November 17, 2024 • 42 mins
What happens when frustrated doctors decide to fix healthcare technology themselves?
In this fascinating episode, we dive into the world where medicine meets code with two remarkable physician-entrepreneurs who are tackling one of healthcare's biggest headaches: broken software systems.
Meet Dr. John Manning and Dr. Grey Faulkenberry, who split their time between saving lives in emergency rooms and building MayJuun, a startup that's revolutionizing how medical data moves between hospitals. They share raw insights about why doctors are drowning in after-hours paperwork (known as "pajama time"), and how this digital burden is pushing medical professionals to their breaking point. But this isn't just another tech story. It's about doctors building for doctors, understanding the pain points from the frontlines, and creating solutions that actually work in the real world of medicine. Our guests break down complex topics like FHIR (healthcare's data standard) into digestible insights and share their vision for a future where technology enhances rather than hinders patient care.
Key Conversations:
#StartupsDilSe #HealthTech #DigitalHealth #MedTech #StartupStories
In this fascinating episode, we dive into the world where medicine meets code with two remarkable physician-entrepreneurs who are tackling one of healthcare's biggest headaches: broken software systems.
Meet Dr. John Manning and Dr. Grey Faulkenberry, who split their time between saving lives in emergency rooms and building MayJuun, a startup that's revolutionizing how medical data moves between hospitals. They share raw insights about why doctors are drowning in after-hours paperwork (known as "pajama time"), and how this digital burden is pushing medical professionals to their breaking point. But this isn't just another tech story. It's about doctors building for doctors, understanding the pain points from the frontlines, and creating solutions that actually work in the real world of medicine. Our guests break down complex topics like FHIR (healthcare's data standard) into digestible insights and share their vision for a future where technology enhances rather than hinders patient care.
Key Conversations:
- Why current healthcare software is failing both doctors and patients
- The real cost of poor tech on physician mental health and patient care
- How MayJuun is building doctor-friendly solutions that just work
- The open-source revolution in healthcare tech (FHIR-FLI)
- What's really happening behind the scenes in hospital IT systems
#StartupsDilSe #HealthTech #DigitalHealth #MedTech #StartupStories
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:00):
This is the Startup Still Say podcast.
Speaker 2 (00:02):
Thank you for tuning in favor subscribe over LinkedIn and
be sure to give us your predom. I hope you
enjoyed this episode.
Speaker 1 (00:11):
Hello everyone, this is Anthony Prakashi, your host for Startups
They'll Say. Here we are with another brand new episode,
and I'm super excited today to bring to you a
couple of my friends. We were colleagues or partners in
crime a couple of years earlier, but we are all
members of a flood of community right now and I'm
(00:34):
excited to bring to you their story, their mission and
vision for something super important in the healthcare world, both
in the US as well as the entire world that
they're trying to solve, and it is Their company is
called May June. So John, doctor John Manning here is
the CEO of May June and doctor Gray Fuckinberry is
(00:56):
the CEO of May June. So if about anything, I mean,
they always inspire me because both of them are doctors
by day and sometimes by night, and then the rest
of the twenty four hours, they are technologists and hardcore technologists.
I must say so a warm welcome John, and great
(01:16):
to startups Bills. Why don't we kick it off maybe
with John. You can go first and tell a little
bit more about yourself and your journey of how you
started off me.
Speaker 2 (01:26):
Yeah, sure, and thanks for the introduction Anthony. We've gone
a long time back, so it's great to have the
opportunity to speak here. So I'm John Manning. I'm an
emergency medicine physician by training. I teach medical students and
residents how to care for patients in one of the
Level one trauma centers in Charlotte, North Carolina, and have
(01:49):
done that for almost a decade at this point, getting
close to that point right before coming down here, like
right around that time, I co founded May June with
an interest in building technology solutions that focused on human factors,
usability and just trying to make your life easier from
(02:10):
an end user perspective, the kind of short version of it.
And we can go in depth if interested. But I
used to be ascribed in a medical field in DC.
I hand wrote notes helping emergency positions. Was there when
we transitioned from paper to electronic saw the equivalent of
twenty different health systems across med school, residency, fellowship, etc.
(02:33):
And all with different levels of medical record integration and usability.
A lot of them had a lot of usability challenges,
and my attempt to give back was in building tech
solutions through made focus on that.
Speaker 1 (02:47):
Awesome. Great, why didn't you do the same? Just introduce yourself?
Speaker 3 (02:52):
Great and thanks again for having us. I'm excited to
be here. I've been looking forward to it, even though
I responded like so, I'm Gray Falconberry. I'm CTO trained
in internal medicine and pediatrics. I practice as an adult hospitalist,
and my other part of my day job is I'm
(03:13):
the co chair of the Global Health Informatics Program at
the Children's Hospital of Philadelphia, although I actually live in Atlanta.
That's one of the joys of technology and informatics is
that so much of its virtual. You know, you don't
have to be in the place where you practice, which
is really nice and allows flexibility. And I there's a
longer story involved, but I got into this from the
(03:35):
global health perspective actually and looking at how you can
sort of leverage technology abroad to improve healthcare.
Speaker 1 (03:43):
That is awesome and it's a super worthy cause. And
like I've heard from both of you, I mean, it's
it's obviously even in a country like the US. It
is informatics and the amount of manual effort that goes
into the work aside from you guys, care for patients
I mean is a big part of every doctor's day, right,
(04:04):
I mean, and there's so much of others involved in
it as well. So let me ask you this, I mean,
where did you kind of get the h to say, Okay, man,
this is a big problem. We've got to do something
about it. I mean who started that? I mean, I'm
sure both of you have to think about that vision
and believe in it saying okay, we can do something
to do it. But John, I know you started this
(04:27):
and Gray probably joined later. So what did you tell
the audience about? I mean, where is it that you
noticed this? I mean, I know you said you were
ascribed maybe there, but tell us more.
Speaker 2 (04:38):
Yeah, So it all kind of comes back full circle.
And when you're just trying to think through like how
you can improve and where those options are, you mentioned
usability being a big issue for physicians. Christine Sinski from
the American Medical Association published on that there's a thing
called pajama time, which is the time you spend at
home just charting and catching up on your documentation, and
(04:59):
every specialty has it. It's not fun, it causes burnout.
It leads to burnout, so it's everywhere. The specific path
that led to May June, interestingly was as a resident
in Emergency Medicine, I pitched an idea to the first
ever Emergency Medicine hackathon at their national conference. Got the
(05:22):
grand prize, got a bonus phase prize, and an opportunity
to speak at a public health summit in part of
the keynote in Illinois. So we built an app that
was based on the hackathon idea. Wrote it in Java's
first time that we'd ever written any sort of code.
