Latest / Becker’s Healthcare Podcast / Dave Newman, MD, Chief Medical Officer of Virtual Care at Sanford Health

Transcript

Dave Newman, MD, Chief Medical Officer of Virtual Care at Sanford Health · Becker’s Healthcare Podcast

• 0:00
  Hi. Welcome everyone to the Becker's Healthcare podcast.
• 0:03
  This is Giles Bruce. I'm an assistant editor
• 0:06
  with Becker's Healthcare covering health IT.
• 0:10
  Pleased to be joined by doctor Dave Newman
• 0:12
  from Sanford Health.
• 0:14
  We we just had him on a a
• 0:16
  fireside chat at our our AI summit at
• 0:19
  at Becker's CEO CFO roundtable. So pleased to
• 0:22
  have him again for a a podcast to
• 0:24
  talk more things virtual health and AI.
• 0:27
  And, Dave, can you please start by, you
• 0:29
  know, introducing yourself, telling me a little bit
• 0:31
  about your your role and what you do
• 0:33
  at Sanford? Yeah. Absolutely. So I'm Dave Newman.
• 0:35
  It's good to see you again, Giles. I'm
• 0:37
  an endocrinologist and informaticist at Sanford Health, which
• 0:40
  is the nation's largest rural health care system.
• 0:42
  We're headquartered in the Dakotas. We serve a
• 0:45
  huge amount of territory in the Midwest. We've
• 0:47
  got hospitals, clinics, just about everything every single
• 0:50
  subspecialty,
• 0:51
  and we are dedicated to being, the leading
• 0:54
  rural health care system in the in the
• 0:56
  nation.
• 0:58
  Excellent.
• 0:59
  So yeah. So I mentioned you're did a
• 1:01
  fireside chat with us at the the AI
• 1:03
  summit, and we really we're pleased that you're
• 1:05
  able to join us. But you you asked
• 1:07
  a provocative question there. I found it interesting.
• 1:09
  Basically, when is it irresponsible
• 1:11
  not to use AI? So so what in
• 1:13
  indicators will will tell us we've reached that
• 1:15
  point in health care? And are there any
• 1:17
  specific clinical areas,
• 1:19
  insulin dosing, you know, chronic kidney disease detection
• 1:22
  perhaps where you believe we're already close to
• 1:25
  that point?
• 1:26
  Okay. So let me start with a story
• 1:28
  about when I first started thinking about this.
• 1:30
  So as an endocrinologist, I treat a lot
• 1:32
  of type one diabetes. So type one diabetes
• 1:34
  is a condition where your pancreas doesn't work.
• 1:36
  So you need to take insulin,
• 1:39
  and it for years was an unbelievably
• 1:42
  hard condition to treat.
• 1:44
  Patients had to have a large amount of
• 1:46
  education to know their body, to know how
• 1:48
  many carbohydrates they're eating, to dose their insulin,
• 1:50
  when to take more, when to take less,
• 1:53
  when to eat more carbohydrates.
• 1:55
  And I had patients that I was seeing
• 1:56
  every three months that I could not get
• 1:58
  a handle on.
• 2:00
  Insulin pumps came out a long time ago
• 2:02
  and they sorta helped, but still the patient
• 2:04
  had to be armed with a huge amount
• 2:06
  of knowledge,
• 2:07
  in order to continue to be successful with
• 2:09
  these. Even with the insulin pump and continuous
• 2:12
  glucose sensors, which always check your blood sugar,
• 2:15
  patients were still poorly controlled because there was
• 2:18
  there wasn't great communication.
• 2:20
  I have a patient. We'll call him Bruce.
• 2:22
  I had followed him for over ten years,
• 2:24
  and I had not gotten him well controlled.
• 2:27
  That was on me. We had gone through
• 2:29
  everything. We had done diabetic education. He'd been
• 2:31
  with nutritionist. He knew everything about carbohydrates,
• 2:35
  But, unfortunately, it was a really hard condition,
• 2:37
  and he was being admitted to the hospital
• 2:39
  for hypoglycemia
• 2:40
  or high blood sugars frequently.
