through us as a healthcare provider, but it's not a traditional prescription.

And then there is the traditional prescription model that we have to prescribe it just like we would prescribe a pharmacotherapy. And this is going to be different depending on what country that you're in, and there's a lot of different variability there. So the bottom line, down at the bottom of the slide, it may or may not be covered by various insurances, state-run healthcare programs, but there are some DTx that are provided as part of employee or health plan benefits. So we may have coverage for some of these digital therapeutics. We won't know until we try.

So with any new type of intervention, I think that the clinical trial design is really difficult, because if you're designing a clinical trial for a medication, well we've been doing that for many decades now. And, not that it's simple, but at least the basic format is, you kind of understand, okay, you have the pill with the active medication, you have the patient put it in their mouth, swallow it every day or multiple times a day as it may be. Or they take a placebo, which is something inert. We used to call it a sugar pill. It's not actually just sugar, but it's all inert, inactive ingredients. And then you randomize them, and there's some statistics go along with it, and there's lots of nuances in between. But fundamentally, that's the basic setup for all pharmacotherapies that are orally administered. 

Dr Craig Chepke: But digital therapeutics are a little bit more difficult than that. And since this is a nascent category, they haven't worked out all the kinks yet, because It is really challenging, specifically in the control conditions, because how do you design a placebo for a digital therapeutic? So let's take a look at some of the reasons why this is so difficult, because that will really help inform as to interpreting results. I think of it kind of like the Olympics when they do gymnastics, they don't just get 1 score based off of 1 single metric. They take degree of difficulty into account, and if you attempt something more difficult, you get more credit. I think that makes a lot of sense. And so if we really understand more about what it takes to design a clinical trial for digital therapeutics, I think we'll have a greater appreciation of the results. 

The thing is that just like in pharmacotherapies, the exact mechanisms that are producing therapeutic effects in psychological interventions also aren't fully understood. We can't know for sure about that, which makes that difficult from the very get go. And then the digital therapeutic may or may not have disease management features, and that could be behavior or goal tracking, that interact with the disease-specific treatment to produce a really complex multifaceted intervention. It might not just be one straight simple thing. There could be lots of different facets to it, various interventions, and we'll have some examples of that later. 

And then there might be different types of placebo mechanisms that people might have, certain beliefs that the technology is good. Some people may have the opposite. It could be more like a no-cebo effect, that the technology could be harmful, or it could be neutral. And that's very different, just like it is different for medications, but it could be different ways in different people, and we don't really understand all those impacts yet.

And then just the digital mode of delivery itself, we don't know what the efficacy of that is. I mean if we are interacting with something on a regular basis or we're getting prompted to with text messages for engagement, well how's that going to affect someone? I mean someone, if you think about major depressive disorders specifically, often these people are more isolated and maybe no one is calling to check on them. So if they're getting a text message reminding them to engage with their app, that might make them feel like someone cares about them that they wouldn't get in a true placebo, which would be that they didn't have any intervention at all. 

Dr Craig Chepke: So that brings up, what are the different levels of what the control conditions could be? So the wait list, that is probably the closest thing to what we would think about in medicine as, in pharmacotherapy, as a placebo—doesn't do anything. Well, if someone's on a wait list, they're not getting any sort of treatment whatsoever, and that means they're not getting any contact with anyone, which is actually more restrictive than clinical trials with placebos because they do have the interactions with the clinical trial site. But it's very reflective of what we have in the real world though, because we know that psychiatric providers are in short supply and so people wait and wait and wait. So that could be one, but that has some problems too. 

Treatment as usual, that means that they just get traditional psychiatric care, and that can vary based off of who they're seeing and regionally, things of that nature. 

Active control, well that has a known effect. So this we see in pharmaceutical trials as well that the investigational med will be compared to placebo, but there may be an active control in there as well, an established medication that has been approved and is believed to be reliably effective in that specific condition. An antidepressant that's known to work, an antipsychotic that's known to work, whatever the disease state is. So there can be some comparisons and we can test, was this a negative trial or a failed trial? Because if the established treatment doesn't, as the active control doesn't separate from placebo either, well that's a failed trial instead of just a negative one. 