It was a Java developer friend of mine who had
not written anything in Android before, neither had I. He
(05:45):
lived a few months away or a few blocks away,
and the better part of my final year of residency
training was building this application. All through fellowship, got a
lot of opportunities to work with other startups and do
a bit of mentoring and working with a group called Matter,
a health tech incubator, and time unfortunately broke a little
(06:09):
bit of the app, and when we were trying to
think through am I going to do this in Cotland?
Because Cotland is new. Am I going to switch over?
Am I going to just update this in Java? That
was almost exactly the time that Flutter was announced, so
December of twenty eighteen. The next month, I was in
Boston giving a live coding session, handshaking for five minutes,
(06:32):
showing how you can really make this because it just
fell in love with Flutter. A few months later, it
was a big informatics conference at their annual symposing or
their clinical Informatics conference, and in two hours introduced Flutter
and live coded how to create the conference's app in
real time while just talking and writing code. And one
(06:57):
of Gray's predecessors, his co fellow one year ahead of him,
tap through that conference. It was like seventy eighty people
in that two hour event. And when he joined Informatic
like that Informatic fellowship a few months later, he was
interested in global health and maybe Android or you know,
trying to build and develop for this, and she looked
at him, she was like, well, why don't you just
(07:17):
look into Flutter? And he started looking into this, and
there was a shared Slack channel with a lot of
Informatic fellows that were going through this new training, and
we started collaborating, collaborating every day or slack over years
and started building these solutions that would then allow other
(07:38):
people that are in and around the health tech space
to use the standard for interoperability FIRE within Flutter applications,
and much of that goes to Gray and all of
the time and effort that was spent in this, I
was I just make apps and Flutter and I live
code and I can talk and you know, it's the
emergency medicine brain. I can talk and write code at
the same time, you know. And that's my thing. But
(08:01):
a lot of that underpinning and architecture that's open source
and has been used by developers a lot across the
world that goes with many things to correct.
Speaker 1 (08:10):
That is awesome. John, So you touched on a lot there, right,
I mean you talked about fire obviously informatics and great,
maybe you can wear the CTO had I've always wondered.
I mean, you know, you guys have very little time,
yet you've found the time to learn a technology like Flutter,
(08:31):
decide and pick on it saying Okay, this is the
best choice to move forward with what we're trying to build.
I mean, for I mean, obviously you don't need to
be an engineer to learn flutter and use flutter anymore.
I mean you are perfect examples for that, right. But
how did that love for even technology come in to
both of you? I mean, obviously you need a spark
(08:53):
within you, you need that push within you to say, Okay,
this is not something that I can outsource to a
company and say build it. I want to do this myself, right,
I mean, how did you sort of make that choice
great to say okay, I mean, you guys want to
be you know, getting your hands and feed wet with
building this whole thing.
Speaker 3 (09:13):
So I my love of technology and sort of science,
I think at John's two, you know, is not not
a new phenomenon we liked growing up. In college, I
was a physics major and I'm minored in computer science,
so I had a little bit of a background, but
then I didn't really use it. I went to medical
school and was in medical school when EHRs really started
(09:36):
coming around and sort of everybody switching over from paper
to electronic and so I saw that transition, and at
the time it was exciting. I was looking forward to
it because again I like technology, and at the time
it seemed mostly good you know, the not having to
fight for the chart, having one chart for a patient,
having eight people trying to look at it. You know,
you didn't have to worry about that anymore, having to
(09:57):
walk across or down fifteen floors and across campus to
write an order and a chart. You could just put
it into the computer. So there were some really nice
things about it, and then there were a lot of
not nice things about it. It was obvious that, as
John said, the usability was not a sort of forefront
in anyone's mind. You know, with the design of these
(10:17):
and any sort of study or survey done on physicians
now you'll see that. You know that there is burnout,
and so you work with that long enough and it
becomes a source of frustration because it really felt like
the people who designed it hadn't ever seen a patient.
And that was probably true, you know they and you
(10:39):
really need to understand the workflow. You need to understand
how it feels and what you need and how it
should be organized to take care of, you know, a patient.
And I think what drove a lot of us into
informatics is the feeling that we want to be involved
in this, We want to help make this better. This
is something. This is a skill set that we feel
like we have and we can develop and we can
(11:00):
actually contribute and sort of change things for the better.
So much of medicine is already sort of set in place,
and there's there's so little opportunity to sort of change
any of it. But informatics, I think is different. It
is still new, it is still changing. It is a
lot of it is sort of set, but a lot
of it's not, I think enough of it's not that
(11:24):
we can sort of affect that change. And I think
that's what's exciting about it is the feeling that I
actually have the ability to hopefully, you know, help medicine
and healthcare be provided in a better way through technology,
you know. And that was sort of how informatics I think,
came into my life. Yeah.
Speaker 2 (11:44):
So if you think of like as an example in
medicine of something that's never changed and it will never change.
Coagulation cascade, which sure whatever, Okay, you don't have to
know about it. One of there's a dozen factors. One
of them is literally just calcium and it was named
its own factor. Sure, and another one factor six. They
(12:07):
found out, oh actually it was just another version of
factor seven, and so there is no factor six. There
isn't one. So it's labeled up to a dozen or
a thirteen, but you're missing a whole number. And since
we've all learned it this way, I'm sorry you have
to learn it that way too, even though it makes
no sense. Welcome to healthcare. Here's a chance where we
can actually build things that match the workfload that's there.
(12:31):
To Gray's point, if you don't see it, if your
brain isn't literally into the weeds, and it's not there's
one thing to be shadowing somebody who was doing this,
there's another where you were putting all of your cognitive
effort and load into the care of a person. So
you don't make those mistakes, and so that you know,
medicine always has mistakes, but so that you minimize that
(12:52):
and yet you allow for the right processes and everything
to help. Well, if you have a very expert, complex
system that's pulling all of your attention away, you're gonna
get distracted and things will happen, and it's you know,
it can lead to it can lead to harm.
Speaker 1 (13:07):
If you.
Speaker 2 (13:09):
What was the the book that was talking about the
digital doctor from like ten years ago or ten to
fifteen years ago, talks about all the processes of how
a fifteen year old child swallowed thirty eight and a
half pills of a single antibiotic and seized, and all
(13:33):
of the workfload things, as well as the medical issues
that were a part of your electronic health record that
was almost like a Swiss cheese effect where everything was
perfectly aligned to a point of a nurse in a
room with a fifteen year old by himself, with mom
or dad not nearby, as he is individually swallowing almost
(13:54):
forty pills. There's a lot to it. So any ways
that we can improve that medical perspective I think is helpful.
