Latest / Becker's Healthcare Behavioral Health / Turning Behavioral Health Data Into Actionable Evidence

Transcript

Turning Behavioral Health Data Into Actionable Evidence · Becker's Healthcare Behavioral Health

• 0:00
  - Hello, everyone. This is Erica Spicer Mason with
• 0:02
  - Becker's Healthcare. Thank you so much for tuning
• 0:05
  - in to the Becker's Healthcare podcast series.
• 0:08
  - Today, we're going to talk about Whole Musks
• 0:10
  - health systems and care networks partnership model.
• 0:14
  - And joining me for this conversation, we have
• 0:16
  - Holmusk's Alex Vance, who is the chief clinical
• 0:18
  - data officer.
• 0:20
  - Alex, it's so great to have you with
• 0:22
  - us, and I look forward to learning more
• 0:24
  - about you and Holmusk today. Yeah. Glad to
• 0:26
  - be here. Thanks. Yeah. So just to get
• 0:28
  - us started, would you mind sharing a little
• 0:30
  - bit more about yourself and the work that
• 0:32
  - you're leading at Hallmark? I'd love to know
• 0:34
  - what problem you're ultimately trying to solve in
• 0:37
  - the behavioral health space.
• 0:39
  - Sure. And thanks again, Erica, for having me
• 0:41
  - on. I'm Alex. I'm
• 0:42
  - the chief clinical data officer at Hallmark.
• 0:46
  - My background is as a clinician, so I
• 0:48
  - practice as a mental health clinician before I
• 0:50
  - moved into health care data and real world
• 0:52
  - evidence.
• 0:53
  - And at Hallmark, I sit right in the
• 0:55
  - middle of clinical data
• 0:57
  - strategy, AI, and research,
• 1:00
  - and all that is in collaboration with our
• 1:02
  - health systems.
• 1:03
  - The problem we're ultimately trying to solve, is
• 1:05
  - that behavioral health generates an enormous amount of
• 1:08
  - clinical clinically meaningful information
• 1:11
  - through re routine care. But in the past,
• 1:14
  - we haven't been very good at turning that
• 1:15
  - information into evidence. And that's particularly true in
• 1:18
  - neuropsychiatric and neurodegenerative
• 1:21
  - disease where so much of what matters clinically
• 1:23
  - isn't captured neatly in a lab value or
• 1:25
  - a structured field.
• 1:27
  - So no one's clicking a button on the
• 1:28
  - other end to tell you what's going on
• 1:30
  - always. It's usually buried in clinical notes. It's
• 1:33
  - in descriptions of symptoms and cognition,
• 1:36
  - notes about functioning, treatment response, social context, and
• 1:40
  - clinician judgment. It makes the data incredibly rich
• 1:43
  - clinically, but it makes it really, really difficult
• 1:45
  - to study at scale.
• 1:47
  - So at the center of what Hallmark has
• 1:49
  - built is what we call NeuroBlue, and that's
• 1:50
  - just our real world clinical database.
• 1:53
  - It's really the foundation on top of which
• 1:55
  - all of our products and services are built,
• 1:58
  - and we bring together,
• 2:01
  - deeply characterized,
• 2:02
  - de identified
• 2:03
  - longitudinal data across a very large patient population.
• 2:07
  - And so we can start to answer some
• 2:08
  - of these questions that are extremely difficult to
• 2:10
  - answer from any individual dataset or health system
• 2:13
  - alone.
• 2:14
  - And so our goal really is to close
• 2:16
  - that gap that I talked about to take
• 2:18
  - the information that's already being generated through routine
• 2:21
  - cares. We're not asking the health system to
• 2:23
  - change its practices,
• 2:24
  - and we responsibly turn that into evidence that
• 2:27
  - helps us better understand patients,
• 2:29
  - disease progression, treatment patterns,
• 2:31
  - ultimately, what treatment works,
• 2:34
  - for whom, and under what circumstances.
• 2:36
  - And importantly, we want that process to benefit
• 2:38
  - not just the researchers or life sciences companies,
• 2:41
  - but also the health systems and clinicians who
• 2:43
  - generate that information and ultimately the patients who
• 2:47
  - receive care.
• 2:48
  - Alex, it's great to learn more about you
• 2:51
  - and also what what HomeSke is doing in
• 2:53
  - this space.