Then finally, the sham or placebo control. So this is the one that's actually the closest to a pharmaceutical one, but the sham is harder to create than a placebo tablet or capsule, because what about the time and attention? I mentioned an example of that, that if you're interacting with something, you're giving attention to not just the device but your symptoms and thinking about what's going on and thinking about managing your illness, well that could have a therapeutic effect. We don't really know what the full ramifications of that are. Can you match for that? I don't know if you can or not. That's difficult. We don't have that kind of information yet. And patient blinding as well. Are you going to give them, is the sham arm Candy Crush? Well, they're going to know that it's not a digital therapeutic if it's a game that they've already played or looks a lot like it. So how can you make sure that they're blinded? If they know they're not getting the actual intervention, that's going to depress their effect in the placebo arm, or if people know they're getting it, well that could inflate theirs inappropriately. So there's no universal definition of a digital sham. There's lots of different proposals, but there's no real consensus. Then there's also even inconsistencies with the nomenclature. No surprise, this is psychiatry, we can't name anything right or agree on it even. So it's really complicated and it makes the interpretation of the results a little bit difficult. 

Dr Craig Chepke: So a few more questions to ask. We want it to mimic the actual digital therapeutic as closely as possible, only different in the active component, which is what we try to do in pharmaceutical studies. So is the sham going to identify the active components? That would be kind of like reverse engineering it. If you have the digital therapeutic and you don't, it seems to work, but you don't know why, how do you isolate the variables and figure out which one it is? Is that what the sham is going to help? If you do everything but one aspect of it? Well, and then if it doesn't work, okay, it's not that, try a different one. Okay, that one now works as the isolated one. That must be what's efficacious. Or is it to demonstrate the efficacy over a matched sham, and is a sham even validated? Has it been used in other studies in the past? Because if not, who knows if it's a good sham or not. Do we know what the active components are? Are they different from what's in the sham? They could be relatively similar and it might not be so obvious if they are. 

And digital therapeutics often include adaptive levels that as someone gets better at it, it increases the degree of difficulty that I brought up before. So that degree of difficulty, well, does the sham have that too? Because people realize that, oh, I'm getting harder tasks, I must be getting better. Maybe that leads to some improvement because it inspires some hope that, hey, well I actually made progress. Does the sham have that too? Because if it doesn't, is it just as fair? I don't know. 

And then are they going to be blinded, and is it designed to match the engagement and user satisfaction? Because if people don't have any satisfaction, if it's, well, a video game is not a good analogy for that sham that I brought up before, but if there was a really boring one that people don't get anything out of as opposed to something that they feel they're actively engaged with, it could change some of the control situation. So wouldn't it be great to have a 3-arm study? So you have a wait list or treatment as usual, and you have a sham, and then you have a digital therapeutic? Well, maybe that would be nice to have more data, but then that makes it a lot more complex to the design and the analysis. 

So I think you can see by now that this is not something easy, and we're going to have to learn along with the investigators who are doing these studies. 

Dr Craig Chepke: And then how do you interpret the data? And this is a problem with all clinical trials, because you could have something that is statistically significant and clinically meaningless, because it could be a very small difference. And just because it's a statistically significant finding doesn't necessarily mean it's an important, clinically meaningful finding for us or more importantly for our patients. 

So the FDA has thought of this and they've come up with some recommendations, and I think these are really exciting. It's something called MWPC, the meaningful within-patient change, and between-group difference. And usually we're used to thinking about the mean, what is the mean difference between the medication or the intervention group and the placebo group, and we compare that. Well, I don't know about you, but I don't treat 300 patients en masse at a time and give them all 1 intervention or half of them 1 and half of another. I treat 1 patient at a time. And that's what the within-patient change looks at is each individual in the study. Yeah, they're part of a group, but each individual has a change. Hopefully it's a change in the right direction, could be a change in the wrong direction, they could get worse while taking the treatment. And then they look at all of those and then they have certain defined thresholds of what's clinically meaningful and what percentage of patients hit that mark. And we'll take a look at some examples of that later. But I think that's a pretty cool way to look at it, because it brings each individual with lived experiences' journey into clear relief, because that's what we're doing is treating individual patients in our clinics.

