Latest / Becker’s Healthcare Digital Health + Health IT / Building Ethical, Equitable AI in Public Healthcare with Divya Pathak of NYC Health + Hospitals

Transcript

Building Ethical, Equitable AI in Public Healthcare with Divya Pathak of NYC Health + Hospitals · Becker’s Healthcare Digital Health + Health IT

0:00
Exciting things are happening at Becker's Healthcare. Stay
0:04
ahead of industry trends with the new Becker's
0:06
CFO plus Revenue Cycle podcast,
0:08
your go to source for insights from top
0:11
healthcare finance leaders. Tune in wherever you get
0:14
your podcasts.
0:15
And don't miss the tenth annual health IT
0:18
plus digital health plus RCM conference, happening September
0:22
30 to 10/03/2025
0:24
in Chicago.
0:25
Join thousands of
0:27
executives, engage with industry leaders, and explore the
0:30
future of health care innovation.
0:32
Learn more about our upcoming events at beckershospitalreview.com.
0:36
See you there.
0:38
This is Laura Dierda with the Becker's health
0:40
care podcast.
0:40
I'm thrilled today to be joined by Divya
0:42
Patek,
0:43
chief data and AI officer at New York
0:46
City Health and Hospitals. Divya, it's a pleasure
0:48
to have you on the podcast today. So
0:50
I'm excited for our conversation because I I
0:52
know there's so many cool things that you're
0:54
doing at New York City Health and Hospitals
0:56
and certainly a a very important community that
0:59
you're serving there. So, I'm excited to dive
1:01
in. But before we do, I'm wondering if
1:03
you could tell me a little bit more
1:04
about
1:05
New York City Health and Hospitals and what
1:07
makes the system unique.
1:09
Thank you, Laura, for inviting me to this
1:11
podcast.
1:12
My name is Divya Pathak. As you mentioned,
1:14
I'm the chief data and AI officer for
1:15
the New York City health and hospitals,
1:18
the largest public health system in our country.
1:21
And I provide executive leadership and strategic direction
1:24
to the data and AI organization,
1:27
to support our network of hospitals
1:30
and its partners
1:31
with data, digital tools, and AI capabilities and
1:35
insights
1:36
that drives quality and performance improvements,
1:38
enterprise operations,
1:40
and enhances patient outcomes.
1:43
And to speak about our health system,
1:46
our network includes 11 acute care hospitals and
1:49
a network of over 30 community based clinics,
1:52
long term care facilities,
1:54
and trauma centers
1:56
in the five boroughs in New York City,
1:58
allowing us to offer a full spectrum of
2:00
health care services.
2:02
Beyond just medical care, we focus on improving
2:05
the overall well-being of our patients
2:07
by addressing the social determinants of health, such
2:10
as access to housing, food, and other critical
2:13
resources.
2:14
What makes us unique is our unwavering commitment
2:17
to providing comprehensive
2:18
care to all individuals
2:20
irrespective of their
2:22
socioeconomic background,
2:23
racial, ethnic status, and legal status.
2:27
And as a safety net provider, we see
2:30
a diverse and often
2:31
underserved population ensuring that essential health care is
2:35
available to everyone in New York City, including
2:38
the most vulnerable communities,
2:40
and I'm very proud to be part of
2:42
it.
2:42
Well, that's amazing to hear. You know? And
2:44
as you mentioned, just such a a great
2:46
mission that you have there in New York
2:48
City to serve so many,
2:49
different unique and diverse populations.
2:52
It really I know it makes a big
2:53
difference for the health and and safety of
2:55
the community.
2:56
Now I'm curious. What, could you tell us
2:59
about an accomplishment that you're most proud of
3:01
from the last year?
3:02
Sure. So there are several,
3:05
but to speak of it, I'm very proud
3:08
of
3:09
a collaboration with revenue cycle services
3:12
department at health and hospitals. And one of
3:14
the key initiatives
3:16
we focused on is in building an AI
3:18
driven one stop benefits application.
3:21
Given the diverse and underserved
3:24
population,
3:25
this project has been more
3:27
meaningful and deeply impactful
3:30
in our community.