• 2:43
  About two or three years ago,
• 2:45
  Bruce got something called a closed looped pump
• 2:48
  and sensor,
• 2:49
  where it's an artificial intelligence enabled device that
• 2:51
  you wear in your belt that when your
• 2:53
  blood sugars go high, it gives you more
• 2:55
  insulin, and when it goes low, it stops
• 2:58
  giving you insulin.
• 2:59
  This was mind blowing to both me and
• 3:01
  Bruce.
• 3:02
  All of a sudden, it was like every
• 3:04
  eight seconds, he had an endocrinologist
• 3:06
  sitting with him that could make decisions.
• 3:09
  Based upon how his blood sugars were acting,
• 3:11
  it would automatically give him more insulin.
• 3:13
  He only really had to be responsible for
• 3:16
  charging his pump
• 3:17
  and filling with insulin every three days.
• 3:20
  I have decreased my visits. I used to
• 3:22
  see him every three months. Now I see
• 3:23
  him every six months. We might be able
• 3:25
  to go to every year visits because the
• 3:27
  AI is doing the work. It's better than
• 3:29
  I am. It took me realizing that I
• 3:32
  had to partner with the technology to do
• 3:33
  that with all my patients.
• 3:35
  As we move forward, we're gonna see this
• 3:37
  with other conditions.
• 3:38
  Whether it is in pathology,
• 3:41
  whether it is when we see peer reviewed
• 3:43
  studies where it's AI is better at diagnosing
• 3:45
  celiac disease. We have to educate our pathologists,
• 3:48
  and they've gotta be the ones that lead
• 3:49
  this charge. We want a human in the
• 3:51
  loop, so we want
• 3:53
  nephrologists, we want primary care doctors verifying that
• 3:56
  the CKD
• 3:57
  models
• 3:58
  are accurate.
• 3:59
  But at some point, we have to partner
• 4:00
  with the technology.
• 4:02
  I kinda like the anecdote is when I
• 4:04
  go to my accountant, I would be really,
• 4:06
  really skeptical if they're using Abacus to do
• 4:08
  my taxes. Right? Like, I want them using
• 4:10
  Excel. I want them using some sort of
• 4:12
  program.
• 4:13
  I'm gonna be skeptical if my doctors aren't
• 4:15
  using AI if it's the best available tool.
• 4:20
  Yeah. Great example there about this technology is,
• 4:22
  you know, really having some some benefits for
• 4:24
  patients.
• 4:26
  You also reported near universal
• 4:28
  provider satisfaction with ambient documentation and even joked
• 4:32
  that you'd have a a riot if it
• 4:33
  were removed. So why does ambient AI succeed
• 4:36
  where where past documentation tools have failed? And
• 4:39
  and how would you say it's changed the
• 4:41
  patient experience? You mentioned a moment you described
• 4:43
  when patients said it was the, you know,
• 4:45
  quote unquote, first time you you looked actually
• 4:47
  looked at them.
• 4:49
  Yeah.
• 4:50
  So I I I joked about having a
• 4:52
  riot. I've now I think I would have
• 4:54
  a riot if we took away Ambi and
• 4:56
  Experian.
• 4:57
  I get lots of emails. I get way
• 4:59
  more emails when our Ambi and Experian tool
• 5:01
  goes down than any time else, because it
• 5:03
  has improved so many doctors' lives.
• 5:05
  I've got a buddy. We'll call him Sam.
• 5:07
  He's a pediatric oncologist.
• 5:09
  Unbelievable
• 5:10
  doctor, but has never been good at writing
• 5:12
  notes.
• 5:14
  So he spend a lot of time with
• 5:15
  his his kids,
• 5:17
  and parents going over treatment plans, going over
• 5:19
  all the things that doctors do.
• 5:21
  But because of this, he wasn't great at
• 5:24
  entering everything into the medical record.
• 5:27
  He was always on what we call our
• 5:28
  naughty list, so overdue documentation.
• 5:32
  After two weeks of having the ambient experience,
• 5:34
  he was no longer on the naughty list.