And this is an elaboration of what I just mentioned about the statistical versus clinical significance. And it gives an example here. So over on the left of the 2 arrows that face in towards the center is, one way that you may get clinically insignificant results from a statistically significant difference is that if you have a very large sample size, you can prove almost anything if you get the sample size large enough. Even a tiny, tiny little difference that is insignificant clinically, if you have enough people in the trial, you can make it statistically significant. It'd be hard to design and costly, but it could be done. But on the other hand, you could have a statistically not significant result that could have actually a very clinically relevant importance. So there could be a large difference in studies with a small sample size. So the difference could be huge, but they didn't have enough people to show statistical significance. But that difference—man, that would mean a lot to an actual patient if they got that much difference in their illness. 

And so these results should be interpretable and help not just us but the patients as well to make treatment decisions. So we always have difficulties in figuring out what do these results mean to me as a clinician, and what do they mean to my patient who could be taking this intervention? So ultimately it boils down to how much change does that 1 individual or does each individual need to see to be able to say that it's working and it's worth taking? So that's an interesting question. I like that. 
 

Dr Craig Chepke: So how effective are digital therapeutics? Well, there's been a lot of digital therapeutics trials, 182 trials that were in clinicaltrials.gov between January 2010 and December 2019. And as you can see, a quarter of them fully are psychiatric. So the single largest group of, the single most frequent group of indications are those for psychiatric indications. I would say probably it’s because of the huge unmet need that we have. Not that there's not unmet need in other conditions, but in my perspective, I think psychiatry has more unmet need than cardiovascular, endocrine, et cetera. And a lot of them actually have been sponsored by academia, over three-quarters of them. 

And then looking at the actual efficacy, there have been several meta-analyses. We picked out just one here, because it was the largest: 176 randomized controlled trials. So overall the study findings were that there were significant although small effects on depression and anxiety compared with the controls. And then there were some sub-analyses. If they were based on CBT, then those apps had larger effect sizes than just the global mean score. And then the effects were larger even when the apps used chatbots, which I thought was very interesting. And this can give us some indication that the overall concept of digital or mobile delivery of digital therapeutics may provide benefits for people with moderate-to-severe MDD. So the concept is good. We need good implementation though as well.

Dr Tarolyn Carlton: Great. Thank you, Dr Chepke. That was a lot of information, very good thorough background on the digital therapeutic landscape. And I know this is new information for a lot of clinicians, so definitely much needed education and we'll probably continue to provide this type of education. Now, you mentioned about a couple times, right, about digital therapeutics being regulated by the FDA. So I thought it'd be great now to switch gears and discuss Rejoyn, which is the first and only prescription digital therapeutic for the treatment of depressive symptoms. So I'll let you take it over.

Dr Craig Chepke: Yeah, thanks Tarolyn. This is a really exciting development in digital therapeutics that we're about to talk about. So as Tarolyn mentioned, Rejoyn is FDA cleared, that's number one, but we're going to go into more detail obviously. So it is a smartphone app and it is prescription based, so it's a digital therapeutic for the adjunctive treatment of major depressive disorder symptoms in adult patients age 22 years and older who are already on antidepressant medications. And the intent of the Rejoyn is to reduce MDD symptoms. And so it's adjunct to clinician-managed outpatient care as well, is one important aspect of it. Rejoyn provides benefit to patients with MDD symptoms, adjunct to antidepressant medication, no related treatment-emergent adverse events during a clinical trial. 

So a couple important safety information tips here. Rejoyn isn't intended to be used as standalone therapy, so it's not a substitute for medication. So as the indication said, it's for adjunct to clinician-managed outpatient care for people who are already on an antidepressant. So not for monotherapy, they should continue their current treatment as directed. 

And also, it's not going to be monitoring the patient's symptoms, clinical status. If someone is starting to get worse, as I mentioned is possible previously, we're not going to get pinged, we're not going to get alerted or warned about it. So the same thing that they would need to do, the same thing they would do if we were prescribing a regular medication, call us or call another healthcare professional, call 911, go to an ER if they're getting worse, if they have suicidality, things of that nature. 

Dr Craig Chepke: How do we think Rejoyn works? Well, we have to say just like we do with pharmacotherapies, we don't exactly know, but we have some theories, and Rejoyn, we think it helps depressive symptoms by enhancing cognitive control over the emotions. And it's a combination of things we see over to the right that complement each other. There's one specific training exercise, the Emotional Faces Memory Task. So it's cognitive-emotional training. So as I mentioned, enhancing cognitive control over emotions. We want to train the person to help have better cognitive-emotional balance. And so that is by targeting the amygdala and the dorsolateral prefrontal cortex. We'll go into the neurobiology more in just a minute. That's portion number 1. 

