The Mental Health Forecast explores cutting-edge developments shaping the future of mental healthcare. Each episode features conversations with researchers and innovators about emerging therapies, technologies, and approaches to emotional wellbeing. From AI-powered therapy tools to advances in neuroscience, we examine how these changes might transform mental health treatment. Join us as we investigate what's on the horizon and discover how tomorrow's solutions could help address today's mental health challenges.
Arjun Nanda (00:01.384)
All right. Welcome back to the podcast where we explore the cutting edge of mental health care and the minds behind all the innovations. I'm your host, Dr. Arjun Nanda. Today, we are thrilled to have Dr. Carlene MacMillan, who is a distinguished child and adolescent psychiatrist who's bridging the gap between traditional psychiatry and modern technology. Awesome.
Carlene MacMillan (00:27.694)
to be here.
Arjun Nanda (00:30.6)
Dr. MacMillan is currently working with Osmind, which is an electronic health record company at the forefront of innovation in general and interventional psychiatry. So thanks so much for coming on.
Carlene MacMillan (00:44.974)
Yeah, excited to be here and love the love the podcast idea. It's great.
Arjun Nanda (00:48.552)
Great. Awesome. So maybe we can dive into some of the things that we've connected about over on Twitter or X. I know we, who knows? I know, right? It's so tough to actually know what to actually call it, but yeah, I mean, there's something in my heart that refers to it as Twitter, but I'm like, no, it's X now. So.
Carlene MacMillan (01:01.742)
Yeah, what do we call it these days? I don't know. You still call them tweets? I don't know.
Arjun Nanda (01:18.696)
Yeah, I think we connected recently over the DSM -5 cross -cutting measure. I know that you've been working on that. So I wanted to start off by exploring a little bit about that, about the DSM -5 cross -cutting measure and your work on that.
Carlene MacMillan (01:36.43)
Sure, yeah. So actually, you know, back in 2013, the DSM -5 came out. And I was actually a trainer for the APA. I went around Massachusetts at the time, and it was sort of pre -Zoom, so we would go around and explain the DSM, and they had a whole little slide deck. And so I got really into understanding, like, what had changed from DSM -4, and why, and why not. And I really, you know, recall in Section 3, which was for their emerging,
diagnoses and measures, they had a number of really interesting things. And one of them was the DSM -5 cross -cutting measure. And it covers like 13 domains of psychopathology. There's a version for adults, as well as adolescents and kids. And it's meant to be almost like, I think of it as like a psych review of systems. Like it's meant to be a way to have a conversation with patients that you're just meeting, also when they're coming in for a follow -up to have them fill it out and kind of guide what you're going to
talk about and also to do additional screening and measurement based care depending on how they answer. So I thought it was really cool. There's some other ones in there as one for personality, all of this, but I didn't really think much of it beyond, that's neat because it was just in the book and then it was just on the APA's website as a PDF. And I'm not going to print out a PDF, how many of us sent to the patients. And a lot of us are like that.
we maybe want to do measurement -based care, we want to use some of these tools, but unless they're made like really easy in our workflow for both us and patients, this is not going to happen, right? It's just you have to be, it's not going to happen. So we at Osmind have been collaborating with the APA around, you know, a number of different things around measurement -based care and then just sort of moving the field forward. And one of our initial things is actually to introduce this thing to the world. So we added
it to the Osmind EHR, made it really easy to use, did some education around it, and we hope also beyond Osmind that other people will start to use it. So that's what we did, that's what we started doing, and it's going really well. We're hoping to get some interesting feedback, as well as just the data itself is helping me with my patients when I see them. So it's a really cool tool. I wish I had it for the past decade.
Arjun Nanda (04:03.432)
Yeah, I think ever since we connected about this too, I've been using it in my practice and it's been very helpful to get a snapshot of what's going on with the patient before they come in. Just as a general screening tool, it's been incredibly helpful. Yeah.
Carlene MacMillan (04:24.782)
That's good to hear. Yeah, that's what we're finding too with the folks that are using Osmind that many clinics are not only trying it, but they're now just doing it for all their patients. They're just starting to use it. So it shows me that, look, there's value here. And we're able to show to the APA, like, hey, this is a valuable tool. It's just it needs to be executed properly in terms of how people can use it in their practice. Are you just using the PDF? Or how are you doing? And I'm curious.
Arjun Nanda (04:50.92)
I work in the community setting. I'm just using a PDF that I'm printing out and giving to patients, and then reviewing it prior to them coming in. We're still relatively old school, but the information that's coming in is incredibly valuable. Hopefully, we can get things more up to date.
Carlene MacMillan (04:59.63)
Mm -hmm.
Carlene MacMillan (05:13.166)
That's great. That's great to hear.
Arjun Nanda (05:20.552)
from a technology standpoint, but you know how it works, I'm sure.
Carlene MacMillan (05:26.478)
Yeah, I mean that's part of why I actually joined Ozmind.
like two and a half years ago, almost three years now, because I had a large private practice in New York and we did a lot of TMS and esketamine. And for that, you have to do mission. You have to send PHQ9s. You just have to do it in terms of insurance. And so the way we were having to send it, we were using like Google Forms and Google Sheets. And it was a whole fiasco. I mean, it's very fractured and very common of what people were doing in this space. And I didn't, I just didn't like it.
like that at all and I was trying to find something that would solve it and I really couldn't until I met the founders at Ozmind and they initially were just building this like engine for measurement informed care to make it like really easy and then they said actually we're gonna be making like this whole EHR. that's good you smile I have a weird dream lifelong dream of building an EHR. We all kind of say that we hate these EHRs I just want to build my own well.
