She Helped Design Pres. Biden’s AI Bill of Rights—How Does It Protect Patients?

This transcript was exported on Apr 03, 2024 - view latest version here .

Mark Masselli ( 00:04 ):
As we start a new year, we're also living in the brave new world of artificial intelligence and healthcare.
But what's at stake in terms of privacy and other civil rights?
Alondra Nelson ( 00:15 ):
It begins with a set of principles about the American vision for the use of AI. And driving innovation is
key to this, but we're going to do it in a way that is always upholding our values around the privacy of
patients in the healthcare case, that systems should be safe and effective and secure, and they shouldn't
be harming people, they should enhance their lives.
Margaret Flinter ( 00:38 ):
Acclaimed scholar and writer, Dr. Alondra Nelson says all Americans deserve an AI Bill of Rights, and she
took on the challenge of helping to draft the blueprint for President Biden. We'll ask for an update on
the promise and the challenges of making automated systems work to improve our health.
Alondra Nelson ( 00:55 ):
But it's also the case that some of the challenges that we're facing with regards to bias in AI are about
questions that it almost takes an outsider or somebody who's a little bit removed from the building of
the algorithm to help think through some of these challenges.
Margaret Flinter ( 01:12 ):
This is Conversations on Healthcare.
Mark Masselli ( 01:22 ):
Alondra Nelson, welcome to Conversations on Healthcare.
Alondra Nelson ( 01:26 ):
Thank you so much for having me.
Mark Masselli ( 01:28 ):
You served as the Deputy Assistant to President Biden, acting director of the White House Office of
Science and Technology Policy. I wonder if we could start out broadly by asking you to give our listeners
an overview of the president's executive order on AI, which was released on October 30th, 2023, and
how it deals specifically with healthcare.
Alondra Nelson ( 01:51 ):
So this is, I think, the most extensive and longest executive order in American history. So it gives you a
sense of just how important and significant this executive action, I think, is in the larger scheme of
things. It covers everything. AI is a general purpose tool that's going to affect all facets of life, and it is
the culmination of a couple of years of work at the very least. You could think about it as a longue durée
of work in the policy space, and it really attempts to offer a vision for how we want society, American
society, to both be deeply innovative but also use the tools in a safe way. So this, of course, includes the
national security space, this includes workforce and how we're going to not just allow the use of

Transcript by Rev.com

This transcript was exported on Apr 03, 2024 - view latest version here .

automated systems and AI just to take over work, but how government can help and communities and
industry can work together to shape how we want that to look and to augment that work.
[NEW_PARAGRAPH]And that work includes the work of doctors and nurses and researchers, biomedical
researchers, and the like. And then there's specific call-outs of course to Secretary Becerra and his team
at HHS. And the healthcare space is really interesting because organizations, agencies like the FDA had
been regulating AI as medical devices for a very long time. So I think there's probably lessons to be
learned. So it's asking, the executive order, for healthcare and agencies and the federal government and
the executive branch that use AI for healthcare to really lean into their existing powers and authorities
around healthcare, and then also to explore what the limitations of some of these tools might be and
what can be done on the research side and in the regulatory side to mitigate some of those.

Margaret Flinter ( 03:47 ):
Well, Alondra, you're really just zeroed in on one of the most interesting things going on in the country
today, and I thought one of the most interesting elements of the Biden approach is that you're trying to
learn from what you call the cautionary tale from social media regulation, noting that perhaps the
government did not quickly or quickly enough regulate it. Most of us probably understand what you
mean, but explain to us from your perspective what happened, is happening, in social media that you
want to try to avoid with artificial intelligence.

