Latest / Becker’s Healthcare Podcast / Why AI Alone Won’t Transform Healthcare

Transcript

Why AI Alone Won’t Transform Healthcare · Becker’s Healthcare Podcast

• 0:00
  - This is Scott Becker with the Beckers Healthcare
• 0:03
  - Podcast.
• 0:04
  - I'm thrilled today to be joined by one
• 0:06
  - of my favorite leaders, a brilliant, brilliant leader.
• 0:09
  - We're joined today by Seema Verma. Seema is
• 0:12
  - the executive vice president and general manager of
• 0:14
  - Oracle Health and Life Sciences. She's had a
• 0:17
  - brilliant career.
• 0:18
  - We're gonna ask Seema to talk about
• 0:21
  - what Oracle's focused on now, where they're going,
• 0:24
  - and a lot more.
• 0:25
  - Seema, can you take a moment for anybody
• 0:28
  - that's new to either yourself or to Oracle
• 0:30
  - Health? Tell us a little bit about yourself
• 0:32
  - and about Oracle Healthcare.
• 0:34
  - Sure. Well,
• 0:36
  - gosh. Talking about yourself is always a a
• 0:38
  - little bit, awkward. But,
• 0:41
  - so my, you know, my career journey has
• 0:43
  - been around spending a lot of time in
• 0:45
  - in health policy. And,
• 0:47
  - before coming to
• 0:49
  - Oracle, I was the administrator for CMS.
• 0:53
  - And so most, like I said, that that
• 0:55
  - was sort of the the pinnacle of a
• 0:57
  - lot of time on policy issues, working with
• 1:00
  - state governments, working with,
• 1:02
  - the federal government,
• 1:04
  - and also, you know, being on the front
• 1:06
  - lines working at a at a public hospital.
• 1:09
  - And so all of that I think all
• 1:11
  - of that experience
• 1:12
  - was about trying to solve the problems in
• 1:14
  - health care and then try to do that
• 1:16
  - at the highest level,
• 1:18
  - at CMS on a policy level. So, you
• 1:21
  - know, I'm taking this turn
• 1:23
  - into the IT world and tech world,
• 1:27
  - and I'm very hopeful that this is the
• 1:29
  - place where some of these long standing problems
• 1:32
  - that we've had in health care can can
• 1:34
  - get solved in a way that isn't quite
• 1:37
  - possible,
• 1:38
  - with government, with policy regulation,
• 1:41
  - and even with funding.
• 1:43
  - There are just some things that technology, I
• 1:45
  - think, is better suited to solve.
• 1:48
  - Thank you. And talk a little bit about
• 1:50
  - last year, we had a chance to talk
• 1:52
  - a lot about
• 1:53
  - Oracle's vision for next generation
• 1:56
  - EHR.
• 1:57
  - Since then, Oracle Health has announced the new
• 2:00
  - EHR is live for ambulatory surgical centers, ambulatory
• 2:03
  - organizations,
• 2:05
  - of all sorts.
• 2:07
  - Why was getting highly focused into ambulatory
• 2:10
  - important?
• 2:11
  - And tell us a little bit about why
• 2:13
  - the interoperable
• 2:14
  - foundation
• 2:15
  - will help health systems make progress with AI
• 2:18
  - and modernization. So the movement towards ambulatory,
• 2:21
  - then how will this impact health systems?
• 2:24
  - Sure. So, you know, I think we started
• 2:26
  - with ambulatory. I mean, if we think about
• 2:29
  - what's gonna be in the EHR, there's a
• 2:32
  - lot of different workflows. And with ambulatory,
• 2:35
  - there's a smaller number of workflows than we
• 2:38
  - have with the acute.
• 2:40
  - So a little bit less complex. And, essentially,
• 2:42
  - this is all gonna build up a lot
• 2:44
  - of the work that we're doing for the
• 2:45
  - ambulatory will also feed into the acute EHR.
• 2:49
  - But, you know, I think what's different here,
• 2:51
  - the approach that we're taking is very different.
• 2:54
  - What we're seeing in the market today
• 2:56
  - is older technology
• 2:58
  - and bolt on
• 3:00
  - AI solutions.