• 2:54
  - You know, you mentioned that health systems are
• 2:56
  - generating this incredibly rich,
• 2:59
  - valuable
• 3:00
  - behavioral health data. But as you noted, it's
• 3:03
  - been challenging for organizations
• 3:05
  - to really
• 3:06
  - use it to its full potential.
• 3:08
  - So I wanna talk a little bit about
• 3:10
  - specifically the approach Wholesk is taking here. It
• 3:13
  - it sounds like it's largely building around a
• 3:15
  - partnership approach rather than just treating
• 3:19
  - this health care data as a commodity.
• 3:21
  - So can you say a little bit more
• 3:23
  - about what this model looks like and what
• 3:25
  - makes it different than what other organizations are
• 3:27
  - doing?
• 3:28
  - Yeah. Absolutely.
• 3:29
  - So I think it's it's really important it's
• 3:32
  - a really important distinction between, what I call
• 3:34
  - a transaction and a partnership, and that that's
• 3:37
  - really important, in understanding our model. There's a
• 3:39
  - really common model in health care data that
• 3:42
  - is fair fairly transactional.
• 3:44
  - It works something like a health system has
• 3:46
  - data, Somebody wants access to that data, and
• 3:49
  - that's where the relationship ends. It usually involves
• 3:52
  - the exchange of cash or revenue share or
• 3:54
  - something that's just a one time, you know,
• 3:56
  - you give me this. I'll take that. And
• 3:58
  - and then both entities move on. That, as
• 4:01
  - you can imagine, is not really sustainable, and,
• 4:03
  - it it sort of falls apart as long
• 4:05
  - as soon as one person,
• 4:07
  - isn't able to keep that relationship up. So
• 4:10
  - when we partner with a health system, we're
• 4:12
  - thinking about a long term relationship and what
• 4:15
  - we can build together.
• 4:17
  - Our partners are contributing to the continued develop
• 4:20
  - of the Neuro Blue database, and Hallmark brings
• 4:23
  - infrastructure technology,
• 4:24
  - clinical expertise,
• 4:26
  - data science, connections to life science companies, and
• 4:29
  - research capabilities that are needed to transform
• 4:32
  - complex clinical data into something that can support
• 4:35
  - rigorous research real world evidence.
• 4:37
  - But the health system in this relationship is
• 4:40
  - not, a sim simply just a source of
• 4:42
  - data. They're a partner in the broader research
• 4:44
  - ecosystem.
• 4:45
  - So for us, that means asking
• 4:47
  - what questions are important to you, how is
• 4:50
  - health care delivered, are you a very involved
• 4:53
  - clinical research institute, or are you more concerned
• 4:55
  - with,
• 4:56
  - squeezing out value of every grant funded dollar
• 4:59
  - that your, you know, services are depending on?
• 5:02
  - And so what we can help them understand
• 5:04
  - about their own patient population really matters, what
• 5:08
  - research opportunities we can pursue together, and how
• 5:10
  - we can use the capabilities that we're building
• 5:12
  - to generate insights that are useful to clinicians,
• 5:15
  - researchers, and health system leadership.
• 5:18
  - For me,
• 5:19
  - personally, the underlying principle is that this needs
• 5:22
  - to be a, a mutual partnership in which
• 5:24
  - we're benefiting
• 5:26
  - and the health system is benefiting, and and
• 5:28
  - that benefit is not just theoretical.
• 5:30
  - And so we ask our health systems to
• 5:33
  - continuously give us feedback.
• 5:35
  - And I think the proof in in our
• 5:37
  - value that we provide our health systems is
• 5:39
  - that, they come back to us year after
• 5:41
  - year and keep the partnership up.
• 5:44
  - What a
• 5:45
  - interesting approach, Alex. You know, it it's fascinating
• 5:48
  - to me that we can be talking about
• 5:51
  - this high, high volume of health care data
• 5:53
  - that that you're helping organizations
• 5:55
  - essentially use
• 5:57
  - and leverage, but you're also able to take
• 6:00
  - this kind of individualized
• 6:02
  - approach to that partnership. So,
• 6:05
  - I wanna explore that just a little bit
• 6:06
  - further.
• 6:07
  - I understand Whole Musks partnership network now represents
• 6:10
  - more than 47,000,000
• 6:12
  - patients.