Portion number 2, so a completely separate intervention that is part of Rejoyn is our CBT-based video lessons for patients to learn and then apply the key therapeutic skills. Would it be great if there were enough live human CBT therapists out there for every single person with major depressive disorder? Of course, but we know there's not. There's huge shortages. It's hard to get people in. And Rejoyn includes CBT-based video lessons for the patients. 

And then the third component of Rejoyn are personalized reminders and messages. So they do a couple of things. One, they reinforce the skills that we're taught in those lessons up above and they also encourage people to complete the program and stay on task, almost like a coach type of mentality. So it's those 3 things together that create Rejoyn. So it's a tripartite type of intervention, and it's believed that the mechanism of action may leverage neuroplasticity to modulate the hyperactivity in the amygdala and improve the top-down emotional control by the prefrontal cortex. So let's talk a little bit more about that. 

Dr Craig Chepke: So there's 2 parts of the brain that we think are giving that top-down cognitive control over the limbic system. So I like to think of things with the hand model of the brain. This is something that I didn't learn in medical school or residency. My oldest son's therapist actually taught this to me a number of years ago, but ever since I learned it, I use it with patients all the time.

So it starts with our arm, that's our spinal cord going up to it at the bottom of our hand, that is the brainstem, and then our thumb would be the limbic system. Now anatomically you would fold that in. And then what makes that person human? What makes us human compared to earlier things in evolution? It's our huge neocortex, and that's our prefrontal cortex right there that folds down over. So it would be kind of like this anatomically because we have 2 lobes of the brain. But the way that the therapist taught it was that if you flip your lid, if your prefrontal cortex isn't fully controlling and giving that top-down regulation, then your limbic system has more room to be able to take control. And then you maybe flip your lid totally, and then that's when your limbic system is totally in the driver's seat.

So back to the story. So the DLPFC is part of the prefrontal cortex that's involved in working memory. So that's a cognitive skill, very rational, right? The dACC is connected to the prefrontal cortex, and that helps in a regulatory role. It helps regulate the cognitive and the emotion processing. And then the amygdala, part of the limbic system—emotion processing. The amygdala helps us to, what we take in through the senses, it makes a judgment as to what is a perceived risk or threat in the environment, whether it's verbal or visual, whatever it is, auditory—is this a threat or not? And it's always pinging. The amygdala always wants to fight, and it has to be controlled, because if you always want to fight, that's not good. And so you have to have control over that. And then the dACC can help to be kind of like a judge to regulate the cognitive and the emotional processing.

So in major depressive disorder, there may be top-down control that's impaired. And so if you're not regulating things as much, well, like I said, that gives the limbic system more free rein to be able to go where it wants to go. If you have decreased DLPFC activity, which is seen in patients with MDD, that could cause some problems. The dACC, which isn't part of the hand model, but also there can be dysfunction there. And I like the way in the diagram it has these arrows that are going all over the place, because it's just kind of confused. It doesn't know how to judge between the cognitive processing and the emotional processing, and it makes more errors, and things slip through the cracks that maybe shouldn't have. And the amygdala, it's increased as well, so it's perceiving more risk and threats. But with depression it might be that when the person sees threats and risk, instead of fight, they freeze, and they withdraw and they retreat back into negative cognitions, things of that nature.

Dr Craig Chepke: So EFMT, the Emotional Faces Memory Task, it simultaneously activates the DLPFC by something called N-back working memory task. So N is a specific number one back. So you see a face now and I'm going to show you the next one. If it's one back, what emotion was on the previous face, if it's 2 back, I'm going to show you a face, then another face, a third face. And then after that third face, does this match the face that you saw 2 faces ago, 3 back, 4 back, et cetera. So you identify the emotions, and that may alter connectivity between the brain regions, and we think that it's something to do with the emotions that the model is showing. And they're very big, bold, bright emotions, and they're usually easily recognizable by people, and they're taught what they are ahead of time. And engaging both the rational part of the brain and the emotional part of the brain can retrain it is the theory, and that, we believe, may involve neuroplasticity. And it's modulating that hyperactive amygdala, so it's bringing it back into check and then improving the top-down control, bringing the lid back over it to be able to give that control over the amygdala to suppress the inappropriate fight-or-flight responses that the amygdala could be sending upwards.