• 3:01
  - And the reason why I think that's problematic
• 3:04
  - is that when we think about AI and
• 3:06
  - how AI works,
• 3:08
  - AI works as well as the data that
• 3:10
  - it has available. So how the model was
• 3:13
  - trained, but it also needs access to that
• 3:16
  - real time data. And I think we at
• 3:17
  - this point, most of us have used AI
• 3:20
  - tools
• 3:20
  - that we understand. Right? It's only as good
• 3:22
  - as what you're telling it to do and
• 3:24
  - the information that it comes that that that
• 3:27
  - has available.
• 3:28
  - And so if we think about health care,
• 3:30
  - in health care,
• 3:32
  - a lot of the decision making is multifaceted.
• 3:35
  - Right? We're not making decisions off of one
• 3:35
  - area. So if we think
• 3:42
  - complete
• 3:43
  - health record? You may or may not. So,
• 3:45
  - again, the AI is only as good as
• 3:47
  - what information is right in front of it.
• 3:49
  - But with health care, the decision making goes
• 3:52
  - beyond that. Right? It's it's the patient's preferences.
• 3:55
  - It's it could be genomic information.
• 3:57
  - It could be what the insurance company is
• 4:00
  - going to pay for. So you need a
• 4:02
  - lot of information. You need access to that
• 4:05
  - information,
• 4:06
  - in a real time way for the
• 4:09
  - AI to be more than just a a
• 4:11
  - research tool. Right? I mean, for it to
• 4:13
  - be really effective and deliver and address some
• 4:15
  - of these long standing problems in health care,
• 4:18
  - you need to have access to real time
• 4:20
  - data. Older technology
• 4:22
  - that is built on older databases,
• 4:25
  - they're not built
• 4:26
  - to retrieve real time information. They are systems
• 4:29
  - of record.
• 4:30
  - They're not systems of retrieval.
• 4:32
  - And that's really the the difference here. What
• 4:35
  - Oracle's building is something that is built for
• 4:37
  - AI. Right? We're not bolting on solutions. We're
• 4:40
  - not retrofitting something that's older. We're saying this
• 4:42
  - is the future
• 4:44
  - of AI, and we're building our databases and
• 4:47
  - the underlying
• 4:48
  - infrastructure,
• 4:49
  - to be able to retrieve that real time
• 4:51
  - information. We are able to pull information from
• 4:54
  - different systems.
• 4:56
  - So it could be that, you know, supply
• 4:58
  - chain information or HR.
• 5:01
  - You know, you need to have a lot
• 5:02
  - of information, the ability to retrieve it in
• 5:04
  - real time and to be able to access
• 5:06
  - it from different sources. And so that's the
• 5:08
  - foundation
• 5:09
  - of the EHR that we're building.
• 5:12
  - It is live on ambulatory,
• 5:15
  - and we continue to take input from our
• 5:17
  - customers. And,
• 5:19
  - you know,
• 5:20
  - it's, it's exciting. And I think our customers
• 5:22
  - are really excited about what they're seeing because
• 5:25
  - they can see the potential
• 5:27
  - of a different way and to make their
• 5:29
  - jobs easier,
• 5:31
  - to make,
• 5:32
  - the ability to deliver high quality care,
• 5:36
  - a lot easier than it has been in
• 5:38
  - the past. So,
• 5:39
  - a lot of excitement, and, we are
• 5:42
  - also going to be talking about different other
• 5:44
  - modules. Right? So this is ambulatory. We'll be
• 5:47
  - talking a lot about oncology.
• 5:49
  - But let's talk about that now, oncology.
• 5:52
  - Because on oncology is an area
• 5:55
  - where data, research, and care delivery have to
• 5:57
  - come together quickly.
• 5:59
  - In fact, we see so many places in
• 6:01
  - in the situation currently
• 6:02
  - where oncologists almost seem like they're overwhelmed because
• 6:06
  - there's only so many oncologists versus 2,000,000 new
• 6:09
  - diagnosis a year.
• 6:10
  - How can Oracle or or technology or oncology
• 6:14
  - providers
• 6:15
  - how can better connected data and intelligence
• 6:18
  - help that work and support more informed timely
• 6:21
  - oncology care?
• 6:23
  - Well, that's the foundation of it. And if
• 6:25
  - we think about the data that we're trying
• 6:27
  - to bring together in oncology, it is a
• 6:29
  - lot more complex. There's a lot more pieces
• 6:31
  - to it. So we can start with genomic
• 6:34
  - data. We're looking at tumors. There's a lot
• 6:36
  - of that understanding the
• 6:38
  - genetic makeup of those tumors, and that really
• 6:41
  - does inform the treatment.