• 6:14
  - But its sustainability
• 6:15
  - does depend on health systems continuing to see,
• 6:18
  - I'm sure, meaningful value from participating in that
• 6:21
  - network. So
• 6:22
  - how does your organization
• 6:23
  - provide back to its partners, and can you
• 6:25
  - say a little bit about how you make
• 6:27
  - sure that the value exchange remains actually
• 6:31
  - meaningful and mutual?
• 6:33
  - Yeah. I'd be happy to go into that.
• 6:35
  - So you mentioned forty seven million patients. We're
• 6:38
  - extremely proud of that scale, and that scale
• 6:40
  - is important.
• 6:41
  - You know, our network now covers forty seven
• 6:44
  - million patients who have a diagnosis of a
• 6:46
  - mental health disorder or a neurological disorder and
• 6:49
  - have received care somewhere
• 6:51
  - in The US for that. But that scale
• 6:53
  - by itself is not our objective. The real
• 6:55
  - question is what becomes possible with that scale
• 6:57
  - and whether the organizations who are contributing to
• 7:00
  - that network
• 7:01
  - continue to see meaningful benefit in what they're
• 7:04
  - doing from participating with Hallmark.
• 7:06
  - So I tend to take, to think about
• 7:08
  - value that we provide back to our health
• 7:10
  - systems in two broad categories, and I'll call
• 7:12
  - them research and insight.
• 7:15
  - On the research side, I mean, research in
• 7:17
  - the broadest possible sense. So that can be
• 7:20
  - traditional academic research and publications through NIH or
• 7:23
  - NIMH grant funding.
• 7:25
  - It can be program evaluation,
• 7:27
  - understanding whether a particular intervention
• 7:29
  - or model of care is actually working, what
• 7:32
  - your system looks like before you take that
• 7:34
  - action, and then after it's implemented, what happens.
• 7:37
  - It could be operational research around things like
• 7:39
  - utilization or access
• 7:41
  - or cost, quality, or variation in care. And
• 7:44
  - then there's the entire spectrum of life sciences
• 7:46
  - research from drug development through post marketing surveillance,
• 7:50
  - natural history, and unmet need, and then a
• 7:53
  - number of other things. So when we say
• 7:55
  - research, we're really talking about using the clinical
• 7:57
  - data generated through routine care to answer a
• 8:01
  - very broad range of questions
• 8:03
  - in a way that you can trust the
• 8:05
  - underlying data.
• 8:06
  - When we talk about insight, we're talking about
• 8:08
  - something, altogether
• 8:10
  - different.
• 8:11
  - There, the goal is to make the information
• 8:13
  - useful,
• 8:14
  - to a much broader group of people inside
• 8:16
  - a health system.
• 8:17
  - Your average clinician, myself included,
• 8:20
  - don't usually know how to code, and they
• 8:21
  - certainly don't know how to code in really
• 8:23
  - complex languages.
• 8:25
  - Usually, our progress in working the data depends
• 8:28
  - on there being someone else in the health
• 8:29
  - system who does know how to do that
• 8:31
  - really well and usually more than one person
• 8:33
  - because as you can imagine, a health system
• 8:35
  - has a lot going on other than what
• 8:37
  - it's doing with Hallmark.
• 8:38
  - And so we wanna create,
• 8:40
  - an environment and have created that environment in,
• 8:43
  - our NeuroBlue platform where someone can very quickly
• 8:46
  - explore the things to matter to them. And
• 8:48
  - thanks to some of the recent capabilities of
• 8:51
  - LOMs,
• 8:52
  - you can do that, by writing an everyday
• 8:55
  - English, in the way that you understand your
• 8:57
  - environment, what's going on with your patients. And
• 9:00
  - over
• 9:01
  - time, that translates into a way to look
• 9:04
  - at our data,
• 9:05
  - and derive insights,
• 9:07
  - that you don't have to, you know, be
• 9:08
  - a a really rigorous research scientist. You don't
• 9:11
  - have to be a really rigorous coder,
• 9:13
  - and you can trust that everything that's going
• 9:14
  - on in the background is accurate both because
• 9:16
  - you can see it end to end and
• 9:19
  - because we check it. So anything our health
• 9:21
  - systems are doing,
• 9:23
  - we're verifying that it's, that it's doing what
• 9:25
  - they intend to do, and giving feedback. And
• 9:27
  - so that's again, just circling back to what
• 9:30
  - I mean when I say when you sign
• 9:31
  - up with almost you have a partnership. So
• 9:33
  - we're very in the weeds, with some of
• 9:35
  - our health system partners.
