Latest / Becker’s Healthcare Digital Health + Health IT / Harnessing AI and Data in Healthcare: Insights from Dr. Nabil Chehade of MetroHealth

Transcript

Harnessing AI and Data in Healthcare: Insights from Dr. Nabil Chehade of MetroHealth · Becker’s Healthcare Digital Health + Health IT

• 0:00
  Philips is a health tech leader focused on
• 0:03
  innovation that improves the health and well-being of
• 0:06
  people. Our health care technology and informatics solutions
• 0:09
  help care teams diagnose, treat, and manage more
• 0:12
  patients with greater precision, speed, and confidence across
• 0:16
  the care journey. With Philips, clinicians are empowered
• 0:19
  with streamlined insights in the moments that matter
• 0:21
  for every patient. Better care for more people.
• 0:25
  Philips.
• 0:28
  This is Haley Recker with the Becker's Health
• 0:30
  Care Podcast. We are live at the 9th
• 0:33
  annual health IT Digital Health and RCM conference.
• 0:36
  And today, I am joined by Nabil Shahadi,
• 0:39
  who is the chief clinical transformation officer at
• 0:41
  MetroHealth.
• 0:42
  Nabil, thank you so much for speaking with
• 0:44
  me today.
• 0:46
  To get us started, can you please share
• 0:47
  a bit about yourself, your background, and your
• 0:49
  role at your organization?
• 0:51
  Thank you, Haley, for hosting this podcast today.
• 0:56
  I'm a a physician, a urologist by training.
• 1:00
  I am the chief clinical transformation officer at
• 1:03
  MetroHealth,
• 1:04
  and MetroHealth is a super safety net hospital
• 1:07
  in Cleveland, Ohio.
• 1:09
  We are about $2,000,000,000
• 1:11
  on revenue,
• 1:12
  5 hospitals,
• 1:15
  pre emergency room,
• 1:17
  and more than 30 outpatient,
• 1:20
  clinic. We serve more than
• 1:23
  60% of our patient are governmental patients, which
• 1:26
  means Medicaid, Medicare patients, and uninsured.
• 1:31
  Alright. Thank you so much for sharing. So
• 1:33
  for my first question, I wanted to ask
• 1:35
  about AI adoption, which is currently exploding in
• 1:37
  health care. So in your view, what's the
• 1:40
  most significant or promising application of this technology
• 1:43
  right now, and how is this informing your
• 1:45
  organization's innovation strategy?
• 1:48
  Thank you for this great question. There's no
• 1:50
  easy answer for it. There's a lot of,
• 1:53
  potential excitement,
• 1:55
  but potentially also so many potential disappointment,
• 1:59
  as well.
• 2:00
  So how do we tackle this?
• 2:03
  I think few things we we can,
• 2:06
  zoom
• 2:07
  into and and have a conversation about.
• 2:10
  1 is,
• 2:11
  into
• 2:12
  data
• 2:13
  and,
• 2:14
  predictive data analytics.
• 2:16
  How can you leverage AI
• 2:19
  into streamlining,
• 2:21
  that data and and predict a disease before
• 2:25
  it becomes,
• 2:26
  a much bigger problem?
• 2:28
  How do you use it to help manage
• 2:30
  chronic disease
• 2:32
  on an outpatient setting?
• 2:34
  How do you use it
• 2:36
  within the inpatient setting to predict a potential
• 2:41
  bad outcome before it happens? Like potential very
• 2:44
  well known,
• 2:46
  sepsis
• 2:47
  sepsis prediction
• 2:48
  or fall,
• 2:50
  prediction on the nursing unit. So how do
• 2:52
  you help your nurses and clinical staff
• 2:54
  to prevent
• 2:56
  a catastrophic
• 2:58
  event before it happens? And can you have
• 3:01
  early warning signs that can help you in
• 3:04
  the management,
• 3:05
  of those,
• 3:07
  potentially disastrous,
• 3:09
  health outcomes for your patient?