• 6:43
  - So there's genomic data. There's imaging data.
• 6:46
  - There is
• 6:47
  - laboratory data. There's surgical data. There's all of
• 6:50
  - these different pieces. And, you know, what we
• 6:53
  - see a lot of times is even if
• 6:54
  - some of the information is available, even some
• 6:56
  - of that genomic data,
• 6:58
  - is often in in the form of a
• 7:00
  - PDF. Right? So it's not as easy to
• 7:03
  - analyze.
• 7:04
  - And then I think with oncology, the other
• 7:06
  - big piece of it is clinical trials,
• 7:08
  - understanding what clinical trials are available,
• 7:12
  - where a patient can be enrolled, and understanding
• 7:15
  - that all at once. And so there's really
• 7:18
  - no other area or no other area of
• 7:20
  - medicine that combines all of those different pieces,
• 7:23
  - and it's very much multispecialty.
• 7:25
  - Right? There's gonna be an oncologist. There's a
• 7:28
  - surgeon.
• 7:29
  - There's,
• 7:30
  - a lot of different providers that are that
• 7:33
  - are involved in one's care when it comes
• 7:35
  - to oncology.
• 7:36
  - So that's why we're starting there. You know,
• 7:38
  - even things like just facilitating
• 7:40
  - tumor board, right, where there's a lot of
• 7:42
  - different providers
• 7:44
  - weighing in on a patient's care.
• 7:46
  - Even those things are difficult because of the
• 7:49
  - information sharing. So part of it is information
• 7:51
  - sharing,
• 7:52
  - but a lot of it is also just
• 7:54
  - the complexity,
• 7:55
  - of the disease itself.
• 7:58
  - Thank you very, very much. You've also, Seema,
• 8:01
  - aside from oncology, ambulatory,
• 8:04
  - we've got a couple more subjects I'd love
• 8:06
  - to discuss. Wanna talk a little bit about
• 8:07
  - life sciences, a little bit about revenue cycle,
• 8:10
  - and also about trustworthy
• 8:11
  - AI in health care.
• 8:13
  - So so you've recently wrote an article about
• 8:15
  - having trustworthy AI in health care.
• 8:18
  - As AI gets deeper and deeper into systems,
• 8:20
  - clinical administrative
• 8:22
  - workflows both,
• 8:24
  - how is Oracle Health approaching
• 8:26
  - human oversight, human in the loop, transparency, privacy,
• 8:29
  - and accountability?
• 8:30
  - And what should health systems think about when
• 8:32
  - they think about the protection of everything and
• 8:34
  - just the the concerns about AI moving on
• 8:36
  - its own?
• 8:37
  - What should health systems expect from technology providers?
• 8:41
  - Sure. Well, this is a really important area
• 8:44
  - because especially when it comes to you know,
• 8:46
  - we're talking about patient care. So it's gotta
• 8:48
  - you know, safety has to be at the
• 8:50
  - first and foremost.
• 8:51
  - And we wanted to put out this information
• 8:53
  - because we wanted under our customers
• 8:55
  - to understand
• 8:56
  - how we are developing it and all the
• 8:58
  - steps that we're taking. And moreover, I hope
• 9:00
  - that it also serves as a standard
• 9:02
  - and that other tech companies out there also
• 9:05
  - have an opportunity to understand how we're doing
• 9:07
  - it and that we're we're setting a bar
• 9:10
  - for the entire tech community. But, you know,
• 9:12
  - it starts with human centered design. Right? So
• 9:15
  - that means that humans are always in the
• 9:17
  - loop, that the AI is not acting alone.
• 9:21
  - It's making suggestions, but the the decision making
• 9:24
  - is always,
• 9:26
  - something that the human
• 9:29
  - has to have oversight. The other areas around
• 9:30
  - transparency
• 9:31
  - and explainability. It's not just saying, you know,
• 9:34
  - here's the conclusion.
• 9:35
  - Right? This patient has x or y or
• 9:37
  - z.