• 9:37
  - And importantly,
• 9:38
  - I don't wanna leave out, there's no cost
• 9:40
  - to the health system for these capabilities as
• 9:42
  - long as they remain a contributing partner. And,
• 9:45
  - one of the things that small stage startups
• 9:47
  - and psych departments have in common is they
• 9:49
  - usually don't come into a conversation with mountains
• 9:52
  - of cash. And so we've we try to
• 9:54
  - build something that is both valuable and sustainable
• 9:56
  - for these health systems.
• 9:58
  - Wow. So interesting. Alex, thanks for laying all
• 10:01
  - of that out. Clearly, a lot of value
• 10:03
  - that the health system partner is is receiving
• 10:06
  - on their end. And I'm curious,
• 10:08
  - as part of this partnership network, what does
• 10:10
  - then the health system,
• 10:12
  - have to bring to the partnership? What's required
• 10:15
  - of them?
• 10:16
  - It's a really important question because,
• 10:18
  - it it is a bit of work, as
• 10:20
  - you can imagine. We're not,
• 10:22
  - moving around things that, are very trivial. We're
• 10:25
  - we're talking about patient data,
• 10:27
  - and, that inevitably
• 10:29
  - involve involves some,
• 10:31
  - some access to PHI.
• 10:33
  - One of the things that we, pride ourselves
• 10:35
  - on is that everything begins with a conversation.
• 10:38
  - So we wanna understand the organization and why
• 10:40
  - they want to work with us, who are
• 10:42
  - the patients they serve, what's the environment,
• 10:45
  - that they work in, and what's their data
• 10:47
  - environment,
• 10:48
  - what are their priorities, and and what they
• 10:49
  - wanna get out of a partnership,
• 10:51
  - with us and in what period of time.
• 10:54
  - So from there, we work through, as you
• 10:55
  - can imagine, the appropriate governance and legal and
• 10:58
  - privacy and technical processes.
• 11:00
  - But there's an important distinction in how we
• 11:02
  - design the model that I think health system
• 11:04
  - leaders should really understand,
• 11:06
  - And that is
• 11:08
  - PHI or personal,
• 11:10
  - health information is never sent to Hallmark.
• 11:13
  - And so our technology does access PHI within
• 11:16
  - the health system to de identify that data.
• 11:19
  - But once that process has been implemented, no
• 11:21
  - deidentified data comes to Hallmark because we don't
• 11:24
  - need it to do what we do. And
• 11:26
  - that distinction for us is really important. We
• 11:28
  - designed it, specifically so that when you sign
• 11:31
  - up over a partnership with us, you don't
• 11:33
  - have to worry about someone hacking us and
• 11:36
  - breaching your data. And so we're very accustomed
• 11:40
  - to work with complex complex
• 11:42
  - EHR data, particularly in behavioral health,
• 11:45
  - which you can imagine is is quite sensitive.
• 11:47
  - And so our goal in that is to
• 11:49
  - do as much of the heavy lifting as
• 11:51
  - possible and to be as secure not only,
• 11:54
  - as what's required, but we go above and
• 11:57
  - beyond to only take in the data that
• 11:58
  - we need to do the work, that we're
• 12:00
  - trying to do within care delivery research, for
• 12:03
  - neuropsychiatry
• 12:04
  - and neurodegenerative
• 12:06
  - disease.
• 12:07
  - Yeah. Alex, thank you so much. And something
• 12:10
  - that's come up throughout our conversation is
• 12:13
  - essentially this idea of trust here. It seems
• 12:16
  - like that's what makes this partnership model sustainable.
• 12:20
  - Health systems have to contribute data. Pullmus has
• 12:23
  - to keep earning their participation
• 12:25
  - by creating some of the value in return
• 12:27
  - that that we spoke to earlier.
• 12:29
  - And, ultimately, that should translate into better research
• 12:32
  - and better outcomes for patients.
• 12:34
  - As the network grows, what becomes possible for
• 12:37
  - everyone who's involved, and how should a health
• 12:39
  - system leader who's interested in becoming part of
• 12:42
  - it get in touch with you?
• 12:44
  - Yeah.
• 12:45
  - So to get in touch with me really
• 12:47
  - simply,
• 12:48
  - my contact is alex@hallmusk.com.