Alondra Nelson ( 04:22 ):
So first I would say that the challenge that we face with artificial intelligence is that as these tools get
more powerful and get more capable, issues that we're already facing become bigger. So the scale and
the velocity of artificial intelligence tools and systems to supersize the issues that we're already dealing
with, this is what we're facing. And so I think many of us, both you and Mark, and I think the listeners,
are already facing or concerned about some of these challenges. It's certainly what I heard in my time in
government. So concerns about young people, concerns about the mental health issues around young
people who spend a lot of time in the social media space, and partly why not only young people but all
of us spend so much time in that space is that we're served and sorted through algorithms that are
becoming increasingly powerful.
( 05:14 ):
Things that resonate with us emotionally, things that we want to see more of, and like a bag of candy we
might like it very much, but the things that we want are not always good for us taken to the extreme. So
that's one of them. Obviously, we've had challenges around misinformation and disinformation, an issue
in the healthcare space, obviously, that Surgeon General Murthy has been working on his entire tenure.
But what that means for the upcoming elections that are taking place in 2024, literally all over the world.
All of the world's biggest democracies will have elections in this year, and what does it mean to not just
have a couple of bots or a TikTok video disseminate misinformation? How does that potentially
amplified and scaled by a smart use of artificial intelligence to do that?
( 06:06 ):
And then we've got in the healthcare space issues of bias. And so the aim for all of this, why we're
working so hard, I think why there's an 111-page executive order, I think the longest ever, is because we
do want to be able to use this to help people improve their lives, these tools, including in the healthcare
space. But if there's evidence, as there currently ism that some of these algorithmic systems being used
and hospitals are not just biased in their use, but exacerbating, making actually not better but worse in
some instances, I'm thinking of course of the important work of Ziad Obermeyer and his colleagues. So

Transcript by Rev.com

This transcript was exported on Apr 03, 2024 - view latest version here .

those are some of the challenges that we already face. Some of them related to social media, some of
them related to... I'll put this in scare quotes, "less sophisticated" AI systems than the generative AI that
we've been living with for about a year.

Mark Masselli ( 07:05 ):
Yeah. I was thinking as you were talking about elections across the globe, I'm wondering how much
referencing you did to other government's activity in this area of regulation. Is there anything going on
elsewhere in the world that you were keeping an eye on? Obviously, we can do only so much here. This
is a global issue, and I'm wondering what your thoughts are about how this might either be an exemplar
for other countries, or did we learn some lessons from other activity going around the globe?

Alondra Nelson ( 07:45 ):
A great question. So we had the great fortune actually of coming into a conversation around AI policy
both in the United States and globally that had been taking place for a few years. So I think things
reached a crescendo in the policy space during the Biden-Harris administration and we're working that
through now. But in prior administrations, both in the Obama administration and the Trump
administration, there had been endeavors to work with the OECD, to work with an organization called
the Global Partnership on AI, to begin to build global benchmarks and a global conversation about AI
tools and systems recognizing exactly as you say, Mark, that all of these issues... I mean, part of the
beauty and the value add of social media in part is that it connects us in then powerful and profound
ways. And so all of these issues, of course, are always global.
( 08:40 ):
So that had already been taking place. There are also bodies that were stood up in the last couple of
years, like the Technology and Trade Council, the TTC, which is an EU and... excuse me, US effort that is
meeting at regular intervals to think about AI policy. And then most recently in the same week actually,
that President Biden issued his executive order on safe, secure, and trustworthy AI, the UN Secretary
General appointed a body called the High Level Advisory Body on AI, which is really tempting to move
from a set of principles. I think organizations like OECD, UNESCO, Microsoft, for example in the industry
space, had asserted over the course of many years ethical principles and guidelines for AI.
( 09:35 ):
And I think what's happened over the last year is that it's become very clear that, to use the metaphor,
we need to put meat on the bone. What are these principles in action? How do you operationalize them
in the different sectors in which our AI tools and systems will be used, including in the healthcare
sector? And how to get traction on that and to do that? And so that body was stood up on October 28th,
and I'm happy to be a member of this UN body. It's 38 people from 32 countries that are attempting to
come to some basic agreement on how we can responsibly operationalize these tools

Margaret Flinter ( 10:15 ):
Because it's very complex on a global stage, but even right here in the United States, we have a lot of
federal sub-agencies involved. We have FDA, we have the Office of the National Coordinator on Health,
just to note two of the most prominent. And I wonder from your expert perch and viewpoint, do you
foresee that we're going to need one overarching regulator for healthcare AI here in the United States,
do you think? I know that's asking you to forecast out, but it is very challenging to think about how this
works if we're coordinating across all the departments.

Transcript by Rev.com

This transcript was exported on Apr 03, 2024 - view latest version here .