• 9:38
  - It's actually telling you why and where that
• 9:38
  - information came from, you know,
• 9:39
  - why and where that information came from. You
• 9:42
  - know? We're saying don't use this drug because
• 9:46
  - the patient has
• 9:47
  - an an allergy. Okay. Well, I didn't know
• 9:49
  - my patient had that allergy, and I don't
• 9:51
  - think that's accurate. We literally footnote everything so
• 9:54
  - you can click on it and go back
• 9:56
  - to the source. Or if it's an external
• 9:59
  - data source, so a certain piece of literature,
• 10:02
  - it's gonna tell you exactly where that came
• 10:05
  - from. So it's explaining
• 10:07
  - why it's coming to that conclusion as well.
• 10:10
  - So I think of it as, remember in
• 10:11
  - math class where it was, show your work?
• 10:14
  - You don't just put the answer on the
• 10:16
  - paper and, you know, you get you get
• 10:17
  - some points taken off if you don't show
• 10:19
  - every step. It's kinda like that with, our
• 10:22
  - approach in AI. So it really tells you
• 10:24
  - that. The other thing I think is that
• 10:27
  - we can't assume
• 10:28
  - that you put it out there. And just
• 10:30
  - like all a lot of technology that's out
• 10:33
  - there, that has to be continuously
• 10:35
  - evaluated and validated
• 10:37
  - to make sure that it's accurate and reliable.
• 10:39
  - So it doesn't stop with, you know, you
• 10:42
  - you kinda put it out there. You're constantly
• 10:44
  - monitoring that. So it's pre deployment validation and
• 10:48
  - then ongoing monitoring.
• 10:50
  - The other big huge piece is cybersecurity and
• 10:53
  - security and privacy.
• 10:55
  - You know, with AI, you're moving data, and
• 10:58
  - we've gotta make sure that at every step
• 11:01
  - that the security and privacy is there. So
• 11:05
  - we have to really think about the privacy
• 11:07
  - by design principles and built in those safeguards
• 11:11
  - into the AI systems at the very, very
• 11:14
  - beginning.
• 11:15
  - Whereas if you're bolting on,
• 11:17
  - it's not gonna be as secure. I don't
• 11:19
  - think safety is a one and done. We
• 11:21
  - continue to monitor and continue to find ways
• 11:23
  - to improve it.
• 11:25
  - Thank you very, very much. And and, generally,
• 11:28
  - people wanna work with
• 11:31
  - companies that have deep resources that could address
• 11:33
  - these issues that aren't just trying to get
• 11:35
  - in the door, but also have the support
• 11:38
  - to back up the privacy and the trustworthy
• 11:40
  - issues and and security issues that are so
• 11:43
  - important today.
• 11:45
  - Sima, let me ask you about revenue cycle.
• 11:48
  - Health systems are obviously constantly trying to work
• 11:50
  - on revenue cycle. I think there's the latest
• 11:53
  - said I saw was 2 to 300,000,000,000 a
• 11:54
  - year spent on revenue cycle.
• 11:57
  - What where is the biggest opportunity
• 11:59
  - near term and longer term
• 12:01
  - that health system providers can use to improve
• 12:04
  - things for revenue cycle?
• 12:06
  - What what we're seeing in the market is
• 12:08
  - a lot of point solutions, which I think
• 12:09
  - are helpful for
• 12:12
  - addressing single tasks. But when we think about
• 12:15
  - revenue cycle and how it works, there's the
• 12:17
  - front office, the middle office, the back office.
• 12:19
  - Right? And all of those pieces have to
• 12:21
  - come together.
• 12:22
  - And we spend a lot of money because
• 12:24
  - there's a lot of coordination that has to
• 12:26
  - happen between the front, middle, and back office.
• 12:29
  - With AI,
• 12:31
  - there's an opportunity to bring all of that
• 12:33
  - data together.
• 12:35
  - I do worry that we're gonna be spending
• 12:37
  - a lot that that there's a, potential
• 12:39
  - for providers to spend a lot of money
• 12:42
  - on point solutions.
• 12:43
  - But, ultimately,
• 12:44
  - it all has to start with common data
• 12:47
  - and to be able so let's think about
• 12:49
  - a denied claim. Well, maybe it was denied
• 12:51
  - because the prior authorization
• 12:53
  - wasn't there. Well, in order to do the
• 12:55
  - prior authorization, you need to be able to
• 12:57
  - get into the clinical record and have the
• 12:59
  - right information
• 13:00
  - at the right time.