So Rejoyn has those CBT-based lessons. So let me go through what those CBT-based lessons are like. So they're derived from a couple of core CBT principles. One, emotion regulation, and this is more or less mindfulness. If people are more mindful of their emotions and how they experience them, and then what they go on and do about those emotions, how they act on them, they may be able to regulate those emotions better. But I also want to activate them behaviorally, because if the person can register something that's a meaningful life activity, that hopefully will let them be stimulated to increase their positive interactions with their environment, go out and seek those situations that could be therapeutic for them. And then those negative cognitions that I mentioned, we want to restructure those cognitively to be able to modify those dysfunctional thoughts and beliefs into something that are healthier. 

So Rejoyn has those 3 components. As I mentioned, the EFMT exercises, it's up to 80 minutes a week. It could be substantially less than that depending on the individual and how they're progressing. So it's not a guaranteed 80 minutes a week, that's a ceiling, not a floor. So it could be a lot less than that. They do 3 of those EFMT exercises a week. Each one is 11 to 26 minutes. So over the 6-week course of the treatment, they have 18 of these EFMT exercises. 

They also have 3 CBT-based lessons a week, and those are up to 12 minutes in total. That means it's 3 to 4 minutes each, so not a huge time commitment there. And they also do 18 of those over 6 weeks. But week 7 through 10, they also have access to those to go back and review them and do them again. 

And finally, the personalized reminders and messages. So it gives those regularly over the whole 10 weeks, and less than a minute just because you're just getting a message reminding you, just as fast as you can read it, that's how long it is, and it reinforces those skills that they've been taught in the CBT lessons and encourages them to keep up with the program and complete it.

Dr Craig Chepke: Now, I want to talk about one specific study in particular. It's called the Mirai trial, and that was a trial that was done to evaluate the effectiveness of a specific digital therapeutic in adults diagnosed with major depressive disorder. Here's what the design is. So they screened people in over the course of 3 weeks. They started with 386, they all were diagnosed with MDD and receiving monotherapy of that antidepressant. Importantly here, they were told they were going to receive one of two digital therapeutics. So they weren't told that they were going to get either a digital therapeutic or a sham, effectively a placebo. Remember sham is the closest thing we have to a placebo in digital therapeutics trials. So everyone who entered the study thought, “Wow, I'm either going to get digital therapeutic A or digital therapeutic B, no placebo, no sham.” But there was a sham. The actual active treatment was Rejoyn plus that baseline antidepressant, or the sham plus the index antidepressant. And they were balanced 192 versus 194, and they had those 3 Rejoyn sessions or 3 sham sessions per week. Primary endpoint at the end of week 6 was change from baseline to week 6 in the MADRS total score or the Montgomery-Asberg Depression Rating Scale. And then the extension period at 4 weeks after they stopped the 3 per week of the digital therapeutics or sham treatment sessions, then they still were made accessible. And then also the patients in the Rejoyn arm could continue with the CBT-based lessons, because they wanted to see the durability of effect at weeks 8 or 10. And overall, throughout the entire trial, patients had weekly remote visits either by video or by phone. 

Here's what these 2 different arms look like. So Rejoyn plus antidepressant, the EFMT exercises, you can see there some of the emotions, the person very sad looking, 2 sad looking in a row, then someone's smiling, very happy, pretty easy to identify those.

Then the CBT-based lessons and the text message lessons. So each Rejoyn session consisted of one of each of the EFMT and the CBT-based lessons. The sham was a Shapes Memory Task. So they did have something, they weren't just staring at a black screen. They had to do something that it didn't have the faces, didn't have the emotion, but they were shapes with the same kind of N-back technique. There were some pilot studies that used this same sham technique here. It was something that was used in previous trials. And so that was a working memory task but didn't engage the emotion, they tried to match to the EFMT for time, attention, the expectation of the therapeutic effect, again, being told that they were testing 1 of 2 digital therapeutics. So there was a lot in this that people were engaging with, and they did get the text messages, but they did not get the CBT-based lessons.