3:31
And to speak to it, what we have
3:34
developed is a holistic financial counseling model
3:37
that proactively addresses patients' financial needs, aiming to
3:41
improve their health,
3:43
well-being, and equity.
3:45
And by integrating AI and advanced data and
3:48
analytics, our financial counselors can now conduct proactive
3:51
screenings for financial assistance and public benefits
3:54
prior to patient appointments,
3:56
and this reduces their waiting times. And this
3:59
approach not only streamlines the process,
4:01
but also maximizes the financial support a vulnerable
4:04
patient will receive,
4:06
ultimately helping them to alleviate some of the
4:08
burdens they face while navigating the health care
4:11
system.
4:12
And it's a truly rewarding project personally as
4:15
well, and that makes a direct impact
4:17
to the lives of our patients.
4:20
And, we're continue continuing
4:22
to to more such,
4:24
efforts in support of our population.
4:27
Well, it's amazing to hear. You know? What
4:28
a great idea of being able to proactively
4:30
identify and then approach those patients who may
4:33
need a little bit more financial support, and
4:36
have a plan in place. I can imagine
4:38
that it's such a relief for so many
4:39
who are are trying to seek needed care
4:42
not to have to worry as much about
4:44
the financial burden that it could present for
4:46
them and their families.
4:48
Absolutely, Laura. And in fact, it's not just
4:50
on the patient side too. Right? A lot
4:52
of these benefits that's provided by our city
4:54
government goes unnoticed
4:55
or not even made aware of.
4:58
So we we take incredible
5:00
effort to make sure that there is a
5:02
lot of proactive support to these patients
5:05
and the appointments and their
5:08
length of stay,
5:09
during their time in health and hospitals is
5:11
seamless,
5:12
and,
5:13
they provide the you know, we provide the
5:15
compassionate
5:16
service that they need the most.
5:19
That's great to hear. You know, and and
5:21
certainly, you know, wonderful to have that type
5:24
of support technology to support your teams and
5:26
then mission to go out there and do
5:28
that. Now I'm curious. Where do you see
5:30
the big growth opportunities for you and your
5:32
team in the next twelve months or so?
5:35
So,
5:36
one of the recent focus areas is the
5:39
establishment of our AI organization,
5:41
under the portfolio of data and AI.
5:44
And this organization is relatively new, and we
5:47
are slowly ramping up our internal AI capabilities
5:51
across our enterprise.
5:53
In doing that, we're embracing AI with a
5:55
very strategic blend of ambition and pragmatism,
5:59
recognizing its potential to improve the experience for
6:01
our patients and clinicians
6:03
and driving operational efficiency,
6:06
but also to make sure we enable equitable
6:08
care delivery for New York City patients, which
6:11
aligns with our mission.
6:13
And in the next twelve months, I see
6:15
significant
6:16
growth in scaling our AI driven solutions, again,
6:19
with guardrails across both clinical and nonclinical areas
6:23
and fostering a culture of innovation and continuous
6:26
learning
6:27
within our AI and health care teams. It's
6:29
it is the biggest growth opportunity,
6:32
and,
6:33
we're following the trend in health care.
6:36
That makes a lot of sense. You know,
6:37
certainly, a lot of possibilities and opportunities with
6:40
AI, as you mentioned, both on the clinical
6:42
side as well as the operational side. And
6:45
as you're building out your AI organization,
6:49
what does that look like,
6:51
if you're able to speak to just briefly,
6:53
you know, what skills are you bringing in
6:54
in house, how are you partnering,
6:56
with other organizations
6:58
as you're kind of thinking about, the AI
7:01
for New York City health plus hospitals.
7:04
You know, what what really
7:06
comes to mind for you in terms of,
7:08
like, how you're building up the team, the
7:09
skill sets, the roles, and what you really
7:11
need,
7:12
from your team
7:14
in a for AI in particular?
7:17
That's a great question as we are in
7:19
the midst of it. So, in supporting an
7:21
AI org, we are talking about,
7:24
organization
7:25
that supports strategic procurement
7:28
and strategic development of AI solutions and supporting
7:31
our workflows.