Arjun Nanda (06:25.768)
Ha ha ha.
Yeah.
Carlene MacMillan (06:31.278)
takes a lot of resources to build your own, particularly if you're not an engineer. But I was like, wow, you guys have the means to do this. You have the backing to do this. Like, this is awesome. And so I ended up doing that. And, you know, I just love it because now I can, you know, get things like the cross -cutting measure out to my colleagues. And, you know, it's great.
Arjun Nanda (06:32.584)
Hahaha.
Arjun Nanda (06:51.752)
Right. Can I ask? Because yeah, I think many of us have this, we all have our own takes and ideas of what could be improved about EHRs and everything. Can I ask a little bit about how that journey has been for you in terms of being one of the people that's actually working on EHRs and developing it and improving it?
Carlene MacMillan (07:16.43)
Yeah, I mean, so I think I had probably used over 10, 10, 11, I guess, 11 counting off mine, EHRs in my career from everything from the hospital systems like like Epic and Meditech, which, yes, definitely looks like it's from the 80s. But it and then, you know, the other smaller ones that are specific to psychiatry or mental health, like Luminello, which recently shut down and simple practice and stuff like that, the sort of
Arjun Nanda (07:33.992)
Right.
Carlene MacMillan (07:46.336)
usual suspects and you know they're all basically EHRs were formed to make it easier to bill in particular to bill insurance and so they really center that and they never really centered the way clinicians think the way that we want to document.
Arjun Nanda (07:58.216)
Mm -hmm.
Carlene MacMillan (08:07.15)
our time, you know, maybe we don't want to click a million boxes, that we all have our own ways of documenting. And so they all bit too lacking. And what I found particularly for psychiatry as a compared to therapy, or just another field of medicine was that the medical EHRs that were out there were like,
Arjun Nanda (08:11.144)
Right.
Carlene MacMillan (08:31.15)
too much stuff that we didn't need and not like a mental status exam template, you know, and not like the therapy stuff. And so if you had, like we had therapists at our practice and they didn't want to use these medical EHRs. They didn't have this, they didn't have treatment, they didn't have the stuff that they were looking for.
Arjun Nanda (08:32.008)
Hmm.
Yeah.
Carlene MacMillan (08:47.886)
didn't have measurement -based care, it was too much. And then the ones that are for therapists, they don't have Eberscribe or Vitals, you know, labs. And so it was like, there's nothing here that was for us. And there's a couple of companies in the space that, you know, now that Luminello's not, there's really much fewer, but they were also like,
Arjun Nanda (08:56.776)
Right.
Carlene MacMillan (09:11.342)
how do I say this, just kind of legacy technology didn't look and feel modern. And it wasn't like at our practice, we really wanted to be able to use the data, like use real world evidence to be able to say, hey, this TMS protocol maybe works better for these people and all of that. And with most EHRs today, even the medical ones, it's very hard to get that information into a form that.
is actually usable. Like you generally have to get a research assistant to like comb through all the charts and it's really terrible. So when we, with Osmind is kind of built from the ground up to have things a bit more structured, but still give, you know, flexibility to clinicians. And we are centering the clinical workflows, particularly those specific to interventional psychiatry, although we have general folks using it as well.
and I think, yeah, we can do the billing, but we're not like the billers dream EHR because those are, those are hated by clinicians. So we're trying to strike that balance. And I also think just in this space of EHRs, we have this concept of note bloat, right? I'm sure you've seen charts that are big and got just copy forward and all this stuff. And you're like, well, I don't even know what happened in that visit. It's all this crap and things get carried forward that don't belong. I actually think that.
Arjun Nanda (10:06.28)
Yeah.
Arjun Nanda (10:18.408)
yeah.
Carlene MacMillan (10:29.824)
is an area where AI is making an impact. And we all like in the past year and a half, a number of companies have come up that are doing the AI scribe thing. And I you know, I've been using those, I think they're really interesting, they're evolving really quickly. And I think that really the holy grail will be when those kind of point solution scribes are able to be deeply integrated within the HR, because the scribe kind of only knows like what you said in the visit, and it doesn't necessarily have all the background.
And so I think the more we can get towards that, the way like the way I chart is just so different now and I don't dread it. And I used to have to carry around that information. If I didn't document with the patient, do all the typing, I'd have to like remember and do it at night or on the weekend. And that was terrible. And the notes weren't very good because I don't remember everything. Now I do them and like I do the AI thing. And then when I review them, it's like pretty painless. Like I sometimes have to fix.
like things in it, little things, but it's sort of the essence of what happened is just there and I think that cognitive load is, you know, really huge to be lifted. So I'm very excited about how that will change the way we think of EHRs. Yeah.
Arjun Nanda (11:47.048)
Yeah, I can imagine that this is music to a lot of people's ears, not having that mental load. I mean, there is obviously the load when we are seeing patients and we're making decisions, but just the dreading writing notes is I think a big problem for most of us. Yeah.