Speaker 1 (14:04):
Yeah, And speaking of a medical perspective, John, I mean,
I know both of you have touched on what you're
trying to make better. But if there is someone who
just is a layman in layman's terms, if you want
to explain before May June's in an after main story
of course, using fire which is fhir for those listening,
(14:27):
how do you explain the work that you guys are
doing to make this thing better?
Speaker 3 (14:34):
So I could take this one. So part of it,
I think is something that we're all familiar with. And again,
in no way as a clinician, I think it's a
patient you see it. We all go to doctor's appointments
or healthcare appointments or therapy appointments where you would like
to be able to move records around. You know, you've
seen someone else before you would like to share those
(14:55):
records with someone else. And it's always painful. It's always
extremely hard. You have to fill out release sheets and
they have to often fax them over the fact that
facts comes into conversation at all anymore is ridiculous, but
it does in medicine all the time. You know, I
regularly get two hundred pages of faxes of someone's medical
(15:15):
record from another hospital, you know, trying to take care
of it, because medical records don't share data very well.
And then about not quite fifteen years ago now, there
was a standard that started being developed called fire Fast
Healthcare Interoperability Resources in order to share it. And again
(15:39):
that's a big deal because just having that information is
so important. So the ability to easily share healthcare data
you know can is not just a convenience issue, it's
it's a safety issue. If I don't know what medications
someone's taking and I give them something that interacts that
obviously can be incredibly dangerous. You know, I don't need
(16:01):
to do certain tests if I know that they were
just done somewhere else. And so having the ability to
more easily share that data from place to place is
incredibly important. And I think it helps both from sort
of again a frustration and a time sort of commitment,
as well as a pure safety issue. And that's something
that again we've we've sort of touched on. As John
(16:25):
pointed out, we both sort of lean slightly different directions
on this, but again trying to make electronic records that
work better for the people involved. So we want to
be able to share data, We want it to work well,
we want it to be easy to use, you know.
And and so I think both the sharing of the
data and the usability can't be overstated about sort of
(16:47):
how important those aspects are.
Speaker 2 (16:49):
Yeah, adding to that a little bit, and when you
think of fire, so FIRE is now the global standard
for interoperable communication of healthcare data period adopted by the
US governments. For that large electronic medical records, they have
to support at least a core subset of fire data
that patients can get access to. And no cost, and
(17:12):
when you're trying to communicate this data, it is miles
beyond what was there before before the Fire standard. What
kind of happened through all of this sorefly Fire with
Flutter Library Integration FLI, where the packages that Gray had
developed over the course of his fellowship, which is about
(17:34):
ten million lines of code total three million lines of dart.
A lot of it's auto generated, but in a way
that allows you to use any published version of fire
that also goes through all its test cases and then
directly put this into your Flutter application. And that Android
equivalent of this is open heal Stack by Google for Flutter.
(17:55):
If you're building a Flutter app, you should use the
Firefly packages. So you know. To Gray's point, so he's
a medicine pedes trained hospitalist, so he cares for patients
when they're admitted to the hospital, adults and children. Practiced
in eight countries, spend a year in Uganda, wrote the
textbook on pediatric care in that setting, and has a
(18:19):
very high global health focus of trying to help in
areas that are very resource limited. My background is emergencies.
I care for emergencies. I did MS things and a
lot of the things that come around me as an
emergencies and disasters, and how can we improve in those cares,
those care settings that are unscheduled and where a lot
(18:40):
of people are afraid, And both of them have opportunities
to improve in rural areas as well as global. So
to your point that you had mentioned earlier, Anthony, of
what was our story, well, we both had a shared
interest of building things to help in healthcare from usability
perspect with a mobile first approach. Then collaborated on the
(19:04):
Firefly packages, reached out to the Office of the National Coordinator.
The actual National Coordinator at the time was Don Rucker,
who we got through other members of Nyjune. They had
connections and here we are in front of the ONC
talking about Firefly and how cool this is, and they said,
(19:25):
look into social help like social care areas, and so
that is where we built screening tools to help with
social need and multiple states and multiple languages, and we're
still on the Gravity Projects. Main page is their success story,
which is wonderful and cool. But you know, as clinicians,
we're always trying to see how best we can apply
(19:48):
this in a way that helps in rural, underserved and
in global areas.
Speaker 1 (19:54):
Yeah, so you mentioned emergency medicine. One of the things
that always bothered me about, you know, going to a
hospital in the US, even if you go to an emergency,
is it's never an emergency in an emergency, right, I
mean you'd go with you know, so much pain, and
you'd be like, want to be seen immediately. You're invariably
waiting for sixty to ninety minutes before you've seen or
(20:17):
at least somebody comes and asks you something, right. I
mean in a lot of cases it has got to
do with data collection. You know, they are validating it
behind the scenes, so you know, just as a patient
myself or you know somebody who I've seen, even in
the children's hospitals, great, right, I mean when my daughter
was young, I mean, you know, she'd be running very
(20:39):
high temperature and you go there, they make your sit
for forty five minutes, right, I mean why because you're
still collecting data. So in that sort of a let's say,
I mean you don't want to call it life threatening,
but it's still an emergency. Is there something that you know?
Using fire in May June, I mean you can simplify
that process to accelerate the time to which by the
(21:01):
time you walk in through that door, a doctor is
able to see you at least within fifteen minutes or so.
I mean, you know, if sixty minutes is too much,
at least cut that down by yeah, you know, seventy
five percent.
Speaker 2 (21:14):
So yes and no. So I think on the other
side of the curtain. So Atrium Health is where I
practice associate professor, and I care for adults and children,
and it's patients are not seen in the same order
that they arrive. They are seen often by the sickest
ones first. So thankfully you and your family were not
(21:36):
that sick. There's an emergency, of course, we'll address that.
But trust me, the very very like near death patients
and severe traumas and medical and traumatic resuscitations, they're seen
really fast. It's not sixty minutes. And there are workflow
tools where you can even register a patient without any
information quickly, be it medical or trauma, so that you
(21:59):
can give them medication and treat them and get do
an ultrasound or get some X rays or something at
the bedside to help them. All of that stuff has
its own workflow in American medicine that uses the electronic
medical record and the use of majune technology probably isn't
(22:20):
as relevant there if you look in resource limited scenarios.