• 12:51
  - All I need to know is how to
• 12:53
  - get in touch with you and a time
• 12:54
  - to meet, and we can take it from
• 12:56
  - there. And so you talked about our network
• 12:58
  - growing, you know, as more and more systems
• 13:00
  - participate in this, in this thing that we're
• 13:03
  - trying to do. And this is where I
• 13:05
  - think it gets really exciting because behavioral health
• 13:08
  - has historically faced some fundamental challenges when it
• 13:10
  - comes to to generating evidence, and that's everything
• 13:13
  - from fragmented data and,
• 13:15
  - lack of tax incentives and adopting electronic medical
• 13:18
  - record technology.
• 13:19
  - And like I mentioned, just enormous amounts of
• 13:22
  - of what's going on locked away in, in
• 13:25
  - notes buried,
• 13:26
  - buried in the system and just not possible
• 13:28
  - to look at with any kind of scale.
• 13:30
  - And so as our network grows, the kind
• 13:32
  - of questions we ask, become much more sophisticated.
• 13:36
  - We can study smaller and harder to reach
• 13:38
  - patient populations,
• 13:39
  - ones that would take years to find,
• 13:42
  - otherwise.
• 13:43
  - We can understand treatment sequences rather than just
• 13:45
  - individual one off treatments.
• 13:47
  - We can follow patient journeys over long periods
• 13:49
  - of time and and can examine differences not
• 13:52
  - only within a health system, but what is
• 13:54
  - going on in that health system and how
• 13:55
  - it looks in comparison to every other health
• 13:57
  - system who's participating with us.
• 13:59
  - In neuropsychiatric
• 14:01
  - and neurodegenerative
• 14:02
  - disease, in particular, we can increasingly use the
• 14:05
  - information in clinical notes to understand things like
• 14:07
  - symptomology
• 14:08
  - and functioning,
• 14:09
  - cognition, treatment response, and disease progression. Those things
• 14:13
  - can be really extremely difficult to capture when
• 14:15
  - you're using structured data alone.
• 14:18
  - And that creates value for the life sciences
• 14:21
  - companies that we work with and researchers certainly,
• 14:23
  - but it also creates a network effect for
• 14:25
  - participating health systems. So each organization becomes part
• 14:27
  - of a much larger evidence base and creates
• 14:30
  - opportunities to answer questions that wouldn't be possible
• 14:33
  - for any,
• 14:34
  - any one of those health systems on their
• 14:36
  - own or even with in collaboration with one
• 14:39
  - or two health systems.
• 14:40
  - So, ultimately, what we're trying to do and
• 14:42
  - that is shorten the distance between care and
• 14:45
  - evidence.
• 14:46
  - There's an extraordinary
• 14:48
  - amount of information that we can learn from
• 14:50
  - the experiences of our patients.
• 14:52
  - I think the opportunity is actually bidirectional. So
• 14:55
  - we can we can turn routine care into
• 14:57
  - better evidence, but we can also take what
• 14:59
  - we learn from that evidence and make it
• 15:01
  - much more accessible to the people making decisions
• 15:04
  - about care
• 15:05
  - every day.
• 15:06
  - And for me to me as a clinician,
• 15:08
  - that's where what I do becomes much more,
• 15:11
  - much more than just a database. It it
• 15:13
  - it's creating an ecosystem
• 15:15
  - where care generates evidence and evidence increasingly informs
• 15:18
  - care. So for the health system leader,
• 15:21
  - who's interested,
• 15:23
  - please do reach out. Like I said, it's
• 15:25
  - alex@holmusk.com.
• 15:27
  - I I respond really quickly, and we'd love
• 15:29
  - to talk to you.
• 15:31
  - Well, Alex, it's been so great learning about
• 15:34
  - you, what HomeWorks is delivering in this partnership
• 15:37
  - model and approach with health systems.
• 15:40
  - A lot for our listeners to think about
• 15:42
  - and appreciate you making it so easy and
• 15:44
  - accessible to reach you if they're interested.
• 15:47
  - Just wanna thank you again for your time
• 15:48
  - and the insights today.
• 15:50
  - Yeah. Thanks, Erica. This has been a pleasure.
• 15:52
  - And we also wanna thank our podcast sponsor,
• 15:55
  - Holmusk.
• 15:56
  - Listeners, be sure to tune in to more
• 15:58
  - podcasts from Becker's by visiting our podcast page
• 16:01
  - at beckershospitalreview.com.