• 3:11
  So that's on the,
• 3:13
  structured
• 3:14
  data
• 3:15
  side.
• 3:16
  However,
• 3:17
  on imaging,
• 3:19
  and diagnostic,
• 3:21
  there is a
• 3:23
  lot to
• 3:24
  be wishing for,
• 3:26
  with AI. Can AI help us into,
• 3:31
  some of the images in radiology,
• 3:33
  MRI, x-ray, CT scans, etcetera?
• 3:36
  And
• 3:37
  can
• 3:38
  AI
• 3:40
  accelerate the efficiency
• 3:41
  of the provider reading those x rays?
• 3:45
  Can it,
• 3:46
  unravel
• 3:47
  some subtle
• 3:49
  differences in the images, etcetera, that might lead
• 3:51
  to a
• 3:53
  different diagnostic,
• 3:55
  outcome.
• 3:56
  But the same can apply for pathology,
• 3:59
  reading slides for,
• 4:01
  cardiologists
• 4:02
  reading their own images,
• 4:04
  even dermatologists lesion.
• 4:06
  And can can can you look at patterns
• 4:06
  that the naked eye cannot, truly see? And
• 4:08
  can you really help AI with those diagnostic
• 4:10
  tools? Absolutely.
• 4:18
  You mentioned the intersection of AI and data
• 4:21
  in the very beginning there, and so I
• 4:23
  kinda wanna pivot the conversation to data specifically.
• 4:26
  On a daily basis, health care leaders are
• 4:28
  managing greater volumes of data and more devices
• 4:31
  across a growing number of care settings and
• 4:33
  populations.
• 4:34
  In this complex environment, what clinical data integration
• 4:37
  tools or practices are you seeing drive improvements
• 4:40
  in patient outcomes and operations? And then can
• 4:43
  you share an example?
• 4:44
  Sure. That's probably
• 4:47
  the biggest,
• 4:48
  conundrum and issue that we have to solve
• 4:50
  too. All AI tools
• 4:53
  are based on what kind of data we
• 4:55
  have,
• 4:57
  and bad data will lead to bad outcome.
• 5:01
  Biased data
• 5:03
  will lead to biased
• 5:05
  decision support tool,
• 5:07
  that we have today.
• 5:09
  So the opportunity,
• 5:12
  as I stated before, to get a lot
• 5:14
  of, predictive analytics, etcetera,
• 5:17
  only
• 5:19
  if your
• 5:20
  basic
• 5:21
  data gathering is clean.
• 5:24
  And what we struggle today
• 5:26
  as a health system,
• 5:28
  not just integrating
• 5:30
  data
• 5:31
  that come to us from all kind of
• 5:33
  different sources from outside of our organization.
• 5:37
  Because on an average,
• 5:39
  a patient that sees us at MetroHealth will
• 5:42
  only get 50% of their care at MetroHealth,
• 5:44
  but that the other 50% is somewhere else.
• 5:48
  So how can you account for that data?
• 5:50
  Also, the payers so we're the health system.
• 5:53
  There's the payer, and the payer have the
• 5:55
  those data. And can you
• 5:57
  bring,
• 5:59
  that data and integrate it with your,
• 6:01
  electronic medical records so you have that nice
• 6:04
  base of data that that you can
• 6:07
  use to build your your AI tools on?
• 6:12
  But that's only one part of the equation.
• 6:13
  The other part of the equation
• 6:15
  is within your own health system.
• 6:18
  There are all those peripheral
• 6:20
  and devices
• 6:22
  just like EKGs
• 6:25
  or or monitors on the inpatients,
• 6:27
  that are really collecting data?
• 6:30
  How do you interface all of those and
• 6:32
  bring them into a a a nice,
• 6:35
  data warehouse? And finally,
• 6:37
  are the wearable,
• 6:39
  with the consumers?
• 6:40
  And this is one of the biggest challenge.