7:32
And in doing that, we also understand that
7:36
there are several other front runners in our
7:38
health care system, several provider organizations. So there's
7:41
a lot of learnings,
7:42
that we want to leverage,
7:45
including
7:46
consortiums both in industry and academia.
7:48
For example, Coalition for Health Care AI,
7:51
Health AI Partnership, Valid AI. These are key
7:54
consortiums that we are part of.
7:56
And,
7:57
additionally, we want to also,
8:00
as I mentioned, leverage the blueprint from existing
8:02
organizations, but also have an organizational
8:05
design with the multidisciplinary
8:07
skills needed to support AI adoption and enablement.
8:10
And in addition to that internal
8:14
capabilities, we also want to have an open
8:16
ecosystem of partnerships
8:18
that spans across
8:20
big to small
8:22
companies,
8:23
startups,
8:24
cloud providers,
8:25
because, you know, this is such an evolving
8:28
space, Laura, that we see that we have
8:31
to be truly multi cloud. We need to
8:33
have options to work with multi multiple vendors,
8:37
and more importantly,
8:38
also have a data strategy, support that open
8:41
data sharing,
8:42
with AI vendors. So there's a lot of
8:44
focus related to open ecosystem of partnerships.
8:48
So in terms of skills we are trying
8:49
to bring in, we want to support the
8:51
entirety of AI life cycle.
8:53
Right? All the way from ideation to development
8:56
to deployment
8:58
to performance and monitoring and also looking at
9:00
evaluation and outcomes. So we're in the process
9:03
of bringing a multidisciplinary
9:05
team,
9:06
some through our professional services, but slowly,
9:09
you know, develop the inbuilt in house talent,
9:13
across data science,
9:15
AI, engineering,
9:17
ML ops,
9:19
UI, UX development, and most importantly, right, integration,
9:22
developers in support of integration with our existing
9:25
systems.
9:27
That's really a helpful outline. Thank you so
9:29
much for going through that with us, Didier.
9:31
Now
9:32
I wanted to ask you as well, lots
9:34
of opportunities ahead, but also challenges. What are
9:38
some of the big challenges that you're anticipating
9:40
over the next year or two?
9:42
Oh, yeah. I mean, always challenges and opportunities
9:45
go in hand. Right? But I take challenges
9:47
as opportunities as well. So one thing, Laura,
9:50
that, you know, I've been really spending some
9:52
time on is to see how we integrate
9:54
AI into our health care systems
9:58
and how do we balance the return on
10:00
investment.
10:01
Yes. There is AI everywhere,
10:04
and there's proven,
10:06
efforts across other health systems.
10:08
But when we talk about return on investment,
10:11
the hard ROI such as cost savings or
10:13
potential revenue,
10:15
operational
10:16
efficiencies,
10:17
or direct improvements in patient out outcomes
10:20
will be more straightforward to measure and track
10:22
as we implement AI. And we need to
10:24
make sure that we have success metrics along
10:27
the
10:28
line.
10:29
But let's talk about areas where there is
10:31
soft ROI, which includes
10:33
more intangible benefits,
10:35
like improving clinician
10:37
satisfaction, reducing burnout, enhancing patient
10:41
experience, and fostering a culture of innovation.
10:44
These are much more difficult to quantify and
10:47
measure.
10:47
So achieving a balance is essential
10:50
for the long term success of AI initiatives.
10:53
So we need to be ensuring that while
10:54
we focus on hard metrics like cost savings
10:56
and efficiency, we also measure and prioritize softer
10:59
human centric aspects of AI that contribute to
11:02
better outcomes,
11:03
trust, and a sustainable
11:05
people first approach to AI integration.
11:08
Additionally, overcoming potential resistance to change,
11:12
ensuring data privacy and security, and upskilling our
11:15
workforce for workforce adoption and trust
11:18
alongside AI solutions will be, you know, hurdles
11:22
that, you know, I anticipate to address.