• 13:02
  - So it it starts with being able to
• 13:05
  - bring together a foundation of all of that
• 13:07
  - data and all of those systems together so
• 13:10
  - that the AI can work across it and
• 13:12
  - it isn't just, you know, focusing on one
• 13:15
  - particular solution. So think about coding.
• 13:17
  - So maybe you've got an AI tool that
• 13:19
  - helps you with coding, but that's only part
• 13:21
  - of the revenue cycle. Right? You gotta code
• 13:23
  - it. You gotta submit the claim. If there's
• 13:25
  - a denial, you need to be able to
• 13:27
  - respond to it, you need to be able
• 13:28
  - to go back to the clinical record and
• 13:30
  - pull the relevant information,
• 13:33
  - to justify the claim.
• 13:35
  - So, again, it starts with that common data
• 13:37
  - foundation and the AI being able to work
• 13:40
  - across the system and not just in the
• 13:42
  - EHR.
• 13:43
  - There's some clinical data, but there's also financial
• 13:46
  - data. And then I would say the last
• 13:48
  - piece of it is ERP. Right? Ultimately, this
• 13:51
  - is all financial information that goes into an
• 13:54
  - ERP,
• 13:55
  - and you wanna be able to make sure
• 13:57
  - that that whole system can work together.
• 14:01
  - Thank you. Talk also,
• 14:04
  - Oracle Health and AI
• 14:06
  - is is
• 14:08
  - bigger and bigger in life sciences overall,
• 14:11
  - in research and bringing new therapies to patients.
• 14:15
  - Where do you see the greatest opportunities
• 14:17
  - in sort of the life sciences areas?
• 14:20
  - Sure. And that's a, you know, that's a
• 14:22
  - great point. And the way Oracle is coming
• 14:24
  - at health care is not just with one
• 14:27
  - point solution on the EHR. Right? It's a
• 14:30
  - it's a much more it's a much more
• 14:32
  - larger play,
• 14:33
  - and it's focused on targeting
• 14:36
  - and transforming
• 14:37
  - the entire industry.
• 14:39
  - Because I think there's a realization, you can't
• 14:41
  - transform one part of it and think that
• 14:43
  - you're gonna do all of it. It's not
• 14:45
  - gonna work that way. And that's why, you
• 14:47
  - know, a point solution, an EHR, it it's
• 14:51
  - not going to get at it. And the
• 14:52
  - the clinical trials world and life sciences is
• 14:55
  - is very, important to health care.
• 14:58
  - We think about all the new treatments
• 15:00
  - that are out there.
• 15:01
  - We gotta make sure that providers on the
• 15:03
  - frontline
• 15:04
  - know about those treatments and therapies.
• 15:07
  - And also,
• 15:08
  - in order to accelerate innovation,
• 15:11
  - we need to make sure that people are
• 15:13
  - aware of the clinical trials that are out
• 15:15
  - there. And that's that's a, you know, that's
• 15:17
  - an area that we're very weak on. One
• 15:18
  - to two percent of the population actually participate
• 15:21
  - in clinical trials. And usually, those are the
• 15:24
  - ones that are coming out of the larger
• 15:25
  - academic health centers.
• 15:27
  - And so there's an opportunity to really accelerate
• 15:30
  - research, cures, and treatments if we have more
• 15:33
  - real time information.
• 15:35
  - And it's bidirectional.
• 15:36
  - I think for for pharmaceutical companies and med
• 15:39
  - tech companies, they need to be able to,
• 15:42
  - enroll patients into trials. They also need to
• 15:44
  - be able to have access to data to
• 15:46
  - understand
• 15:47
  - the impact of the therapeutic intervention. So whether
• 15:50
  - it's a drug or a device,
• 15:52
  - what happened five years later?
• 15:55
  - I think there's a lot of information out
• 15:57
  - there that we are not able to,
• 16:00
  - to understand,
• 16:01
  - and there could be a lot of lessons
• 16:04
  - learned just from the data that we're not
• 16:05
  - using today. And what I mean by directional,
• 16:08
  - let's say there are new treatments and therapies.