Carlene MacMillan (12:08.206)
Yeah, yeah. I mean, and even like, you know, since I've experimented with, you know, various different scribes, and they all have their strengths and weaknesses for psychiatrists. But I used to just like, be really reluctant even to see new consultations and only do a certain number of weeks because all this documentation is just like the note, seeing the patient was fine, coming with the treatment plan was fine, but the note was a real drag.
Now, it's, you do these beautiful intake notes and I just have to make sure they look accurate and then they're, they're much better than what I was doing.
Arjun Nanda (12:45.608)
Right. That's phenomenal. And I think that, yeah, a lot of people would look forward to using something like that, compared to whatever they're using now.
Carlene MacMillan (12:56.622)
Yeah, yeah, we're seeing it more common. You know, Epic is partnered with one of these tools. Different things are thing. And at first, people were like, this is weird. And you know, patients are not going to like this. Patients don't seem to really mind at all, because also the clinicians are more present during the thing. And sometimes I'll even share what it wrote and be like, isn't this funny? It thought this. Or this is the impression it had of our conversation. And it's not like I've written that thing. It's like a moment where we can reflect together.
how something might be coming across.
Arjun Nanda (13:28.712)
Yeah, yeah, yeah, absolutely. I've used some AI scribes before just to try it out with my work and I've found that for the most part, it really accurately gets at what is really happening in the session. I do like writing my own assessment and plans because that's where a lot of the integration of the session happens and...
Carlene MacMillan (13:51.374)
Yes.
Arjun Nanda (13:58.728)
I found that it just works better with understanding what exactly is going on with this patient. So everyone has their own ways or methods, I guess, of integrating these sort of technologies into their practice.
Carlene MacMillan (14:13.006)
Mm hmm. Yeah. And that's the assessment and plan part. It's like they can't say what's it doesn't know what's in your head. So I think I've actually even maybe changed some of my way I'm interacting. I think I always try to sort of share that assessment with a patient in the plan. And so I'm trying to be like showing my work and like showing how I'm thinking, because I know if I don't, then I'll have to write later. And, you know, say, yeah, I think, but that's the little things like medication doses of like, how is that 20 milligrams of Prozac treating?
Arjun Nanda (14:34.44)
Alright.
Carlene MacMillan (14:42.912)
you, I might otherwise have said, how's the Prozac going, right, but I needed to know that it's 20 milligrams. So we'll get to a point where we're better at that. And some of them like even kind of learn your style and you can make templates of like what you want. Like I have one where the way I like to do my mental status, the way I like to do my risk assessment, I'm able to really kind of fine tune it. And so the note that like sounds like me, like if you use this and saw the same patient, like the note will look different, your style. And
Arjun Nanda (14:46.984)
Sure.
Arjun Nanda (15:12.488)
Yeah. Yeah, that's awesome. Very exciting time to be in this. Awesome. With regards to Ozmind, and you were saying that there's a lot of measurement -based care that it focuses on. I wanted to get your take on...
Carlene MacMillan (15:12.816)
I think that's cool.
Carlene MacMillan (15:16.942)
Yes, very exciting time to be in this health check. Yes.
Arjun Nanda (15:36.104)
you know, how it's being used right now, even just generally, not even within Osmind, but in general, because I also find that it's not used as much as it could be. So I'd love to get your take on that.
Carlene MacMillan (15:40.014)
Mm -hmm.
Carlene MacMillan (15:49.166)
Yeah, so I think, you know, a number of studies, they always come out with this like 20 % number for mental health clinicians, including psychiatrists that are using it and 80 % that are not. I'm hoping that will change and maybe even has changed. We just haven't measured it yet, but it's definitely a lower, lower number. And many patients that I see me for the first time coming from somewhere else, they've never done these. The people that have done them are people that have gotten ketamine, esketamine, and TMS.
Because all of those like procedural things where either you know patients paying privately and you need to kind of show it's working or you need to show to insurance Like those of us in interventional are just like doing this like the web whether we're just using it for the insurance or we've really like Gone to town and really leaned in that that group is much more
just all in on that because we kind of have to be but at the American Psychiatric Association, I'm part of a smaller measurement based care work group that put out a white paper around this last year and now we're doing a survey of all the members and even non members anyone that's a mental health clinician can fill it out about like what are what are they using? Why aren't they using? What's the biggest barrier? So we're hoping to have you know read out of that within the year and that's where we're going to say okay this is
Arjun Nanda (16:43.752)
Mm -hmm.
Carlene MacMillan (17:10.4)
where you really need to lean in and improve. And I think it's like you have to do something that's valuable. So if it doesn't give valuable information and you're just doing it for like insurance or something or to add a code to your bill, like that's not intrinsically valuable. So like the PHQ -9 people don't love it. Like some of the questions like, you know, it doesn't say if they have less sleep or more sleep or less appetite or more appetite or it's not really very fine tuned. So you have to kind of ask the thing.
anyway so you know is that really going to be a value add?
Whereas something like, I think the DSM crosscutting measure, for example, people that that is a value add because it's covering a lot of ground efficiently and it's leading them to have deeper conversations. It's not just this rote thing. And then other things like people like the Y box for OCD. Like if you see a lot of OCD, people use that Y box and it's very detailed and it really kind of gets it gets that thing. So we have to think, does this measure like is a clit actually going to want it? You know,
And then is the patient actually going to want to fill it out and not feel that it's like, you know, offensive to them in some way or too invasive? And that's always kind of a balancing act. And then I think a big thing that's a barrier and very common, and there's actually some studies on this, that there's not a feedback loop. So a patient will fill something out that the doctor sent and then the doctor either never looks at the thing or doesn't actually comment on it in the session.