• 5:37
  He came up and he gave me a
• 5:38
  big hug and he said, I can't believe
• 5:39
  that I've got my life back. That he's
• 5:41
  able to see his patients, do the things
• 5:43
  that he did, and it's easy.
• 5:45
  The problems with the old note taking tools,
• 5:48
  the problems with the other softwares
• 5:49
  is that it required lots of clicks. It
• 5:52
  required
• 5:53
  a somewhat unorthodox
• 5:54
  workflow that doctors had to do, and it
• 5:57
  wasn't easy. Ambient is easy and it gives
• 5:59
  your life back.
• 6:00
  Some statistics.
• 6:02
  And in medicine, we never reach 100%.
• 6:05
  A 100% of the providers that have used
• 6:07
  this have said they want to keep doing
• 6:09
  it. They want to stay.
• 6:10
  Ninety percent have reported greater job satisfaction and
• 6:14
  ninety five percent have reported less mental fatigue.
• 6:18
  We firmly believe that this is gonna be
• 6:20
  a tool that is gonna help,
• 6:21
  retain the best talents in the Midwest. That
• 6:24
  if we don't have these tools, you're not
• 6:26
  gonna be able to recruit and retain doctors
• 6:28
  going forward.
• 6:30
  Yeah. Yeah. Great stuff. And, yes, similar to
• 6:32
  some of the benefits I've been hearing hearing
• 6:34
  from other leaders across the country on, you
• 6:36
  know, clinician
• 6:37
  satisfaction and and burnout, etcetera.
• 6:41
  And next question. You you gave an example
• 6:42
  at the conference of of patients having to
• 6:44
  drive six hours for a a ten minute
• 6:46
  visit, you know, very pow powerful anecdote.
• 6:49
  So what would you say in the way
• 6:51
  of, virtual care and telehealth is is having
• 6:53
  the biggest impact on reducing some of that
• 6:55
  travel burden and and miscarare in rural communities?
• 6:59
  Yeah. So, our statistic at Sanford is virtual
• 7:02
  care saves patients an average of a 176
• 7:05
  miles of travel. That's just the miles saved.
• 7:08
  So it is time off of work. It
• 7:10
  is meals. It is hotels. It is putting
• 7:13
  your dog,
• 7:15
  in doggy daycare. It is missing your kid's
• 7:17
  gymnastics events. It's all those things that really
• 7:20
  add up that we're really bringing the power
• 7:21
  back to the patients. One of my best
• 7:23
  examples
• 7:24
  is we have a phenomenal doctor in Fargo.
• 7:27
  His name is Seth, and he's a bone
• 7:29
  marrow transplant doctor.
• 7:31
  He has revolutionized bone marrow transplant by leveraging
• 7:34
  virtual care.
• 7:35
  In the past,
• 7:37
  patients that did a bone marrow transplant required
• 7:39
  to stay in Fargo inpatient for five to
• 7:41
  six weeks because they needed so many cares.
• 7:44
  Now he's able to have them inpatient for
• 7:47
  two to three weeks
• 7:48
  and then
• 7:49
  send them home and do virtual visits for
• 7:52
  the rest of the time. So they don't
• 7:54
  have to be in the hospital environment for
• 7:56
  that long.
• 7:57
  I've got lots of patients that I have
• 7:59
  to see every three months for medical requirements
• 8:00
  for diabetes. They no longer have to come
• 8:02
  to see me. They can do it from
• 8:04
  satellite clinics or they can do it from
• 8:05
  home.
• 8:07
  If now patients can get care close to
• 8:09
  home in the Midwest instead of going to
• 8:11
  a massive tertiary referral center, which they don't
• 8:13
  wanna do. They wanna stay with their doctors
• 8:15
  that know them, and they have their that,
• 8:17
  have their
• 8:19
  records and know have been knowing them for
• 8:21
  over ten years.
• 8:23
  Great stuff. Yeah. And I know from our
• 8:25
  our previous reporting,
• 8:26
  at Becker's, you know, how big of a
• 8:28
  investment and commitment that that Sanford has made
• 8:30
  to virtual care.