So another example, I helped with some of the hurricane
relief efforts in western North Carolina where it was multiple
groups together and some of us were using paper and
some of us were using another technically medical record. That
(22:42):
will also wasn't really designed by a clinician, and it
was clear that there were a lot of opportunities to
improve with that. That's a place where we can inject
and help and where we have the skill set and
the expertise and the domain knowledge to improve in those settings.
And we've had conversations with groups that were interested and
our services from there. So from the over arching perspective
(23:05):
of you know, just emergency medicine care, usually it's going
to be in the low resource or disaster response or
international area where we can improve rather than the traditional
brick and mortar emergency department. But medicine always changes, you know,
if we build something and then it expands and then
(23:28):
it continues to expand. And that's right up both Gray
and my alley of ways that we can improve in
areas that also build on the background expertise that we have.
Speaker 3 (23:38):
That's perfect, and I'll say a slightly different sort of
view of it. I again not sort of the emergent
view in terms of how quick we can get you in,
but I think that some of the stuff we do
can make the time we spend with you more valuable.
And I think that you know, most of the time,
the about of time that we spend with a patient,
most of it is collecting data and very little of
(24:01):
it is actually spent talking to a patient. And so
I think using technology, including some of the June you know,
questionnaire data gathering, structured data capture sort of stuff, a
lot of that could be done beforehand. And so if
you have a fifteen or a twenty minute appointment, or
you know, whatever John has available to be able to
see you in the emergency room before the next trauma,
(24:25):
you know, instead of spending all that time or the
bulk of it trying to gather all your information, make
sure that we have all of your medications correct, that
we know all of your history. If that can all
be shared or gathered more efficiently ahead of time, then
we can sort of review it quickly and then we
can spend more time talking to you. We can spend
more time ensuring that we didn't miss anything that's important.
(24:48):
We can ensure that you understand the plan, and so
the time that we do spend with you can be
actually spent with you and so sort of caring about
what the problem is instead of just asking you a
whole lot of questions. And so I think there are
some ways, even in sort of a traditional setting where
(25:09):
you know, we could be beneficial that. Again, I think
there's lots of opportunities to sort of improve that.
Speaker 1 (25:15):
Yeah, thank you for that.
Speaker 2 (25:15):
I mean the term doctor, go ahead, John, Oh, go ahead.
The term doctor comes from dose air, which means to teach.
It is literally your role is to map somebody's health
literacy and then help explain and work through that. And
the other point from Gray is if you've ever seen
there's a twenty twelve article in Jamma called the Cost
(25:41):
of Technology by Toll and it was a drawing of
a seven year old in their pediatricians clinic where you know,
they're smiling, Mom's smiling, big sisters holding baby child like
baby sibling. Everybody's smiling. Then the doctor is smiling but
has their back turn and a staring at a computer.
And this is over thirteen years years ago when that
(26:01):
was done, and it's still relevant today.
Speaker 3 (26:04):
And again I think that's again to dive back into
the technology. I think that's where Flutter is so nice
and actually part of the reason we chose it as
the health stack. Again, I was looking at sort of
a global perspective and everything global it has to be mobile, right,
you know, it has to be at least available, and
most EHRs or not most of them were never designed
(26:25):
to be and so you know, we're looking at what's flexible,
what can be done, you know, and multiple platforms, multiple devices.
And that's sort of what sold me on Flutter because,
as John said, one of my colleagues said, well, she
had seen this talk about Flutter. Maybe I should look
into it. So I did, and I got on Slack
(26:47):
and I started working with John. And the punchline to
his story was that it was about two years later
when I realized that the speech my colleague had seen
was given by John that then got that, then got
the end of flu. So it was yeah, it's a
full circle.
Speaker 1 (27:03):
Yeah, full circle indeed. And I mean, thanks for sharing
the procedure about how the emergency rooms work, right, I mean,
I know everybody comes in there and there's nothing more
frustrating than waiting, right, so all your anger levels go
up and you see a lot of people walking up
(27:24):
to the poor receptionists and saying, I've been here for
ninety minutes. You know why is nobody seeing me? So
that part is super helpful. So shifting years a little bit,
John and Greg, obviously you started May June with a vision, right,
how much of that have you achieved? And what are
the next couple of years looking like for you? Guys
with May June, either of you can take it.
Speaker 2 (27:47):
So I'll say that our vision always shifts forward. We're
always working to build new things, better things. Everything we've
built in May June has been contracted work for hire
to satisfy a specific problem that needed some extra usability.
(28:07):
Flare that was cross platform by design and frequently had
one group or one solution that didn't match the initial expectations.
So a good example of the first thing we built
was for any MS agency that had protocols to help
in rural areas that the initial the content was excellent,
(28:30):
that was something they built and curated, but the delivery
needed help, and so we helped and you know, we
continue to help and support them for years. We're always building,
and we're always doing this in a way that can
scale into other solutions and can apply some of the
(28:50):
backgrounds and experiences that we have as as practicing clinicians
and informatis and we have other informatics clinicians that are
a part of MADE as well, and in addition to
more tech first people as well. I think for us,
we're always trying to find that best opportunity where we
(29:11):
can have the biggest impact on the broadest scale in
a way that pushes things forward in a good way.
And I think that's always where we've strived. And so
are we there yet? I'll say no, Are we always
going down that path?
Speaker 1 (29:25):
Yes?
Speaker 2 (29:26):
And you know, it's one opportunity after the next and
one solution after the next as we learn and make
our code better and as we provide that service of
Firefly for the global community, in addition to building upon
that and harnessing it with the tools that we that
we then deploy got it.
Speaker 3 (29:47):
Yeah, Well, I think that a lot of what we've
done so far was still sort of foundation building to
an extent. It was sort of learning and building out
our skills building out. I mean, some of it was
certainly sort of technology building out, the tech stack and
that sort of thing, and some of it was also
sort of finding where we wanted to be. You know,
(30:07):
We've done you know a number of different projects in
a number of different places, sort of in and sort
of slightly out of healthcare, always intertwined, you know, And
so I think we're we've sort of found our niche
and sort of where we want to be. And I
think that's again it's sort of exciting, and a lot
of times it's places that other you know, groups don't
(30:29):
want to go. And again, sometimes there's a lot of
times there's a monetary barrier to it, you know, whether
it's sort of rural areas or low resource settings or
disaster relief one of the ones that that again John
brought up earlier. You know, these are all really important
to have good health tack tech stacks for if you're
going to actually collect data and be able to show
(30:50):
what you're doing, you know, who you're treating. You know,
it's important to be able to record all that. It's
really hard, though, but I think you know a lot
of that fit what our interests and what our passions
are and our knowledge and our expertise, and so I
think we're sort of excited and looking forward to being
able to use you know, who we are and what
(31:12):
we want to do it, you know, even more effectively
as we move forward.