So those were the differences between the 2 arms. 

Dr Craig Chepke: Inclusion/exclusion criteria, I'll hit the highlights. Twenty-two to 64 years old. When we think about adults, we generally think about 18 to 64, and then elderly by definition is 65 and above. But when it comes to devices, FDA mandates it as 22. So that's what we are looking at here with Rejoyn. And they had to have a primary DSM-5 diagnosis of MDD without psychotic or mixed features, monotherapy antidepressant, and with a reported history in the current depressive episode of inadequate response. And they used the MGH-ATRQ to define what that was, make sure it was adequate dose and duration. Inadequate response, less than 50% reduction in the depressive symptom severity. It had to be a stable dose of the antidepressant for at least 4 weeks prior to the baseline. And then to screen people in to make sure they had the right severity, the Hamilton 17 had to be at least 18. 

Remember the primary endpoint that they used in the trial was MADRS, but they screened people in with the HAMD. It's good to not give people the same screen in as they use for the primary endpoint, because you can learn, people get better at taking the test the more times they take it. So they use a different one to avoid inflation of the scores. 

Then the exclusion criteria, if they had any lifetime psychotic or bipolar disorder diagnosis, current PTSD, panic, OCD, substance/alcohol use disorder, if they had inadequate response to more than one adequate trial of the antidepressants in the current episode, this was not for treatment-resistant depression. This was for inadequately treated depression. And then if they had any augmentation treatment within 90 days of the screening, so ECT, neuromodulation, psychotherapy within 90 days. And then again, not just not treatment resistant but not treatment refractory based off the investigator discretion. And then if they were a significant risk of suicide, this is not a treatment for people who are a significant suicide risk. 

Dr Craig Chepke: Primary endpoint was the modified intent-to-treat population. We'll go over what that means. And it was change from baseline to week 6 in the MADRS, and then they assessed durability of the MADRS assessments at week 6, 8, and 10. 

Also had a number of secondary endpoints. The GAD-7, had assessments of those, and durability also at 6, 8, and 10. MADRS response/remission, CGI-S, PHQ-9, the WHODAS 2.0, the EQ-5D-5L. And then the overall individual and participant and healthcare provider satisfaction, were they satisfied with the intervention? And obviously safety, the adverse events, how frequent/severe they were, what the protocol-defined serious AEs were, and if they discontinued because of an adverse event. 

Demographics, mostly were people in their early 40s. They were predominantly female, predominantly white in the clinical trial, and they were balanced between the groups overall.

These are the statistics about the severity of the illness based off of the MADRS, GAD-7, CGI-S, PHQ-9, and the Hamilton Depression 17 scale. Let you take a look at those. I'm not going to read it, but these are people who had a fair amount of pathology. 

Finally, let's get to the efficacy data. Here's the primary endpoint displayed 2 ways. On the right, the protocol-defined primary endpoint, the modified intent-to-treat, and the full intent-to-treat. So, there were a few people that were not in the modified intent-to-treat. They had a definition there of what those were in terms of, did they have enough interventions, enough assessments. And this is common. We see this, the modified intent-to-treat population used in most studies these days in pharmaceuticals, and that's what the primary was here. So the change here was a difference of 1.78, and that was not statistically significant.

Dr Craig Chepke: Now if you look at the overall intent-to-treat population at week 6, well that was statistically significant with the change of over 2 points on the MADRS. A -2.12 for the Rejoyn, which is the lighter blue arm, the royal blue arm versus the dark blue, midnight blue arm, which is the sham arm plus antidepressant. And that one would have met statistical significance had that been the primary endpoint. So, in the modified intent-to-treat group, the treatment effect did persist past week 6, with a trend that favored the continued improvement, and the mean change from baseline to week 10 in the MADRS total score was -10.96 in the Rejoyn and antidepressant compared to -9.93 in the sham plus antidepressant group. 

Now here's the responders and remitters here. So that you can look at it first and have both the ITT and the mITT partial or full response, Rejoyn in the lighter blue, darker blue for the sham, all plus antidepressant, and full response. So both of those were nominally significant in the ITT group. The partial or full response was in the mITT, and then trends favoring the Rejoyn plus antidepressant group in partial response and remission in both ITT and mITT.