• 8:32
  It's turning to AgenTek AI, which is still,
• 8:34
  you know, kind of in its infancy in
• 8:36
  in health care and elsewhere. But but you
• 8:37
  mentioned some of your,
• 8:39
  AgenTic,
• 8:40
  outbound campaigns, improving connect rates from 40% to
• 8:44
  56%,
• 8:46
  you know, full twenty or so minute conversations
• 8:49
  with sixteen percent of patients. Pretty remarkable stuff.
• 8:52
  What what has surprised you most about how
• 8:53
  patients are responding to AgenTek AI, and, what,
• 8:56
  you know, what do you think this has
• 8:57
  potential as far as, you know, closing care
• 9:00
  gaps for for cancer screenings, primary care, chronic
• 9:03
  disease, etcetera?
• 9:05
  Yeah. I think the thing that surprised us
• 9:07
  us the most was the satisfaction. So,
• 9:09
  we are scoring over eight out of 10
• 9:11
  on patient satisfaction, which we never thought we
• 9:13
  would do. You know, from getting these these,
• 9:15
  bot calls, sometimes they can be quite challenging.
• 9:18
  When it comes to health care, they know
• 9:20
  that we have are their best interest.
• 9:23
  Would I rather talk to a human sometimes?
• 9:25
  Absolutely.
• 9:26
  But would I rather talk to an agentic
• 9:28
  agent to get my colonoscopy scheduled than not
• 9:31
  get my colonoscopy?
• 9:33
  Absolutely.
• 9:34
  That's the feedback that we've been getting. People
• 9:36
  want to close their care gaps. They wanna
• 9:38
  get their lipid panels done. They wanna have
• 9:40
  colon cancer screening. They wanna have all these
• 9:42
  things done to make them healthy. The rate
• 9:45
  limiting step now is we don't have the
• 9:47
  workforce.
• 9:48
  So right now, we're able to reach,
• 9:50
  thousands more patients using an agent and get
• 9:53
  them connected to the person that they need
• 9:55
  to get their their, visit scheduled or to
• 9:58
  get their specialty visit scheduled.
• 10:02
  Yeah. Great stuff. And and we'll definitely
• 10:04
  continue following that story as as AgenTek AI
• 10:07
  evolves in the years to come.
• 10:09
  And and you also spoke candidly at the
• 10:11
  the conference about governance,
• 10:13
  separating clinical and operational committees, ensuring,
• 10:17
  you know, operators and clinicians also sit on
• 10:19
  each other's boards. What what does good AI
• 10:21
  governments look like day to day at Sanford,
• 10:23
  and and how do you build clinician trust
• 10:25
  without slowing innovation?
• 10:27
  So this is a story that happened to
• 10:29
  me. I went on a call this week,
• 10:30
  and I was getting lunch at one of
• 10:32
  our flagship hospitals in the Doctor's lounge. And
• 10:35
  I literally had a doctor come up to
• 10:36
  me and say, hey. I've got this AI
• 10:39
  blood loss app.
• 10:41
  How do I use it?
• 10:42
  And I was like, that's super cool. Let's
• 10:44
  run it through governance. But he never heard
• 10:46
  of governance. So it's up to us to
• 10:48
  communicate the strategy with everybody that has these
• 10:50
  products.
• 10:51
  We have to make sure they're safe and
• 10:52
  effective, which means that they have to be
• 10:55
  safe and effective through a doctor's law lens.
• 10:57
  But there also has to be an operational
• 10:59
  lens. Cybersecurity,
• 11:01
  that we've gotta look at that. We've gotta
• 11:02
  look at if it is fiscally responsible for
• 11:04
  the system to do. We've gotta look at
• 11:06
  how we can have this these apps work
• 11:08
  not just for one provider, but how do
• 11:11
  we scale it through the entire organization.
• 11:14
  We also have to look at how we
• 11:16
  monitor outcomes,
• 11:17
  not only for the first six months, but
• 11:19
  going forward. How is this safe and effective
• 11:22
  for patients in the long term? We're talking
• 11:24
  about having real time dashboards that give insights
• 11:27
  into how these are trained and how they're
• 11:29
  performing. We need to make sure that the
• 11:31
  the results are accurate for our patient populations,
• 11:34
  not just the patient populations that the the,
• 11:37
  AI ventures were tested on. This is a
• 11:40
  lot of work, and it requires a lot
• 11:42
  of people around the table,
• 11:43
  but we are very, very committed to doing
• 11:45
  what's best for the patient.