Speaker 1 (31:17):
That is great to hear, especially the part that you're
always kind of looking forward, right. And I try to
bring up AI in most of our conversations because I mean,
it seems to be affecting everything I mean, and especially
I mean, I've heard in healthcare there's a lot of
use cases and John, you started off with scribing. I've
heard a number of you hospitals using those livescribers sort
(31:41):
of thing. But with your vision for me and kind
of looking forward and the impact that AI can make.
Or the other trend I've been hearing about is you know,
taking the hospital to the home, right, I mean, if
there's critically ill patients, I mean, you're moving a lot
of equipment into the patient's house and still collecting data
(32:03):
and things like that. Are there's a lot of you know,
devices that the patients can carry home, but as a doctor,
you still want to collect all that information. So looking
at some of these new trends, I mean, what are
your thoughts on you know, how May June will evolve.
Obviously you would have given that thought. I mean I second,
maybe a second part of the question is, I mean,
(32:25):
how much of AI have you actually tried to bring
into the whole tech stack and try to do something
with it.
Speaker 2 (32:31):
There's a few things from that. So the first one
is AI from the perspective of the clinician and the
second is AI from the perspective of the developer, and
they're totally different buckets. So let's take the first bucket. So,
first one, I've given local talks, given talks at Amya
about this, and I've tested this and I also co
(32:52):
chair our service line Innovation and Informatics group with an Atrium,
and we've tested this from the very beginning and with frequency,
there is a difference between So as a scribe, you
mentioned this, there's a difference between a good scribe and
a bad scribe, and a good scribe speeds you up.
A bad scribe has things interjected that gives you almost
(33:13):
twice the amount of time that you would be using otherwise.
Tech always changes. Even a year ago, there were a
lot of bad scribe ambient dictation tools out that a
lot of it's improved even in the last eight months.
And if you want something to look at, there was
a AI tech sprint from the VA, the Veterans Administration
that was hosted at the start of this year and
(33:34):
then announced early Q two, and if you just search
for the AI tech sprint within the VA, you can
see some of the companies that were well embedded in
that perspective. I think ambient dictation is going to have
a high opportunity to help clinicians caring for patients, albeit
in limited resource settings. That's going to be a challenge still,
(33:55):
just because all of these need a lot more cloud
services of a level to them, So that's going to
be your mileage made very depending on where you practice,
but there's definitely opportunity. The other aspect for decision support
from the clinician's perspective, I think is great. What used
to be called clinical decision support is now being rebranded
(34:17):
by the ONC as a decision support Instrument, which includes
AI and also requires you to have to get past
that black box of where did this algorithm come from?
Is this actually a hallucination?
Speaker 1 (34:29):
Or am I like?
Speaker 2 (34:31):
Where am I getting this recommendation? And I think that
guardrail is just as important as the solution itself, because
medical hallucinations of artificial intelligence and machine learning programs can
cause those same issues that I talked about where the
fifteen year old is swallowing almost forty pills. We don't
want that, and so there's that. The other perspective of
(34:53):
building within AI tools as technology specialists and developers, using
a model to help you as you're writing code is
essentially your own personalized pair programmer. I think is really
useful if you have at least a basic understanding of
the tech that you're trying to build in. It's not
going to supplant your knowledge and information, and you have
(35:15):
to work through some of the suggestions that are high
confidence and low accuracy, otherwise your code will suffer from that.
But there's really a lot of opportunities as you're building
to help speed that up, and time has shown us
that as well as from the perspective of the clinician
as you're caring for patients of the bedside.
Speaker 3 (35:36):
And I again, and I think there's some ways to
use it that have not been are not sort of
classic when you think about it, that could end up
being really useful. We know, for instance, that most research
has to be published in English, it has to be
written in a certain form of English, and much of
the world doesn't speak English as their first language. And
so if you can't write a research paper in the
(35:57):
proper pros, your research is not going to be a
to it, just even if it's high quality and so,
but having AI help you write a research paper, you
know where your research and your methods are sound, but
because you may struggle with English, you can't get it published.
That's a barrier to entry that shouldn't be there, and
AI might be really helpful for, you know, to help
(36:18):
you with that. One of the ones we were actually
looking at was sort of what John was talking about,
where you teach people some basic data science and then
use AI to help you know, analyze your data. You know,
there's lots of places where they don't have people that
can do it themselves, and sort of AI can help
with that. And so I certainly think AI has a
(36:39):
good role to play. I think you have to be
really careful with it. I think that most AI, especially
in a healthcare setting, is untested. I think implementing a technology,
you should study it. We do with all the rest
of the technology, and yet AI I don't think is
undergoing the same sort of you know, scrutiny, which again
I think is going to have some difficulty and some
(37:02):
problems that it's going to introduce. It has to be
trained on something. Lots of the old healthcare data is
not great and a lot of it's biased, and I
think that's the other thing to really be cautious about
is that if you're going to train on it, we
know AI tends to amplify sort of biases and data,
and so you have to be really cautious. I think
(37:23):
using AI for especially for decision making.
Speaker 2 (37:26):
The classic example for that was the likelihood of a
skin lesion being cancer was increased from the AI got
algorithms if you had a ruler right next to it,
because the images they were trained on whenever you're measuring
the malignant lesion, they all had rulers, and so the
ruler was a predictive variable.
Speaker 1 (37:47):
That's not right. Yeah, I did AI course at Stanford
this summer and that's the exact same story that professor
told us about. So I do understand. I mean, you're
going to be super castle, but there are some merits, right,
I mean, if you test it to grace point and
put it through the rigor of going through all that
(38:09):
unbiased data and things like that. I mean, there is
some benefit to it. So switching gears a little bit.
How much of adoption have you seen here? And I
know for a fact that, I mean, you've tried to
take this outside of the Americas as well to some
of the countries. Who is using whether it's Firefly or
(38:32):
the overall stack in general, who is using it today?
Speaker 3 (38:35):
So last week we checked again there are certain developers
that were implementing it in Brazil, Colombia, Mexico, Germany, Italy, India,
and Nigeria, Unity Candidas.
Speaker 1 (38:53):
It's great.
Speaker 2 (38:54):
I'll say that's good adoption from what we've developed over
the time. I always think that's.