Alondra Nelson ( 10:52 ):
I think my answer is, I think very much Margaret in your question, I do think we need coordination. I
think for me it's an open question whether or not we need an agency, a separate agency, because the
details of the domain-specific uses of artificial intelligence systems are so specific. So the rules and
regulations of HIPAA, the rules, regulations, and powers that FDA has, that other kinds of organizations
and sub-agencies at HHS have, versus thinking about the Federal Trade Commission or in the financial
sector. So in some ways, and you could imagine for example, that all of these agencies and sub-agencies
might be using generative AI.
( 11:38 ):
They might be using the same baseline tool. And so I can appreciate why someone would think, "Well,
we just need one big regulator." But the sort of granularity in which different agencies and offices will
use these tools and their particular use cases and regulatory apparatus that already exist around them
and might need to be developed, I think is distinct enough that what we probably need and can hope for
is better coordination across government, which is a clarion call, I'm sure, that you have heard many
times about many different issues on this show, and I think is only just made more acute by the power
and fast movement that we're seeing with AI.

Mark Masselli ( 12:22 ):
Well, in the interim, it's a little like herding cats. And I'm thinking about... and I'm not sure if the glass is
half full or half empty, but I'm going to take the positive side that at the end of 2023, there were 28
major healthcare companies that signed the president's voluntary commitments. It's a big industry. I
don't know if 28 is a number we should be happy for or not, but tell me what did they sign up for and
what gets covered in those commitments?

Alondra Nelson ( 12:56 ):
So what get covered in those commitments is some of a continuation of the earlier voluntary
commitments that President Biden was able to negotiate earlier last summer, in the summer of 2023,
with the large tech companies. And these are just sort of basic principles around privacy, around
providing some transparency in the systems that are used and accountability to the public about their
tools and systems, about ensuring that the tools, whether they're consumer products or algorithmic
systems used in hospital systems, are safe and effective. So the principles... one thing to say, Mark,
about President Biden's executive order is that it begins with a sort of set of principles about the
American vision for the use of AI and driving innovation is key to this, but we're going to do it in a way
that is always upholding our values around the privacy of patients in the healthcare case, that systems
should be safe and effective and secure, and they shouldn't be harming people.
( 13:57 ):
They should enhance their lives, help them have better lives and better health, and better workplaces,
for example. So those are the things that these companies agreed to, and I think it's important in the
regulatory space for us to really remember that this is really one of the... it's been the case in healthcare
for a long time, and so I think there's a lot to learn from the ways that algorithms and medical devices
that are algorithmic have been used and regulated thus far by the FDA. So I also think the healthcare
space is a space for learning for the rest of the federal government around these issues. And it's
important to say that, but what we need in this moment, because so much of this is driven by industry,
is of course for industry to make voluntary commitments. So I think to Margaret's earlier question about

Transcript by Rev.com

This transcript was exported on Apr 03, 2024 - view latest version here .

the social media space, so did we drop the ball, I think, at a moment where government and regulate...
we had perhaps a bigger window to sort of just do regulation around the use of algorithms?
( 15:02 ):
I think that is the case. I think more could have been done, but it is also the case that in the intervening
10 or 15 years that organizations like OpenAI, like Anthropic, like the sort of really frontier generative AI
laboratories really know the systems better than government ever will, and there's a particular need for
partnership there that may not have existed in other kinds of regulatory regimes or moments before. So
we need companies to step up. We need industry standards, we need industry norms, and I think the
voluntary commitments are a benchmark that companies that want to understand themselves to be
leaders in the space, responsible, and to be demonstrating that they are going to be accountable to the
public and want to be good corporate actors, I think can sign on to those. And that is a necessary partner
to regulations and other tools and levers that government uses.

Margaret Flinter ( 16:01 ):
Well, this is such a complicated area, but I am thinking we have decades and decades, certainly my
whole career, of learning about the bias in research when we are under-representing people of diverse
race and ethnicity and backgrounds. We're trying to tackle some of that with the NIH All of Us project
right now in real time, one of the largest studies going in the United States. But it seems that that really
has some lessons for us here. And I think you referenced this earlier, a recent study that found a clinical
algorithm hospitals were using to decide which patients needed care showed evidence in retrospect of
racial bias, deciding that black patients or showing that black patients had to demonstrate more criteria
for how sick they were than white patients in order to be recommended for the same care.
( 16:53 ):
That's a brief summary of it. Where is this avoiding the bias and health inequity that has characterized
so much of our healthcare system that I think there's so much commitment to change, but where does
the responsibility for that lie in the development of AI in clinical practice? Is that in the hands of
industry? Is that in the hands of the academics? Is that government's responsibility to ensure that
before these things come to market, so to speak, they've been tested from that perspective? What are
your thoughts?