11:25
And that makes a lot of sense. You
11:26
know, I I really love that idea of
11:28
having those soft metrics as well as the
11:30
hard metrics to think about the return on
11:32
investment and keep the humanity,
11:34
within, you know, the health care system, which
11:37
is such a a human to human industry.
11:39
When you look at those, you know, soft
11:41
ROI,
11:43
how do you think about that? How do
11:44
you know that you've been successful?
11:47
Yeah. So, any initiatives that,
11:50
we we prioritize,
11:52
and I can give an example. Right? Ambient
11:54
listening is a great example,
11:56
that we are actually
11:58
looking to see how we can,
12:00
procure and roll it out because it's a
12:02
very proven solution in the market.
12:04
One of the things we are doing is
12:06
we're making sure we understand the metrics for
12:09
success early on
12:10
even prior to actually moving moving to proceed,
12:14
with procurement. So having those metrics, having baselines,
12:19
you know, again, baselines that cannot be quantified
12:21
but can be qualified. So having baselines,
12:24
making sure we we have inputs from the
12:27
multidisciplinary
12:28
team who's going to eventually use the solution.
12:31
And having a baseline is critical to a
12:34
project success because that's how we're able to
12:37
then see how we're actually improving care, how
12:39
are we
12:40
improving the efficiency, how are we actually improving
12:43
patient experience. So,
12:45
there's there's a very deliberate thought
12:47
put in place by design to have baseline
12:51
for every AI initiatives. And, software initiatives is
12:54
another reason. Right?
12:56
And which is very qualitative, but, certainly, we
12:59
we make a deliberate attempt to capture that
13:02
early on.
13:04
That's helpful to know. I I appreciate your
13:06
time and kind of going a a bit
13:08
deeper there as well. Now before we wrap
13:10
up with our last minute or two, I
13:12
wanted to ask, what is the number one
13:13
thing that you're doing right now to set
13:15
the health system up for long term success?
13:18
That's a great question, Laura. To set New
13:20
York City Health and hospitals up for long
13:22
term success,
13:23
What I'm focusing on right now is to
13:25
build strong AI governance framework
13:28
that ensures equity,
13:30
fairness,
13:31
and transparency in the use of AI technologies.
13:34
And given the diverse and vulnerable patient populations
13:39
populations we serve, it is essential that our
13:41
AI solutions are fair, ethical, and free from
13:41
bias.
13:43
So it's a number one priority for our
13:45
system. So by embedding governance practices, we aim
13:48
to prioritize patient safety, reduce the disparities
13:52
of care, and enhance the delivery while ensuring
13:55
AI aligns with our mission of providing equitable
13:57
care. So this approach,
13:59
we hope. Right? The plan is to actually
14:02
foster more trust
14:04
and improve patient outcomes and ensure our AI
14:06
solutions benefit all populations.
14:10
So governance is key.
14:13
That makes a lot of sense. I I
14:14
know how important it is to set up
14:16
that governance, have the right stakeholders at the
14:18
table, and, really truly make sure that AI
14:21
is doing,
14:23
so much good for the organization and and
14:25
not, getting into some of the risks involved
14:27
as well. That's right, Laura. I think having
14:29
a multidisciplinary
14:32
diverse stakeholder team is key to good governance
14:35
so that we have the diverse perspectives. And
14:37
in health and hospitals, this includes legal, compliance,
14:41
IT, clinical, and business leaders.
14:44
And,
14:45
it's been going great.
14:47
That's amazing to hear. Didya, thank you so
14:49
much for joining us on the podcast today.
14:51
This has been such a informative and inspiring
14:53
conversation, and I look forward to connecting with
14:55
you again soon. And, also, you know, seeing
14:58
you in person in Chicago at our health
15:00
IT digital health revenue cycle event, in October,
15:03
I am really looking forward to your session
15:05
and certainly will be excited to to see
15:07
how these initiatives are continuing to evolve and
15:10
grow.
15:11
Oh, it's been absolutely a pleasure talking to
15:13
you to today, Laura, and, looking forward to
15:16
meeting you in person as well.