So I want to come back to that meaningful change analysis. I mentioned it before. This is looking at individual patients, not just the overall mean. Because the overall mean is what we just looked at, and that can obscure people who do really well, because not everyone obviously is going to get the mean. That's what makes it a mean. Some people will get less than that. Some people get no benefits, some people will get worse, some people will get a little bit more than the mean out of it. Some people get a lot better than the mean. And that kind of nuance I think is important to capture, and that's part of what the meaningful within-patient change tries to start to look at. So for the Mirai trial, the meaningful within-person thresholds, they used anchors for these, and the CGI-S, a 1-point improvement on the CGI-S is pretty universally considered a clinically meaningful one.

So what difference in the Mirai trial, people who had a 1-point improvement in the CGI-S in the Mirai trial, what improvement did they get in the MADRS? It was 8. People who had a PHQ-9 improvement of 6, which is defined as a clinically meaningful improvement, what score change did they get on the MADRS? Well that was a 10, so that's how they anchored it in from both sides. One from the CGI-S, one from the PHQ-9, usually well-defined clinician- and patient-reported anchors, and it comes up with an 8- to 10-point improvement on the MADRS to be clinically meaningful within-person for this clinical trial. 

Dr Craig Chepke: So how many people met this? Well, if you’re looking at that, it's a 5.4% between-group difference. So we can look at it based off the odds ratio, which would translate into they were 24% more likely to achieve an 8-point improvement compared with patients in the sham plus antidepressant group. Or if you look at it in terms of the relative risk, then they were 12% more likely to experience the improvement compared with the sham plus antidepressant group for an 8-point meaningful change threshold. So if you looked at the 8, that's what it was. If you looked at what was the even larger improvement, what percentage of people got to 10-point improvement between the two? Well, that was more in favor of the Rejoyn plus antidepressant group, and that one they were 47% more likely to achieve a 10-point improvement and 26% more likely to experience that improvement compared with the sham plus the antidepressant group. 

You can visualize that yet another way, because those are just 2 endpoints. The 10 and the 8, those are the two in the center vertical dash lines, the 8 and the 10, and the Rejoyn is the blue arm that is almost always above the red arm. The red arm is the sham plus antidepressant group. And it's almost always above, because if you get far enough into the deterioration, which is to the right as indicated by the arrow, then the sham plus antidepressant group supersedes it, because going up means that there is a greater proportion of patients who met that threshold. So that's what we would expect, that the sham plus antidepressant group, probably more people should get worse than an active treatment that works. And then going towards improvement towards the left, the blue should be over the red because that means that it's an active treatment that is working, that it does favor the Rejoyn plus antidepressant group. And that's what we see here. If you go out even further to the left, there are some places like the 20-point improvement, that has a pretty wide separation there. So if you look at only the percentage of people who got to 20%, a 20-point improvement, it's not a lot of people in either arm, but the Rejoyn group was quite a bit above, the Rejoyn plus antidepressant group, was quite a bit above the sham plus antidepressant group there. So I think these are fun graphs to look at because whatever you think the clinically meaningful improvement is, dial up your own score, and then you pick what that is and look at it; is the active-treatment group greater than the comparator group? 

And a couple of exploratory endpoints, this is the PHQ-9. So if we have anyone out there who works in primary care especially, you probably use the PHQ-9 on a daily basis. These are probably going to be numbers that are going to resonate with you. So in the ITT population, you can see the difference there, -5.15 for the sham plus ADT and 6.93 improvement in the Rejoyn plus ADT, and respectively a 5.10 and 6.68 in the modified ITT group.

So clinically meaningful improvements and categorically, the people in both populations went from moderately severe to mild in the Rejoyn + ADT group. Well that's good. In the Rejoyn group, they went from moderately severe to mild. That's great. In the sham plus antidepressant group, they went from moderately severe to moderate. Well that's better, but it's not as great as going from moderately severe to mild. So 2 categorical shifts versus just the 1 in the PHQ-9 for the sham plus antidepressant group. The CGI-S, yet another one. Overall, the mean within-group change in both the ITT and mITT populations represent a clinically meaningful improvement and categorically going from moderately ill to mildly ill. So whichever one you're looking at, ITT or mITT, moderately to mildly ill. 