Speaker 1 (39:00):
And speaking of adoption, John and Gray, I mean, I'm
assuming all of this is open source as well, and
everything you do, you put it out there for people
to contribute or anything people should know about everything.
Speaker 2 (39:12):
Everything that is the baseline package that we want to
help everybody is open source. So Firefly is open source.
The EMS project that I talked about, that's open source.
Some of the things that we've built were closed, some
have different permissive licenses that allow other people to learn from,
but not necessarily directly compete with, so like functional source
(39:35):
if you've heard about those. But realistically, our goal is
we always want to move the needle forward. We're not
trying to train up other people that then immediately directly
compete with us. But for Firefly, we want this to
be useful for everyone. That is a core package that
(39:55):
everybody needs to have and that will always be open source,
you know, through the duration that we're supporting, and it
needs to be like that.
Speaker 3 (40:06):
Just definitely, and we and again we want it to
be We want Google's Open Health Stack we brought up before,
and the Android SDK is used in a number of standards.
Now it's actually the WHO is creating smart guidelines that
it's going to be used sort of around the world,
which is based on Open Health Stack and Android SDK,
(40:26):
And again we're not competing with them where there's no
reason too, there's lots of work, but we want sort
of a parallel option because again we see the value
in Flutter and being able to create something that could
then be used and implemented again sort of based on
WHO guidelines. And that's again the sort of thing where
we're not We don't want to keep it to ourselves.
(40:48):
We want to share it. We want anyone who's interested
or available or wants to try it or or you know,
work with it, develop, you know with it. We we
want to encourage that.
Speaker 1 (40:57):
Perfect Thank you so much. Any anything else that I
have not asked about Me June that you want to
get out there in the community to not say, If.
Speaker 2 (41:05):
Anybody is interested for us, certainly go to Majune dot com.
If you're interested in Firefly, there's Firefly dot dev. There's
a slack that people can join where you can collaborate
with others across the globe that are building solutions here.
We're pretty easily approachable. We're always happy to collaborate with
others that are deeply invested and interested in the space
(41:27):
and hopefully making things better for patients and families and clinicians.
Just as we go from you know, one month to
the next, one year to the next. We're always happy
to work work together and to collaborate.
Speaker 1 (41:43):
I agreed, well, said John, So that Gray and John
always a pleasure speaking to you both. I'm always sharing
for you guys and talking about me June to whom
I think is relevant. So wish you guys the best.
We appreciate it, and uh you know anything you know.
I'll be sure to publish all the details about me
(42:05):
j you and all your repos and your slack and
everything on the YouTube channel, so I'm sure there'll be
a bunch of questions I'm sure you can answer. But
thanks so much for taking the time amidst your busy schedules,
lovely to speak to both of you, and I learned
a lot. Hopefully everyone listening in and tuning in learns
(42:26):
a bunch as well. Thank you so much.
Speaker 3 (42:27):
Thank you for having us pleasure. It was fun.
Speaker 1 (42:29):
Thank you.
This is the Startup Still Say podcast.
Speaker 2 (00:02):
Thank you for tuning in favor subscribe over LinkedIn and
be sure to give us your predom. I hope you
enjoyed this episode.
Speaker 1 (00:11):
Hello everyone, this is Anthony Prakashi, your host for Startups
They'll Say. Here we are with another brand new episode,
and I'm super excited today to bring to you a
couple of my friends. We were colleagues or partners in
crime a couple of years earlier, but we are all
members of a flood of community right now and I'm
(00:34):
excited to bring to you their story, their mission and
vision for something super important in the healthcare world, both
in the US as well as the entire world that
they're trying to solve, and it is Their company is
called May June. So John, doctor John Manning here is
the CEO of May June and doctor Gray Fuckinberry is
(00:56):
the CEO of May June. So if about anything, I mean,
they always inspire me because both of them are doctors
by day and sometimes by night, and then the rest
of the twenty four hours, they are technologists and hardcore technologists.
I must say so a warm welcome John, and great
(01:16):
to startups Bills. Why don't we kick it off maybe
with John. You can go first and tell a little
bit more about yourself and your journey of how you
started off me.
Speaker 2 (01:26):
Yeah, sure, and thanks for the introduction Anthony. We've gone
a long time back, so it's great to have the
opportunity to speak here. So I'm John Manning. I'm an
emergency medicine physician by training. I teach medical students and
residents how to care for patients in one of the
Level one trauma centers in Charlotte, North Carolina, and have
(01:49):
done that for almost a decade at this point, getting
close to that point right before coming down here, like
right around that time, I co founded May June with
an interest in building technology solutions that focused on human factors,
usability and just trying to make your life easier from
(02:10):
an end user perspective, the kind of short version of it.
And we can go in depth if interested. But I
used to be ascribed in a medical field in DC.
I hand wrote notes helping emergency positions. Was there when
we transitioned from paper to electronic saw the equivalent of
twenty different health systems across med school, residency, fellowship, etc.
(02:33):
And all with different levels of medical record integration and usability.
A lot of them had a lot of usability challenges,
and my attempt to give back was in building tech
solutions through made focus on that.
Speaker 1 (02:47):
Awesome. Great, why didn't you do the same? Just introduce yourself?
Speaker 3 (02:52):
Great and thanks again for having us. I'm excited to
be here. I've been looking forward to it, even though
I responded like so, I'm Gray Falconberry. I'm CTO trained
in internal medicine and pediatrics. I practice as an adult hospitalist,
and my other part of my day job is I'm
(03:13):
the co chair of the Global Health Informatics Program at
the Children's Hospital of Philadelphia, although I actually live in Atlanta.
That's one of the joys of technology and informatics is
that so much of its virtual. You know, you don't
have to be in the place where you practice, which
is really nice and allows flexibility. And I there's a
longer story involved, but I got into this from the
(03:35):
global health perspective actually and looking at how you can
sort of leverage technology abroad to improve healthcare.
Speaker 1 (03:43):
That is awesome and it's a super worthy cause. And
like I've heard from both of you, I mean, it's
it's obviously even in a country like the US. It
is informatics and the amount of manual effort that goes
into the work aside from you guys, care for patients
I mean is a big part of every doctor's day, right,
(04:04):
I mean, and there's so much of others involved in
it as well. So let me ask you this, I mean,
where did you kind of get the h to say, Okay, man,
this is a big problem. We've got to do something
about it. I mean who started that? I mean, I'm
sure both of you have to think about that vision
and believe in it saying okay, we can do something
to do it. But John, I know you started this
(04:27):
and Gray probably joined later. So what did you tell
the audience about? I mean, where is it that you
noticed this? I mean, I know you said you were
ascribed maybe there, but tell us more.