• 6:42
  Every consumer
• 6:44
  is collecting
• 6:45
  data
• 6:46
  and wants to share that data with their
• 6:48
  provider, but that's gonna bring a lot of
• 6:50
  noise about,
• 6:52
  useful data that you can act on it
• 6:54
  or not, but that's probably where
• 6:57
  we're innovating the most, trying to curate and
• 7:00
  try to figure out what is
• 7:03
  the most important
• 7:04
  data element,
• 7:05
  that you're receiving from your patient wearable that
• 7:08
  a provider
• 7:10
  a a team of provider and their nurses
• 7:12
  and assistants can act upon and make a
• 7:15
  difference.
• 7:16
  In your view, how can health care organizations
• 7:19
  better support IT and clinical teams as they
• 7:21
  carry out innovation efforts, and what are the
• 7:23
  common pitfalls that you see here?
• 7:26
  Another good question.
• 7:28
  It's all about governance.
• 7:30
  I truly believe
• 7:32
  you have to have a strong,
• 7:34
  governance
• 7:35
  internally to your health system
• 7:37
  in order to streamline
• 7:41
  what kind of,
• 7:42
  application you're bringing in.
• 7:45
  A lot of those application
• 7:47
  promises
• 7:48
  they promise a lot,
• 7:50
  but most of the time, they've not proven
• 7:53
  their outcome.
• 7:55
  And without a
• 7:57
  true governance in place, how you can streamline,
• 8:00
  how you can allow small pilot to happen
• 8:04
  before you commit your organization into 6, 7
• 8:07
  figure
• 8:08
  expenses
• 8:09
  without any
• 8:11
  data to prove that the health outcome is
• 8:14
  better
• 8:15
  or actually your cost has improved and your
• 8:17
  efficiency have improved.
• 8:19
  I've seen it so many
• 8:21
  time that there is a rush to bring
• 8:23
  in the excitement,
• 8:26
  but then
• 8:27
  you spend a lot of
• 8:30
  resources
• 8:31
  and the outcome is not better and sometimes
• 8:34
  it's worse. So you gotta do baby steps.
• 8:37
  So that's that's one of my biggest thing.
• 8:40
  The other one is you don't wanna over
• 8:43
  govern and not allow innovation to happen.
• 8:46
  You wanna
• 8:47
  be very mindful that
• 8:50
  if you have a good governance structure,
• 8:52
  if
• 8:54
  you allow them to innovate, but take those
• 8:56
  baby steps, you are more likely to figure
• 8:59
  out what is the best that you're gonna
• 9:01
  ultimately invest and and and deploy system wide.
• 9:06
  Alright. Thank you so much for sharing. And
• 9:08
  to wrap up our conversation today, what's your
• 9:10
  top piece of advice for health care leaders
• 9:12
  as they prepare for further advancements in technology
• 9:15
  and greater demands for care?
• 9:18
  Don't be afraid
• 9:19
  of the technology,
• 9:21
  but
• 9:22
  you gotta also be cognizant that within your
• 9:25
  health system,
• 9:27
  people are gonna use technology whether you allow
• 9:30
  it or not. So
• 9:33
  embrace it,
• 9:34
  but embrace it in a very calculated way.
• 9:37
  Have a very strong governance,
• 9:39
  not just of what you adopt and what
• 9:41
  you don't adopt and what you invest in
• 9:44
  and not invest in, but also be very
• 9:47
  cognizant about your data infrastructure
• 9:49
  and invest into a
• 9:51
  good data warehouse. So
• 9:53
  it is not,
• 9:55
  bad data in
• 9:57
  bad outcome, bad decision. You want good data,
• 10:00
  and you want data that is unbiased so
• 10:03
  you don't create a health disparity while you're,
• 10:06
  relying on AI.
• 10:08
  Thank you so much for sharing your insight
• 10:10
  with us. Once again, we are at the
• 10:12
  9th annual health IT digital health and RCM
• 10:14
  conference.
• 10:15
  Nabil, thank you so much for being here.
• 10:17
  Thank you, Hailey. Thank you.