What about the GAD-7? So this is splitting the population to those who had a GAD score at least 10 or higher, meaning moderate or higher anxiety, or less than a 10, meaning they had less than moderate anxiety. And that actually is pretty interesting that the Rejoyn group now going down is better here because you're reducing the MADRS score, and you want the MADRS to go down, lower depressant symptomatology. Those who had the GAD-7 score greater than 10 are all the people in this group. And for these people only then there was a pretty large separation here that was nominally significant at weeks 2, 4, and 6. So 5.39 in the sham plus antidepressant and 9.01-point reduction, going down is better once again, in the Rejoyn plus antidepressant group. So we can't make any statistical comparisons here, because this was not controlled for multiplicity in that way, but it indicates that the people with moderate or higher anxiety symptoms did see a significant improvement nominally in the MADRS. 

Dr Craig Chepke: Now, what if we limit it down to only people who completed at least 12 treatment sessions? Because what if people only completed 3 or 4 sessions? One would expect they wouldn't get as better. So if you limit it only to the people who would be adherent, then you see a potentially greater therapeutic effect. So -7.47 in the sham plus antidepressant, 9.21 improvement in the Rejoyn plus antidepressant group. 

Now what about their impressions about satisfaction? Overall, there was a favorable impression for Rejoyn plus antidepressant in 85% of the participants and 82.5% roughly for the healthcare providers, in terms of convenience of the app for delivering treatment. Those are pretty good numbers, that the participants and the clinicians were satisfied and had a favorable impression of the Rejoyn treatment. 

So the safety data, I already told you the answer is no treatment-emergent adverse event was assessed as related to Rejoyn during the trial. No discontinuations due to TEAEs at all, and 1 discontinuation due to lack of efficacy, but that was in the sham plus antidepressant group. No serious TEAEs, and there was 1 in the treatment period; during the extension period, there was a serious TEAE of a TIA, transient ischemic attack, but assessed as not related to Rejoyn. The only TEAE that was at least 2% and had a greater numerical percentage rate than the sham plus antidepressant was headache—2.1% of those in the Rejoyn plus antidepressant group versus 1.6 in the sham plus antidepressant group. 

In terms of the suicidality, 1 participant, well just going for the suicidality, 1 in the Rejoyn plus antidepressant group had worsening depressive symptoms, 3% roughly had clinically important suicidality compared with 4.8% in the sham plus antidepressant group. So just as a reminder, it's not intended to be used standalone or substitute for medications. They should continue their current treatment. 

Dr Craig Chepke: So let's sum things up and then we'll get to the Q&A. Digital therapeutics are subject to regulation as medical devices, but it's important to define those appropriate control conditions, and that is a challenge. I went through in detail why that's so challenging, so this is a work in progress with these clinical trials. And the meta-analyses of RCTs are showing that CBT-based digital therapeutics and mobile app interventions may be effective reducing the depressive and or anxiety symptoms. 

Rejoyn specifically is an FDA-cleared prescription digital therapeutic for the treatment of MDD symptoms as an adjunct to clinician-managed outpatient care for adult patients with MDD that are 22 years of age or older who are on antidepressant medication. Rejoyn as a reminder is a smartphone app-based digital therapeutic that gives 6 weeks of treatment with 3 portions. The EFMT exercises, CBT-based lessons, and the personalized reminders and messages. The effectiveness and safety of Rejoyn were supported by the results of the Mirai trial that I just went through. So there's a QR code that you can scan if you want more information about Rejoyn, and you can get the QR code scanned. They'll give you the Clinician Brief Summary. It's like the prescribing information but for digital therapeutics. It is FDA cleared as of now and will be available in app stores later this year.

Dr Tarolyn Carlton: Thank you, Dr Chepke. That was a lot of information, jam packed in, but again, I think really important information about how our landscape is changing for depression. So talking about the different mechanism of action, different treatment modalities, and then providing some information about how the data is looked at for these types of treatments. So thank you very much and on behalf of Otsuka Pharmaceutical Development and Commercialization, we thank all of you guys for participating and spending your afternoon with us. Have a wonderful rest of your day. Thank you.

Dr Craig Chepke: Bye everybody.

Dr Tarolyn Carlton: Bye.

© 2024 Otsuka America Pharmaceutical, Inc. All rights reserved. June 2024          20US24EBP0038