Speaker 2 (04:38):
Yeah, So it all kind of comes back full circle.
And when you're just trying to think through like how
you can improve and where those options are, you mentioned
usability being a big issue for physicians. Christine Sinski from
the American Medical Association published on that there's a thing
called pajama time, which is the time you spend at
home just charting and catching up on your documentation, and
(04:59):
every specialty has it. It's not fun, it causes burnout.
It leads to burnout, so it's everywhere. The specific path
that led to May June, interestingly was as a resident
in Emergency Medicine, I pitched an idea to the first
ever Emergency Medicine hackathon at their national conference. Got the
(05:22):
grand prize, got a bonus phase prize, and an opportunity
to speak at a public health summit in part of
the keynote in Illinois. So we built an app that
was based on the hackathon idea. Wrote it in Java's
first time that we'd ever written any sort of code.
It was a Java developer friend of mine who had
not written anything in Android before, neither had I. He
(05:45):
lived a few months away or a few blocks away,
and the better part of my final year of residency
training was building this application. All through fellowship, got a
lot of opportunities to work with other startups and do
a bit of mentoring and working with a group called Matter,
a health tech incubator, and time unfortunately broke a little
(06:09):
bit of the app, and when we were trying to
think through am I going to do this in Cotland?
Because Cotland is new. Am I going to switch over?
Am I going to just update this in Java? That
was almost exactly the time that Flutter was announced, so
December of twenty eighteen. The next month, I was in
Boston giving a live coding session, handshaking for five minutes,
(06:32):
showing how you can really make this because it just
fell in love with Flutter. A few months later, it
was a big informatics conference at their annual symposing or
their clinical Informatics conference, and in two hours introduced Flutter
and live coded how to create the conference's app in
real time while just talking and writing code. And one
(06:57):
of Gray's predecessors, his co fellow one year ahead of him,
tap through that conference. It was like seventy eighty people
in that two hour event. And when he joined Informatic
like that Informatic fellowship a few months later, he was
interested in global health and maybe Android or you know,
trying to build and develop for this, and she looked
at him, she was like, well, why don't you just
(07:17):
look into Flutter? And he started looking into this, and
there was a shared Slack channel with a lot of
Informatic fellows that were going through this new training, and
we started collaborating, collaborating every day or slack over years
and started building these solutions that would then allow other
(07:38):
people that are in and around the health tech space
to use the standard for interoperability FIRE within Flutter applications,
and much of that goes to Gray and all of
the time and effort that was spent in this, I
was I just make apps and Flutter and I live
code and I can talk and you know, it's the
emergency medicine brain. I can talk and write code at
the same time, you know. And that's my thing. But
(08:01):
a lot of that underpinning and architecture that's open source
and has been used by developers a lot across the
world that goes with many things to correct.
Speaker 1 (08:10):
That is awesome. John, So you touched on a lot there, right,
I mean you talked about fire obviously informatics and great,
maybe you can wear the CTO had I've always wondered.
I mean, you know, you guys have very little time,
yet you've found the time to learn a technology like Flutter,
(08:31):
decide and pick on it saying Okay, this is the
best choice to move forward with what we're trying to build.
I mean, for I mean, obviously you don't need to
be an engineer to learn flutter and use flutter anymore.
I mean you are perfect examples for that, right. But
how did that love for even technology come in to
both of you? I mean, obviously you need a spark
(08:53):
within you, you need that push within you to say, Okay,
this is not something that I can outsource to a
company and say build it. I want to do this myself, right,
I mean, how did you sort of make that choice
great to say okay, I mean, you guys want to
be you know, getting your hands and feed wet with
building this whole thing.
Speaker 3 (09:13):
So I my love of technology and sort of science,
I think at John's two, you know, is not not
a new phenomenon we liked growing up. In college, I
was a physics major and I'm minored in computer science,
so I had a little bit of a background, but
then I didn't really use it. I went to medical
school and was in medical school when EHRs really started
(09:36):
coming around and sort of everybody switching over from paper
to electronic and so I saw that transition, and at
the time it was exciting. I was looking forward to
it because again I like technology, and at the time
it seemed mostly good you know, the not having to
fight for the chart, having one chart for a patient,
having eight people trying to look at it. You know,
you didn't have to worry about that anymore, having to
(09:57):
walk across or down fifteen floors and across campus to
write an order and a chart. You could just put
it into the computer. So there were some really nice
things about it, and then there were a lot of
not nice things about it. It was obvious that, as
John said, the usability was not a sort of forefront
in anyone's mind. You know, with the design of these
(10:17):
and any sort of study or survey done on physicians
now you'll see that. You know that there is burnout,
and so you work with that long enough and it
becomes a source of frustration because it really felt like
the people who designed it hadn't ever seen a patient.
And that was probably true, you know they and you
(10:39):
really need to understand the workflow. You need to understand
how it feels and what you need and how it
should be organized to take care of, you know, a patient.
And I think what drove a lot of us into
informatics is the feeling that we want to be involved
in this, We want to help make this better. This
is something. This is a skill set that we feel
like we have and we can develop and we can
(11:00):
actually contribute and sort of change things for the better.
So much of medicine is already sort of set in place,
and there's there's so little opportunity to sort of change
any of it. But informatics, I think is different. It
is still new, it is still changing. It is a
lot of it is sort of set, but a lot
of it's not, I think enough of it's not that
(11:24):
we can sort of affect that change. And I think
that's what's exciting about it is the feeling that I
actually have the ability to hopefully, you know, help medicine
and healthcare be provided in a better way through technology,
you know. And that was sort of how informatics I think,
came into my life. Yeah.
Speaker 2 (11:44):
So if you think of like as an example in
medicine of something that's never changed and it will never change.
Coagulation cascade, which sure whatever, Okay, you don't have to
know about it. One of there's a dozen factors. One
of them is literally just calcium and it was named
its own factor. Sure, and another one factor six. They
(12:07):
found out, oh actually it was just another version of
factor seven, and so there is no factor six. There
isn't one. So it's labeled up to a dozen or
a thirteen, but you're missing a whole number. And since
we've all learned it this way, I'm sorry you have
to learn it that way too, even though it makes
no sense. Welcome to healthcare. Here's a chance where we
can actually build things that match the workfload that's there.