• 16:11
  - The reality is,
• 16:12
  - a lot of doctors and nurses, there's so
• 16:15
  - much going on that they're not always aware
• 16:18
  - or have
• 16:19
  - access to the latest
• 16:21
  - treatments and therapies. And so that clinical decision
• 16:23
  - support becomes
• 16:25
  - really important and to be able to make
• 16:27
  - sure that providers have access to real time
• 16:29
  - data and information.
• 16:32
  - Thank you. Just a couple more questions, Seema,
• 16:34
  - and thank you so much for your time
• 16:35
  - today.
• 16:37
  - Oracle has its health and life sciences summit
• 16:40
  - coming up this year.
• 16:42
  - What should health system leaders
• 16:44
  - prioritize first
• 16:46
  - in in general as they tend to modernize
• 16:49
  - operations and adopt AI?
• 16:51
  - And then in terms of prioritizing,
• 16:53
  - then what do you hope they take away
• 16:55
  - from the Oracle Health and Life Sciences Summit?
• 16:57
  - So what should else assist us prioritize first?
• 17:00
  - And then what do you hope they take
• 17:01
  - away from the Life Sciences Summit?
• 17:03
  - You know, for our summit, first of all,
• 17:04
  - it's a global conference, so we have people
• 17:06
  - coming from all over the world to participate.
• 17:09
  - And we wanted the summit to be where
• 17:11
  - obviously, we're gonna talk about what Oracle is
• 17:14
  - doing and and profile some of our new
• 17:16
  - products, but it's also supposed to be an
• 17:18
  - opportunity
• 17:19
  - for sharing and around,
• 17:22
  - try to all work together around transformation. So
• 17:25
  - bringing some of the greatest minds together. We've
• 17:27
  - got a lot of different speakers that are
• 17:30
  - not talking necessarily about Oracle, but they're talking
• 17:32
  - about what's going on in the market. And
• 17:34
  - we hope that it's a learning opportunity for
• 17:37
  - the participants to learn from others, to understand
• 17:40
  - not only technology,
• 17:42
  - but regulation and policy
• 17:44
  - that's going on because that's gonna impact the
• 17:46
  - work that they do, and it impacts their
• 17:48
  - their decisions
• 17:49
  - around technology as well. So it's really intended
• 17:52
  - to be a learning experience for the participants.
• 17:55
  - The other thing I would say is when
• 17:57
  - it comes to what what do we want
• 17:59
  - health care leaders to take away?
• 18:01
  - And for me, the takeaway is to understand
• 18:04
  - an approach
• 18:05
  - to AI.
• 18:06
  - It's not just about one solution, a bolt
• 18:09
  - on solution, but it really is an enterprise
• 18:12
  - approach,
• 18:13
  - which is how do I use and utilize
• 18:16
  - AI?
• 18:17
  - And so it's not just a buying decision,
• 18:19
  - but it's also, okay, what about my what
• 18:21
  - about the impact of AI on my processes?
• 18:24
  - Right? I may need to do things differently
• 18:26
  - because the AI is able to do this
• 18:28
  - part of the job. We don't need to
• 18:29
  - do that anymore, but what do we need
• 18:31
  - to do? So there's technology.
• 18:33
  - There's the the process
• 18:35
  - transformation.
• 18:36
  - And then the third piece of it is
• 18:38
  - also from that buying decision, understanding that you
• 18:42
  - can't have an AI strategy unless you have
• 18:44
  - a data strategy. It has to start with
• 18:46
  - how do I bring my data together.
• 18:49
  - And so building that understanding of how AI
• 18:51
  - works, we hope that it'll help them
• 18:54
  - when they're making decisions about AI, whether it's
• 18:57
  - purchasing decisions,
• 18:58
  - whether it's governance, which is very important.
• 19:01
  - And then the third area is around transformation
• 19:04
  - of the way that they do their work.
• 19:07
  - Seema,
• 19:08
  - incredible to visit with you again today.
• 19:10
  - I followed your career, it seems like, for
• 19:12
  - a long time. Just a brilliant person, brilliant
• 19:15
  - career,
• 19:16
  - and amazing the work you're doing at Oracle
• 19:18
  - Healthcare.
• 19:18
  - Thank you so much for joining us today
• 19:20
  - on the Becker's Healthcare podcast. What a great
• 19:22
  - pleasure. Thank you so much. Thank you, Scott.
• 19:24
  - Thanks for having me. Hope to see you
• 19:26
  - soon.