Alondra Nelson ( 17:26 ):
It's shared responsibility, full stop. And it's shared in part because algorithmic systems, generative AI,
advanced AI, the really capable systems that we're going to be moving into are iterative. They change
and they're dynamic. So I think a lot of our thinking around regulation of algorithms if we think about
medical devices is that the object is static and you create a regulation for it, and then that's supposed to
be significant or to be mitigating of risk for the whole time, the whole use of the algorithm. But what we
know with advanced AI systems is that you will be continuing to bring data into the systems. You'll be
continuing to refine the algorithm. And so there's responsibility, I think, for all of those stakeholders, I
think for government, before these things are released. So has FDA done its due diligence before this
has even been able to be used in a hospital? Has a hospital system, has a vendor, has a purchaser or
procurer of this system done their due diligence around the tool and whether or not it works? And then
at the clinical practice level, is it working?
( 18:43 ):
We want to, I think... excuse me, certainly in these early stages, leave space for and must leave space for
this to augment the work of doctors, nurses, and other medical professionals and to leave them space to

Transcript by Rev.com

This transcript was exported on Apr 03, 2024 - view latest version here .

have their own judgment. And I think they will be able to tell whether or not an algorithm is being, I
think, accurate in some cases, even as the hope is that algorithms and the use of AI and medical care can
help us discern things that we can't do alone as human actors, doctors, nurses, researchers, and the like.
So I would say yes, Margaret, to your question, but it's both about the full lifecycle of the AI tool or
system that you need to be checking it. And in each of those moments in the lifecycle, there's a different
predominant actor, whether or not it's the federal regulator or the doctor or nurse that's using it. And
then of course the systems change over time and we've got to just keep an eye on that. It's one of the
ways in which having risk mitigation regulation for AI is different from some other medical tools and
devices.

Mark Masselli ( 19:56 ):
I want to pull the thread on Margaret's question about bias. And you've said you can't have good science
without diversity. We'd certainly say from the healthcare world, you can't have good healthcare without
diversity. But I'm trying to think about the workforce that we have and the scientists that we really need
to develop and cultivate. And there's a paucity of people who have PhDs going into this particular area.
But I'm wondering, are you seeing any good exemplars out there in terms of the cultivation of a diverse
workforce in the AI area, which will be another check on the sort of unfettered LLMs that are out there
that really need a number of different eyes looking at them?

Alondra Nelson ( 20:43 ):
Sure. I think we've gotten to a place certainly in government... I mean, in the work that I was doing at
OSTP, part of that portfolio was about really expanding the ability for a larger group of communities to
participate in the STEM fields. And here, I would pause to just honor the life of Congresswoman Eddie
Bernice Johnson, who we just lost a few days ago, who was just such a champion of this on the House
Science Committee. And that meant bringing people, as she would do, from urban communities, from
rural communities, because part of what healthcare, healthcare technology innovation, offers from a
government perspective are opportunities for people, for their lives, for their wellbeing, but also
employment opportunities to your point. I think we have significant data that more diverse teams are
better for innovation. Some of the things, the challenges around bias, that we're seeing in AI, when you
look back and think about them, in some instances don't need more PhDs in computer science. You
need, I think, diverse and multidisciplinary teams to be thinking about them.
( 21:52 ):
So in the instance of the case that the research, the... excuse me, Obermeyer et al research that we'd
been talking about, that algorithm it turned out was using prior healthcare costs as a proxy for illness. In
a for-profit system like ours, prior healthcare costs mean that you're actually well off and you have good
insurance, like probably the three of us. So you're going to be a higher user than people who don't have
use. And so I think if you had brought someone into the room and said, "Oh, we're using healthcare cost
as a proxy," who wasn't an engineer or wasn't an AI developer, somebody in public health, someone in
the social sciences might've said, "Well, gosh, this is what that proxy of cost is actually going to do in the
model."
( 22:42 ):
Similarly, we've got the work of a young woman computer scientist, Joy Buolamwini, Timnit Gebru, and
others who have demonstrated that some of the facial recognition tools, which have all sorts of
problems, but one could imagine them having some safe applications ultimately if they're tested in the
healthcare space, had not been tested on people with dark skin. So I think to your question, Mark,

Transcript by Rev.com

This transcript was exported on Apr 03, 2024 - view latest version here .

certainly we need... and I think Congresswoman Johnson would say, certainly we need like her more
African-American nurses, more doctors, more people working at the intersection who have RNs and also
some computer science training, these kind of interesting folks with multidisciplinary training. But it's
also the case that some of the challenges that we're facing with regards to bias in AI are about questions
that it almost takes an outsider or somebody who's a little bit removed from the building of the
algorithm to help think through some of these challenges.