So the patience, I don't know why I'm filling this out, you know, and so I think it's really important to bring the topic into the session. And even if there's a discrepancy, right, like if somebody is telling me they're fine and they're great, like, well, your PHQ -9 is, you know, 25, so I'm not.
Arjun Nanda (18:49.8)
Yeah.
Carlene MacMillan (19:03.502)
seems what is going on here with this or they tell me that they are, you know, like terrible and they're ph. So those discrepancies are interesting to me as a clinician. And I like to use them and also showing the patients the results. People like to see the graph. They believe pictures, you know, so there are those things we need to do more of. I think our platform makes some of that easier. But at the end of the day, you can lead a horse to water, but clinicians,
Arjun Nanda (19:06.344)
Yeah.
Arjun Nanda (19:16.072)
Yep.
Arjun Nanda (19:23.016)
Right.
Carlene MacMillan (19:33.408)
need to feel that this is going to add to their, you know, their toolkit meaningfully. And I think many of us didn't even get trained on it in residence. Very variable. I don't know in your residency if it was emphasized or not. Certainly it was not in mine. And so people don't do it.
Arjun Nanda (19:48.2)
Yeah.
Yeah. I think for something like the, the PHQ -9 and the GAD -7, those screeners, it will detect obviously depression symptoms or anxiety symptoms. The issue that I've run into is that there's a lot of symptom overlap between, even if they do screen positive with the PHQ -9, how sure are we that it's not as a result of earlier childhood trauma or...
Carlene MacMillan (20:10.062)
Mm -hmm.
Carlene MacMillan (20:19.47)
Exactly.
Arjun Nanda (20:19.816)
or many other things. So that leads to the whole problem of these transdiagnostic concepts and what really, are we fully picking up on the things that we should be picking up on when we're putting some of these scales out? That's the thing that I actually like a lot about the cost -cutting measure is that it tries to pick up on everything. So you're getting all these different symptoms. So.
Carlene MacMillan (20:40.846)
Mm -hmm.
Arjun Nanda (20:46.408)
Maybe we can actually just talk about that for a second.
Carlene MacMillan (20:49.806)
Sure. Yeah, no, I think that so I think the DSM cross cutting like I mentioned, it was like a psychorefuse systems. I think that's at its most sort of basic level. That's what it is. But it actually was designed to be transdiagnostic, you know, meaning that it's to get that whole idea about a person and acknowledging that the DSM five, you know, doesn't really carve nature at its joints, it makes approximations of things. And I think the hope is that in future DSMs, we could perhaps take a more
the NIMH and the R .DOT criteria that they're taking a bit more sort of domain approach. And actually the NIMH, if they're funding grants now, they're actually requesting that this measure be used because it actually can map on through factor analysis to those domains. So it really plays well with that. You can use it to get your sort of classic bipolar and all of that and go down those paths. But I also like to see, if the person is in the
Arjun Nanda (21:39.56)
Mm -hmm. Mm -hmm.
Carlene MacMillan (21:49.632)
person scores really high on it, the score itself isn't really meaningful, but it is relatively, you know, it's like they were really high and then in the follow up, they're really low. I'm basically saying, okay, there's like less overall psychopathology here. Now there's less, less call like a P factor. There's less of that psychopathology and there's actually some interesting factor analysis on, you know, potentially what you could do with it to get a score that isn't just a total score that certain symptoms like, you know,
probably like suicide, for example, are going to get higher rate than some of the others. So I think that's a really interesting way because I just see complex patients and, you know, rather than like a laundry list of five diagnoses, the transdiagnostic approach is like, well, you're having difficulty in this area and this area and this is why the treatment plan is going to be X, Y or Z. If you just do the PHQ -9 and look, we all do it, it's sort of part of life, especially if you're interventional.
Arjun Nanda (22:23.4)
Mm -hmm.
Carlene MacMillan (22:49.408)
You're gonna miss a lot of things like a lot of people just use GAD7 and PHD9 for all their new patients and you're gonna see a lot of depression and anxiety and you're gonna miss the PTSD, OCD, dissociation, you know, personality. I love the questions on the cross -cutting measure. There's two of them about perhaps there's some personality concerns going on there because they're not like in your face. They're not sort of alienating. They're just kind of getting it if you have difficulty interpersonally and feel disconnected and all of that.
And so I think that's a really nice, like that really helps me at intake if people have answered those.
Arjun Nanda (23:24.2)
Right, right. Excellent. So I guess my next question here would be related to the DSM. I guess any thoughts that you have about where you would hope that the DSM goes from here? The research domain criteria, I mean, I'm skeptical if it would be used so much in a clinical setting.
Carlene MacMillan (23:52.238)
Mm -hmm.
Arjun Nanda (23:54.248)
but, and the DSM really is like the pathway for a lot of clinicians to diagnose and assess patients. So I'm curious about your thoughts there about where you're hoping the DSM will go and especially with these trans diagnostic concepts.
Carlene MacMillan (24:11.054)
Yeah, no, I mean, I think.