• 11:49
  Yeah. Good stuff. Advice there. I know,
• 11:52
  yeah, health systems are definitely looking for guidance
• 11:55
  on their their AI governance journey. So, yeah,
• 11:57
  some of that will will definitely help help
• 11:59
  our listeners.
• 12:01
  And your description of orthopedic residents practicing procedures
• 12:05
  with with AI enabled virtual reality was also
• 12:08
  a highlight of our our last chat. How
• 12:10
  do you see virtual training shaping the next
• 12:12
  decade of medical education?
• 12:15
  Will future condition do you think future clinicians
• 12:17
  will need a a different skill set, maybe
• 12:19
  be more like, you know, prompt engineers or
• 12:22
  clinical workflow designers?
• 12:24
  I love this question.
• 12:26
  And for the listeners,
• 12:27
  my mind was absolutely blown the first time
• 12:30
  that I saw the orthopedic surgery,
• 12:33
  AI VR, where it is allows them to
• 12:36
  go through a surgery using virtual reality goggles
• 12:39
  and participate in the surgery before they go
• 12:41
  into the operator.
• 12:43
  For a little bit of of, level setting,
• 12:46
  I have been a doctor for about twenty
• 12:48
  years, and I learned from textbooks.
• 12:50
  YouTube wasn't really a thing even, so I
• 12:52
  couldn't even watch videos of surgery before I
• 12:54
  went in. We have evolved from textbooks,
• 12:58
  from videos and now to virtual reality, and
• 13:02
  the learning is substantially improved.
• 13:05
  The orthopedic surgery residents can look at patient
• 13:07
  placement. They can see how the shoulder is
• 13:09
  supposed to be placed beforehand.
• 13:11
  They can map out anatomical landmarks.
• 13:14
  They can hit a button on their virtuality
• 13:16
  goggles, and the skin will be removed so
• 13:18
  they can look at musculature. They can hit
• 13:19
  another button, the muscles will be removed, and
• 13:21
  they can look at vasculature.
• 13:23
  They can have the AI prompts
• 13:26
  be
• 13:27
  have problem solving. So they can say, oh,
• 13:29
  you know, this vein or this nerve isn't
• 13:32
  in the right spot. How do I deal
• 13:34
  with that? I love your questions about being
• 13:36
  prompt engineers.
• 13:37
  So most like, my kids now. I've got
• 13:39
  a college kid. I've got high school kids.
• 13:41
  They're amazing at using AI. They're great at
• 13:44
  designing their own prompts. Doctors need to get
• 13:46
  good at prompts to be better at doing,
• 13:49
  medicine.
• 13:50
  This is something that we have to be
• 13:52
  educating on. One of the things that we
• 13:54
  see our virtual care initiative doing is being
• 13:56
  one of the the nation's leaders on education,
• 13:59
  that we want people to be good at
• 14:00
  AI. And we've had lots of older doctors
• 14:03
  come to us and say, hey. I've never
• 14:04
  used these things.
• 14:05
  Can you give me a three to five
• 14:07
  minute video or sandbox environment to do this?
• 14:09
  So we designed our own curriculum to help
• 14:11
  people with that.
• 14:14
  Great stuff as always.
• 14:16
  Yeah. With that, we're gonna wrap up. But
• 14:18
  really appreciate the talk today, Dave, and your
• 14:20
  insights on on virtual care and AI. And,
• 14:22
  also, we, look forward to seeing you. You're
• 14:25
  gonna be speaking on a keynote panel at
• 14:27
  our health IT conference,
• 14:29
  next, next year in Chicago. So I really
• 14:32
  appreciate you coming back to another Becker's event
• 14:34
  and and doing the podcast today. I'm excited,
• 14:36
  Josh. Thanks for having me. Awesome. We'll see
• 14:39
  you then, Dave.