Margaret Flinter ( 23:48 ):
Well, that is a very interesting perspective, and I'm going to stay a little bit with workforce generally for a
moment. You're a distinguished senior fellow at the Center for American Progress, and we know the
center has said that there must be a plan to address the economic and job impacts. And now we're just
talking about impact on the workforce generally from the use of AI, and CAP says we should try and
steer AI generation in a manner that complements and doesn't replace workers. So, great goal. How do
you see this playing out in the healthcare sector?

Alondra Nelson ( 24:24 ):
I think we use the levers that we'd often use in, I think, big manufacturing that you might use at, for
example, an auto plant or you can think of how we've been thinking about industrial policy around the
workplace and how that applies now to what we would call white collar workers, information workers,
expert workers in the space of health and technology, which is to say that to the extent that
government... so if you think about something like the VA system or the VA hospital, to the extent that
government and through executive action, if we're talking about that in particular, has the ability to say,
"We're going to create rules for how algorithmic systems are used in our hospital," one of the biggest
hospital systems in the world can help shape behavior and norms in this space. And that means we are
going to... there's often a case with some of these tools that you can... in generative AI, you talk about
turning the dial so you can have the dial give you... one way to thinking about it's more or less or less
mischief or more or less agency in the tool.
( 25:33 ):
And you could imagine thinking about that in the medical workplace as well, which is to say, we could
turn the dial on full automation for this radiological tool or for these other tools, but we actually think
that it's better for workers and it's better for outcomes to have an augmenting experience with these
tools. And those are really choices that we can make. You can make as a consumer, either a hospital
system or the federal government in the case of the VA, about how you want the tools to participate in
the workplace. Those are decisions that can be made. So that's just one example. You can certainly think
about workplace training. Are there ways that government or industry can partner? There was a
recently announced partnership between Microsoft and the AFL CIO around participating together and
creating the technologies and ways that include workers and their perspectives and the way that the
tools are built before they're deployed.
( 26:32 ):
So I think there are creative ways for us to think about this. And it's also the case I mentioned the
radiology example, which is one of my favorite ones, which we've been hearing for not quite a decade,
six or seven years, that radiologists were over, that the whole field is over, it's going to be totally
automated. And of course, we know that that's just not come to pass. We need more radiologists than
we actually have. It's certainly the case that radiologists working with excellent AI systems and tools can
often do their work better, but you really still do need their discretion in getting the optimal outcomes.

Transcript by Rev.com

This transcript was exported on Apr 03, 2024 - view latest version here .

Mark Masselli ( 27:14 ):
Critical thinking continues to be an important role.

Alondra Nelson ( 27:14 ):
Yes.

Mark Masselli ( 27:17 ):
I want to get your crystal ball on what might be happening with this legislation. Advocates say one of the
goals of the Biden-Harris administration was to really spur Congress, which seems not to be functioning
at the highest level, to draft and pass AI legislation. I'm wondering, is that moving forward, and as I said,
Congress still seems to be at a standstill on some basic fundamentals, but is there a hope that there's
some bipartisan consensus that can come about around AI regulations?

Alondra Nelson ( 27:51 ):
Well, I think I'd want to mention here that there's been effort. There's been effort in the 117th Congress.
There has been effort in this current Congress, in the 118th. Obviously, Majority Leader Schumer has
been working in a bipartisan way. There have been several... I think, there've been nine now of these AI
insight forums. I was happy to participate in the second one on innovation and tried to make the case
that one doesn't have to... it's not a zero-sum game between innovation and tools that are safe and
responsible and secure for my part. So there is a good faith effort, and it is some of the leaders and
some of the policy thinkers around science and technology leading this effort that help to get things like
the CHIPS and Science Act over the line. So you've got Senator Young working on this as well.
( 28:39 ):
So to the extent that there's any hope for moving forward, any legislation, Mark, I hear and I think share
your pessimism. I think often given now what we're a couple of weeks out from potentially another
government shutdown, I mean, it's the calendar. There's not a lot of days left in the legislative calendar
for this Congress, and so I'm not optimistic. I would think I probably... it's glass half empty for me on this
one, but there there is a recognition that this is a bipartisan issue that lots of people will care about and
would benefit from doing innovation in a responsible way, that there are US companies, US innovators,
US researchers that are really some of the world leaders in this space. So there are lots of incentives for
Congress to get going on this, and I try to be hopeful, but I think, Mark, it doesn't look great.