I don't think it's going to be like gone away, right? I think we're still going to have these concepts, but every time the DSM comes out, it's bigger and there's more variants and it's more nuanced. And I guess I would hope that rather than sort of exhaustively categorizing things that have no relevance clinically, that we would focus more on like, well, is this diagnosis going to change a treatment plan? Is this going to change a prognosis or are we just like,
being like really specific in our nuanced minutiae, right? Like you don't, when you have cancer, people have different symptoms. They don't sort of say cancer with, you know, nausea and X, Y, like you kind of just say like what stage it is, like how bad is it? And that's gonna dictate, whereas the DSM, I think.
Arjun Nanda (24:59.08)
Yeah. Right.
Carlene MacMillan (25:04.782)
It's always taken this much more just sort of, it doesn't really look at the causes. It doesn't really, it's just looking like, is this a concept that we can describe? And as a result, you can just get more and more granular.
Like there's no, where's the end point of just getting more and more granular? So I think we need to swing back more towards like, is this like depression? Like this is an anxious subtype of depression, not if it's, you know, all of these little qualifiers that we see. Cause this is unwieldy, it's very unwieldy. And in addition, I think, you know, they got rid of the GAF score, the global assessment of functioning, and they hoped that people would use a measure called the HUDAS, but I think we've lost something.
around functional impairment and you know the key is like is this a disorder or not is it functionally impairing and so I hope that the future DSM has a just more encapsulated tidy way to do that I get that the GAAF wasn't reliable over time unless people were trained on it but we've got to be we got to have something that kind of sort of how impaired is this person right and we don't.
Arjun Nanda (26:10.184)
Yeah. Yeah.
Carlene MacMillan (26:16.046)
So yeah, that, and then I also really do hope that, yeah, sure, there's gonna be book. Everyone likes to have the book on their shelf. It's like, are you really a psychiatrist if you don't have the book on your shelf? But I think that there is an app.
version of the DSM, but it's kind of just like the book. So I'm hoping that we'll do more technology wise to make it easier for people to engage with the DSM to use measures like the cross cutting measure that those things can be really kind of centered and not just like in the back of the book and not just this sort of static one dimensional thing. So that's a lot, but I do think, you know, at the APA, they had a panel with some of the folks that are involved in the future of the DSM.
and they had someone from the NIMH as well and someone from Hopkins, which happens has its own thing going on over there. But there was seemed to be a real appetite to kind of think.
Arjun Nanda (27:08.328)
No.
Carlene MacMillan (27:13.71)
Like we're not gonna just do the same thing. We're not gonna just publish a book that's 1200 pages instead of 900. So I'm hoping that that will come true and we'll see. And part of our collaboration at Osmite is actually to hopefully bring the voices of the clinicians and bring the experiences to help inform what we actually want and need.
Arjun Nanda (27:35.24)
Yeah, yeah.
Carlene MacMillan (27:37.55)
Yeah, I don't know. What do you think? What's your prediction?
Arjun Nanda (27:39.688)
Yeah, I mean, I think as our assessments get...
more as there's more and more information about patients. I'm also interested in other aspects like physiologic aspects of depression that we're not able to really pick up on through questionnaires or screeners that I think now with greater integration of technology, I think that we could be able to pick up on that even in just in regular...
Carlene MacMillan (27:59.118)
Mm -hmm.
Carlene MacMillan (28:07.566)
Yes.
Arjun Nanda (28:19.432)
clinical practice, we have access to wearables. I'm impressed that we don't use vitals or things like heart rate variability and routine screening for psychiatric conditions when we know that this is something that happens with patients, for example, with PTSD or anxiety. So yeah, I'm hopeful that that also starts to get integrated into our assessments.
Carlene MacMillan (28:22.574)
Mm -hmm.
Carlene MacMillan (28:49.582)
Mm -hmm.
Arjun Nanda (28:49.64)
I'm not sure if the DSM will integrate that, but I'm still hopeful that we as clinicians will.
Carlene MacMillan (29:00.046)
Yeah, I mean, I saw I think that's I'm glad you brought that up because I think that in terms of the DSM, like I probably don't foresee them saying, well, in order to get to diagnose with depression, then, you know, you have to have like a decrease in the number of steps that you've taken or something like that. But if you look at the criteria, it's going to say things like decreased energy, not moving as much or disrupted sleep or early in awakenings. It's it's not going to really say how we test that. And right now, we just do we ask patients, most of us just ask.
Arjun Nanda (29:29.096)
Yeah. Yeah.
Carlene MacMillan (29:29.952)
patients. Some of us use measurement based care and fewer still use tools like I have an aura ring. A lot of people have Apple watches. They have this data and it really like you know even with Osmind we don't have that in there yet like we really need those tools to get that data at the fingertips. I have many I mean psychiatrists like we treat insomnia like so many people have sleep disturbances and I use
to be kind of just like, okay, well, you know what, how many hours sleep or whatever, I didn't really have any sense of what was going on. And now I've started to really ask patients to show me the sleep data from their, their phone or watch or ring and starting to be like, okay, well, you actually don't have enough deep sleep. So let's work on how we can get that or you have the early awakening. So maybe that's depression or you know, you have no, I have someone recently who just showed me these things and there was like no REM sleep. Like,
Like zero, none. And so we're like, okay, the meds that you're actually taking, you know, they actually reduce REM sleep, right? Like they're knocking you out, but I don't think they're really giving us that result. And, you know, they don't believe you because they like the meds. But now they can see like, okay, yes, I stopped this med and my REM sleep is back now. So I think that's totally amazing. And I think we will need to do more of that.