Margaret Flinter ( 29:38 ):
All right. I'm going to take us back to something maybe a little more hopeful, and that's thinking about
the potential for AI in addressing the behavioral health needs of the people of this country and really
worldwide, certainly from our perspective in a very large, integrated primary care system where
behavioral health is a core component. We've seen the mismatch between the numbers of behavioral
health staff that are coming through the training and the demands of the population, and AI NBH is not
so new. I remember, perhaps Mark, out of England is looking at presentations as much as 10 years ago,
avatars that were there. That technology has no doubt moved on hugely with AI. Do you see this as a
potential to provide some very meaningful support in the country for helping to meet the behavioral
health needs of the population and perhaps freeing up the scarce resources for those who need it most?
What are your thoughts on that? Do you think people will accept that?

Alondra Nelson ( 30:43 ):

Transcript by Rev.com

This transcript was exported on Apr 03, 2024 - view latest version here .

Yeah, I think so. I think it's a really promising avenue. And as I've been sitting in these policy spaces, both
domestic and international, thinking about the potential future uses and often posing the question,
what is the best case scenario for how we might use AI if we can get it above a threshold of privacy and
safety and security and trustworthiness? And this is a use case that comes up quite a lot because it's
becoming, as you probably know better than I, the data on mental health challenges and crises globally
is staggering, actually. And certainly in the US, we know through the pandemic and after, there's been
lots of challenges. So it does offer the opportunity to really scale mental health resources to lots of
different communities, to do it fairly inexpensively to the extent that a lot of the challenges are with our
young people.
( 31:44 ):
They're much more comfortable using their telephones, with... they're much more comfortable with
avatars. They're just digital natives plus plus. And so I think that using the space of AI for this is great. I
guess my caution, the thing that I'd want to caution, is that we shouldn't have to give up any of the
protections that one would want to have or expect in a human clinical encounter. So we really have to
be careful about protecting people's privacy and how are we thinking about not just the privacy and the
use of the tools, but what if there's a data leak of the company that is hosting or sponsoring these
particular platforms? I would really like us to think smartly and really hard around those implications.
There's obviously cybersecurity implications. Can people get access, malign actors, to files about
people's mental health status?
( 32:42 ):
And then there's been some concern, particularly for colleagues who are worried about potential future
risk with generative AI systems, about things ranging from addiction to emotional attachment to AI tools
as they get more compelling and as they become, I think, more... we make them more human-like in the
way they interact with us. So I think we want to think about how to use these tools at scale to ring the
care that people need. And the promise here is tremendous, but also do it in a way that doesn't, I think,
violate privacy, that depending on the system and tool abides things like HIPAA to the extent that that's
appropriate, and also doesn't exacerbate by making people dependent in a way on the particular
intervention and help that they're getting. So it's a complicated space, but I think a space with a lot of
potential for helping people.
( 33:43 ):
And I would just want to underscore that. Why we are engaged in these regulatory battles, why people
are spending a lot of time thinking about AI policy, is because the potential is extraordinary, and we've
got a wonderful opportunity in the early stages to get it right. And so I think, Margaret, to your earlier
question, having not quite gotten it right with social media, it's a great opportunity to get it right and to
really leverage these tools to catalyze much better outcomes for people, which is ultimately with my
former government hat on, the job of what government's supposed to do and how I think government
policy makers think about the role of science, technology, and medicine in particular.

Mark Masselli ( 34:34 ):
Dr. Nelson, thank you so much for joining us today for this fascinating conversation and for your
leadership in this area. We really appreciate you joining us, and thanks to our audience for being here.
As always, you can go online to chcradio.com to sign up for email updates. You can also share your
thoughts and comments about this program. Thank you again so much, and again, best wishes for a
happy and healthy new year.

Transcript by Rev.com

This transcript was exported on Apr 03, 2024 - view latest version here .

Alondra Nelson ( 34:57 ):
Thank you so much.
Margaret Flinter ( 35:02 ):
This copyrighted program is produced by Conversations on Healthcare and cannot be reproduced or
retransmitted in whole or in part without the express written consent from Community Health Center,
Inc. The views expressed by guests are their own, and they do not necessarily reflect the opinion of
Conversations on Healthcare or its affiliated entities.

Transcript by Rev.com