Arjun Nanda (30:30.984)
Yeah. Yeah.
Arjun Nanda (30:36.776)
you
Arjun Nanda (30:50.664)
Yeah.
Carlene MacMillan (30:56.494)
be very excited when that becomes just kind of the standard of care. And I think also we have to be mindful of like, I don't want a patient sending me their sleep data every night. I don't care to receive that, right? I want to review it during a session. And ideally, I might like let's say someone has bipolar, like, yeah, I don't want to know every night, but I would love to have it generate an alert that says, Hey, patient X slept two hours last night, or, you know, all of a sudden, like to
generate for me in surface, like the problems. And that's where I think actually the AI stuff will really come in. Like my aura ring will tell me like, hey, you should take it easy today. Like you didn't seem like your sleep was this restful. Did you have a late night meal? Like that, those kinds of things, I think if we can get them to the place where psychiatrists can use them and not feel like they're just drowning in data, that's, that's going to be the key. That's going to be the key for sure.
Arjun Nanda (31:54.696)
Yeah. Awesome. I feel similarly that having the data at your fingertips is helpful, but having succinct data and usable information will go the farthest way. Any thoughts about...
Carlene MacMillan (32:12.238)
Yes. Yes.
Arjun Nanda (32:18.28)
about this integration of technology and mental health and any sort of exciting technological advancements for you that you're excited about with regards to mental health?
Carlene MacMillan (32:29.71)
excited about. Yeah. So yes, there is actually so speaking of measurement based care, you know, we talked about PDFs and paper and now we have them in the HRs. But that's kind of measurement based care, like 1 .0, you know, and so I think that actually, it's all there's always going to be a value in asking patients what they're feeling and what's bothering them. There's, we're never going to get away from tools like that. But in terms of like questions like about
sleep and all kinds of things. I actually think that's where there are tools now that are being developed in various stages that can basically like, you know, like watch as we're talking now and give us each a PHQ -9 score based on a lot of attributes. And so I think a really interesting use case of this actually is involving Tardive dyskinesia. So with Tardive dyskinesia, it's a movement disorder. And I think we're all pretty good at detecting a very extreme.
case and you might see those folks on the subway or whatever. But the more subtle cases, I think we miss. And I don't know about you, but the AMS exam is pretty long. It's pretty in -depth. I don't think we're doing that every time. Maybe some people are. Bless them. But it's just not very practical. And so there is now this AI study underway that I've been able to be a rater on. And actually, it seems like the AI is generally more cohesive.
than the three Raiders. Like one Raider might be an outlier and this thing is always sort of like right in the pack. And it's just looking for tardifugous kinesia based on like a normal visit. Like you're just talking with the patient, this thing is watching it and no aims, no aims is done and it's still able to basically track with the aims. I think that's like really cool and like really good because like when I was in training we didn't have
Arjun Nanda (34:22.64)
Yeah.
Carlene MacMillan (34:29.296)
any drugs for TD. It was just like, well, I guess you have TD now. Like, let's like, kind of like the boogeyman is here, we just, that's too bad. Now, we have several different medications that actually can treat TD. And I think also, you know, the atypical antipsychotics, I kind of was trained like, there's such a low risk. It's not as low as we might think. And there's much more subtle things that can happen. And so,
Arjun Nanda (34:37.896)
Yeah.
Arjun Nanda (34:54.408)
Right.
Carlene MacMillan (34:59.216)
I think it will be a good public service when tools like that are available to at least lead to a further conversation. I don't want them to replace the clinical judgment, but I think some of these things, they're just going to be better than humans. They're just going to be able to detect subtleties that...
Arjun Nanda (35:11.4)
Mm -hmm.
Carlene MacMillan (35:17.326)
you know, we, we can't like with the naked eye detect, you know. So I'm very excited about those type of applications. And I know even, you know, the folks at Medicare that decide what's ratings for value based and all this kind of stuff. They're interested in that too. They're open. Like, I was at a thing about rating scales with CMS, the Centers for Medicare and Medicaid, and they had a lunch, a special lunch you could go to with the head of that, who decides about all the measures. And it was on AI. And it turns out she's done a bunch of AI stuff in campus.
Arjun Nanda (35:21.448)
Right.
Carlene MacMillan (35:47.232)
detection and she's super excited about getting these types of modern tools you know into the into the repertoire so I yeah I think it's that's so exciting and practical like yeah.
Arjun Nanda (35:48.104)
Hmm.
Arjun Nanda (35:59.848)
Right.
So the human eye, the clinician eye can look for some of these larger scale changes when they are happening. I mean, of course, certainly with movement disorders, but even with regards to depression or anxiety symptoms, generally there's a certain degree of change that needs to happen for it to be picked up by clinician. But what you're saying is that AI...
or other machine learning is able to pick up on these subtleties that we might not be able to get at. That's a genius.
Carlene MacMillan (36:38.638)
That's right.
Yeah, and I think also like one of the scribes now maybe some others they have a thing where they will provide data on the rate of speech, the amount of speech, the percentage of talking the volume, all that kind of stuff. They're not yet doing video, but they provide this data where they'll actually show you like, is like how does this compare to a patient's normal rate of speech. And it doesn't suggest like, maybe this patient is more depressed or maybe that they're manic that they'll
leaving that up to the clinician, but I think that can be really useful. And in particular, when people are not having the best insight into the fact that they might be having a mood episode in either direction, that you can actually say to them, instead of just, I notice you're talking faster. No, I'm not, I'm not talking. Well, okay, it's your word against mine. Or we can see that actually your average is this, and today it is higher. And I think that's gonna be helpful.
Arjun Nanda (37:27.688)
Yeah.
Yeah. Yeah.
Carlene MacMillan (37:40.112)
to help all of us be a little bit more objective around this stuff. And yeah, I think I just very excited that we will get there. We need to do some basic things first. And I do worry there's sometimes technology companies that, you know, excited about all these bells and whistles, but like, they also then just can't like, you know, submit a bill or, you know, do do basic stuff, we need to figure out like prior authorization, you know, like,
we can have all these tools, but if prior authorization is still a pain, then that's going to get our attention, not the fancy tool.
Arjun Nanda (38:17.768)
Right. Which is a perfect role for someone like yourself who can bridge that gap. And maybe we can talk about that. I mean, how you're in a very unique position and working for a technology company as a child psychiatrist. Can I ask how that happened? How did you get into that position?
Carlene MacMillan (38:26.51)
I mean, yeah, I hope so.
Carlene MacMillan (38:45.806)
Hmm.
Arjun Nanda (38:47.88)
And yeah.
Carlene MacMillan (38:49.326)
Yeah, so I think my practice ever since day one was kind of a technology forward practice where I was always very interested in the tools that we could use and how can you do HIPAA compliant messaging. And I wanted a paperless practice from day one. I just.
I did paper. And so I ended up through that and that was pretty early on, like 2012, 2013, all that. And so at that time there weren't as many companies in this space doing this. And so I was able to find these companies and work with them kind of early on.
You know, usually at that time was unpaid consulting. I didn't I just really know it's just like want to help them build the thing I need and so things like spruce health, which is great for messaging and phone like I was there very early days and Helping them say like hey, this is what I need as a psychiatrist. These are safety concerns This is what actually this feature would be really super helpful for my practice And so I got to just kind of do that. Whoops my my headphone there I got to do that in like informal
and sometimes even formally, like they would bring me out and have meetings and stuff. And that was very rewarding, but the end of the day it was still like running the practice, you know? And once we had a practice that grew, large group practice, we had a lot of people working with us, it took a lot of bandwidth to manage that. And honestly that just got a little old that wasn't as exciting, because it was basically like HR at that point, you know, you're just like hiring and.
worrying about all those sorts of things or that the air conditioner broke or you know just all kinds of like the toilet doesn't work. It's not that's not my thing. It's somebody else's thing. And I became increasingly frustrated that our practice was using you know ten different tools, technology tools that were all point solutions that none of them talked to each other. My staff was double entering demographics here and there and it just was like madness. It was just not productive.
Arjun Nanda (40:37.416)
Yeah.
Carlene MacMillan (40:58.544)
but yet we still needed these tools. Like particularly like say Spravato, a lot of workflows, there's the REMS program and our practice was the first in New York in 2019 to offer Spravato. So we were just like pioneering like how do we document this? And it was a huge thing to have all these different systems and so I hated that. And then because I needed that measurement -based care solution though that was so core.
Arjun Nanda (41:02.472)
Hmm.
Arjun Nanda (41:13.32)
Hmm.
Carlene MacMillan (41:25.838)
I ended up just meeting with a bunch of companies just as a prospective customer. And I will, you know, I remember meeting the founders, Jimmy and Lucia on Christmas Eve, actually. And they were very early and I think they were only around for a year at that point. And we both joked how it was like Christmas Eve and like here we were working and talking about measure of base care. Like this is our, we like this. And it was a very exciting conversation. And although that time, like that EHR was just too, too
Arjun Nanda (41:36.52)
Mm -hmm.
Carlene MacMillan (41:55.792)
in its early days to support a practice like ours. They invited me to be a clinical advisor, which was like a couple, you know, hour to a month. Give you a bit of equity, which I didn't really know what that meant. And then, you know, several months later, they said, hey, we, you know, we want to pick your brain about, you know, some roles coming up. And basically then they asked, would you ever think about like not doing your practice and coming to work?
for us. And it was first it was like, okay, I can do a couple hours of consulting. And then before you know it, like within a month or two, it was just like, okay, I'm going to sell my practice and I'm going to work here full time. Because this is where I can have that impact. This is where I basically got to sort of write my job description. And, you know, they didn't know what it would be like to have me there where I could add value. But it turned out that having a strong clinical perspective, and we also have another psychiatrist there who
Arjun Nanda (42:35.848)
Hahaha.
Carlene MacMillan (42:55.584)
who's great and more involved in the Ketamine space. Having us both there is just like decisions get made very differently. Because if you're not of this world, if you're not of our culture, psychiatric medicine, then things that they think will be useful are not useful or they haven't thought of the devil's in the details. So I love being able to learn from them and.
me and it's been really cool and I still definitely see patients we do use Ozmine they call that dog fooding when you use your own product tested out and I love still seeing the patients they inform the needs too and they'll complain you know hey this thing sent me a notification I didn't like or yes I love to get that.
Arjun Nanda (43:33.04)
Yeah.
Carlene MacMillan (43:47.566)
But it's been, yeah, it's a really cool experience. And I think that other colleagues like want to get into this space and certainly with COVID -19 and all the funding that was thrown at anything digital mental health.
There's all these opportunities that they just need a psychiatrist and people would ask me like, is this a good opportunity or not? And like there are ways that you can assess. And one of them is if they just basically want a figurehead or they just want someone to sign off on owning the PC and they just, they don't really care about your expertise. They just need that MD degree and they're not really going to have you involved. Those I don't, this is not a valuable.
Arjun Nanda (44:26.888)
Mm -hmm.
Carlene MacMillan (44:31.2)
thing to do usually and that's going to tell me that they don't really value your expertise versus the ones that...
Arjun Nanda (44:33.256)
bright.
Carlene MacMillan (44:38.638)
really care and are gonna kind of put their money where their mouth is and support the physician coming on board and getting super involved. That's that's what you want. Like the company doesn't have to be founded by a physician. But if they think the physician is like window dressing, that's going to be a huge, huge red flag versus like, hey, I admire physicians, I know we need them to help us and they want to partner. That's like, that's what you want to look
look for, for sure. And those companies definitely exist and or just consulting gigs, things like that. People can like I'm not technical, like I don't know how to program things, but you can still add tremendous value with the experiences and just being I think, as psychiatrists, you know, we're good communicators, we're used to managing conflict and competing demands and ambiguity, uncertainty and all the
Arjun Nanda (45:10.28)
Right.
Carlene MacMillan (45:38.512)
of those things when you're in a startup, in a tech startup, those are important things to do. So when people think they wouldn't be good at it or whatever, they probably would be actually. It's a good thing. Whereas I think maybe some other medical specialties maybe aren't quite as a natural fit. Yeah.
Arjun Nanda (45:49.224)
Right.
Arjun Nanda (45:57.864)
Yeah. So being an expert in your own field is of great value, but also finding the company that will actually value your expertise is just as important. Hello.
Carlene MacMillan (46:13.166)
Yes, and also like I think I'm just the one person. So I have to keep like my views might not be the same as your views. And so they also empower myself and Dr. McGinnis to, we formed a clinician advisory board of our colleagues. And so we bring those voices in and we basically serve as that, you know, kind of connective tissue between our world and the tech world and bringing people to the table. And I think that's been great for clinicians that are not, you know,
know, in the tech world, but still have like great perspectives to share. So that's also kind of cool. Yeah.
Arjun Nanda (46:49.768)
Right. That's awesome. Can I ask just to wrap it up and also get your thoughts on if you were to advise medical students or residents that are interested in getting into this space of technology plus mental health, what sort of advice would you give them?
Carlene MacMillan (47:17.166)
Yes, that's a great question. So actually, the American Psychiatric Association annual meeting, you know, it's quite large, but in the exhibit hall for the past several years, they've had something called the Innovation Zone. And they have their they have a pitch competition that residents can participate in. They have great panels of clinicians that actually work, you know, in industry in some capacity, sort of that meld
of tech and psychiatry. And so if you're at all interested in that space, like go to that and forget about the rest of the conference, just sit there, you know, just go to those talks. And I think meet people and get involved in that way is like a key way. And I haven't seen that necessarily like the child psychiatry meeting, for example, we don't really have that there, but APA does. And then I think, you know, if you're interested in
specific committees, even with ACAP, there's a health IT technology committee that people can be on and meet others in this world. You can get involved as a resident through that committee work. I joined a committee at ACAP when I was a resident, and now I'm the co -chair. It's been a part of my developmental life as a CHAD psychiatrist. I think those are so good. If your program doesn't have what you're looking for,
you can meet people through these professional societies that do and are eager to either have you do an elective or a project. And you can basically kind of choose your own adventure beyond just the walls of your program. And I think that's one of the people wonder what the value of the membership is. Like that's a huge part of the why we pay our dues, I think.
Arjun Nanda (49:08.968)
Right. Well, thank you for the sage advice and wisdom. It was a real pleasure having you on this podcast and just getting to talk with you. And I also hear that you recently accepted a new position at the Clinical TMS Society Board, if I'm not mistaken. So I just wanted to congratulate you.
Carlene MacMillan (49:33.102)
Yes, I've been so yes. So so I'm actually I'm on the Board of Directors of the Clinical TMS Society been on it for about a year, but I actually just took another board position with the Foundation for the Advancement of Clinical TMS, which is a sister nonprofit devoted to like increasing access funding things. And so that's great, because it's just basically like spreading the good word about TMS. And so yeah, very, very excited about that. And I don't know why I just love being
in these professional societies, I've just found that they've been so grounding to me and so great for decisions throughout my career. So, yeah, it's exciting. Yeah. Thank you.
Arjun Nanda (50:13.608)
Amazing. Well, thank you so much again. It was a real pleasure and looking forward to seeing where things go in the next couple of years with Osmind and the rest of your career. Absolutely. Awesome.
Carlene MacMillan (50:26.318)
and the field, right? Where the field's going is exciting. So I'll go wherever the excitement is. All right, well have a good rest of your day and thanks so much for having me. All right, bye bye.
Arjun Nanda (50:36.58)
Yeah, you too.