(12:31):
To Gray's point, if you don't see it, if your
brain isn't literally into the weeds, and it's not there's
one thing to be shadowing somebody who was doing this,
there's another where you were putting all of your cognitive
effort and load into the care of a person. So
you don't make those mistakes, and so that you know,
medicine always has mistakes, but so that you minimize that
(12:52):
and yet you allow for the right processes and everything
to help. Well, if you have a very expert, complex
system that's pulling all of your attention away, you're gonna
get distracted and things will happen, and it's you know,
it can lead to it can lead to harm.
Speaker 1 (13:07):
If you.
Speaker 2 (13:09):
What was the the book that was talking about the
digital doctor from like ten years ago or ten to
fifteen years ago, talks about all the processes of how
a fifteen year old child swallowed thirty eight and a
half pills of a single antibiotic and seized, and all
(13:33):
of the workfload things, as well as the medical issues
that were a part of your electronic health record that
was almost like a Swiss cheese effect where everything was
perfectly aligned to a point of a nurse in a
room with a fifteen year old by himself, with mom
or dad not nearby, as he is individually swallowing almost
(13:54):
forty pills. There's a lot to it. So any ways
that we can improve that medical perspective I think is helpful.
Speaker 1 (14:04):
Yeah, And speaking of a medical perspective, John, I mean,
I know both of you have touched on what you're
trying to make better. But if there is someone who
just is a layman in layman's terms, if you want
to explain before May June's in an after main story
of course, using fire which is fhir for those listening,
(14:27):
how do you explain the work that you guys are
doing to make this thing better?
Speaker 3 (14:34):
So I could take this one. So part of it,
I think is something that we're all familiar with. And again,
in no way as a clinician, I think it's a
patient you see it. We all go to doctor's appointments
or healthcare appointments or therapy appointments where you would like
to be able to move records around. You know, you've
seen someone else before you would like to share those
(14:55):
records with someone else. And it's always painful. It's always
extremely hard. You have to fill out release sheets and
they have to often fax them over the fact that
facts comes into conversation at all anymore is ridiculous, but
it does in medicine all the time. You know, I
regularly get two hundred pages of faxes of someone's medical
(15:15):
record from another hospital, you know, trying to take care
of it, because medical records don't share data very well.
And then about not quite fifteen years ago now, there
was a standard that started being developed called fire Fast
Healthcare Interoperability Resources in order to share it. And again
(15:39):
that's a big deal because just having that information is
so important. So the ability to easily share healthcare data
you know can is not just a convenience issue, it's
it's a safety issue. If I don't know what medications
someone's taking and I give them something that interacts that
obviously can be incredibly dangerous. You know, I don't need
(16:01):
to do certain tests if I know that they were
just done somewhere else. And so having the ability to
more easily share that data from place to place is
incredibly important. And I think it helps both from sort
of again a frustration and a time sort of commitment,
as well as a pure safety issue. And that's something
that again we've we've sort of touched on. As John
(16:25):
pointed out, we both sort of lean slightly different directions
on this, but again trying to make electronic records that
work better for the people involved. So we want to
be able to share data, We want it to work well,
we want it to be easy to use, you know.
And and so I think both the sharing of the
data and the usability can't be overstated about sort of
(16:47):
how important those aspects are.
Speaker 2 (16:49):
Yeah, adding to that a little bit, and when you
think of fire, so FIRE is now the global standard
for interoperable communication of healthcare data period adopted by the
US governments. For that large electronic medical records, they have
to support at least a core subset of fire data
that patients can get access to. And no cost, and
(17:12):
when you're trying to communicate this data, it is miles
beyond what was there before before the Fire standard. What
kind of happened through all of this sorefly Fire with
Flutter Library Integration FLI, where the packages that Gray had
developed over the course of his fellowship, which is about
(17:34):
ten million lines of code total three million lines of dart.
A lot of it's auto generated, but in a way
that allows you to use any published version of fire
that also goes through all its test cases and then
directly put this into your Flutter application. And that Android
equivalent of this is open heal Stack by Google for Flutter.
(17:55):
If you're building a Flutter app, you should use the
Firefly packages. So you know. To Gray's point, so he's
a medicine pedes trained hospitalist, so he cares for patients
when they're admitted to the hospital, adults and children. Practiced
in eight countries, spend a year in Uganda, wrote the
textbook on pediatric care in that setting, and has a
(18:19):
very high global health focus of trying to help in
areas that are very resource limited. My background is emergencies.
I care for emergencies. I did MS things and a
lot of the things that come around me as an
emergencies and disasters, and how can we improve in those cares,
those care settings that are unscheduled and where a lot
(18:40):
of people are afraid, And both of them have opportunities
to improve in rural areas as well as global. So
to your point that you had mentioned earlier, Anthony, of
what was our story, well, we both had a shared
interest of building things to help in healthcare from usability
perspect with a mobile first approach. Then collaborated on the
(19:04):
Firefly packages, reached out to the Office of the National Coordinator.
The actual National Coordinator at the time was Don Rucker,
who we got through other members of Nyjune. They had
connections and here we are in front of the ONC
talking about Firefly and how cool this is, and they said,
(19:25):
look into social help like social care areas, and so
that is where we built screening tools to help with
social need and multiple states and multiple languages, and we're
still on the Gravity Projects. Main page is their success story,
which is wonderful and cool. But you know, as clinicians,
we're always trying to see how best we can apply
(19:48):
this in a way that helps in rural, underserved and
in global areas.
Speaker 1 (19:54):
Yeah, so you mentioned emergency medicine. One of the things
that always bothered me about, you know, going to a
hospital in the US, even if you go to an emergency,
is it's never an emergency in an emergency, right, I
mean you'd go with you know, so much pain, and
you'd be like, want to be seen immediately. You're invariably
waiting for sixty to ninety minutes before you've seen or
(20:17):
at least somebody comes and asks you something, right. I
mean in a lot of cases it has got to
do with data collection. You know, they are validating it
behind the scenes, so you know, just as a patient
myself or you know somebody who I've seen, even in
the children's hospitals, great, right, I mean when my daughter
was young, I mean, you know, she'd be running very
(20:39):
high temperature and you go there, they make your sit
for forty five minutes, right, I mean why because you're
still collecting data. So in that sort of a let's say,
I mean you don't want to call it life threatening,
but it's still an emergency. Is there something that you know?
Using fire in May June, I mean you can simplify
that process to accelerate the time to which by the
(21:01):
time you walk in through that door, a doctor is
able to see you at least within fifteen minutes or so.
I mean, you know, if sixty minutes is too much,
at least cut that down by yeah, you know, seventy
five percent.
Speaker 2 (21:14):
So yes and no. So I think on the other
side of the curtain. So Atrium Health is where I
practice associate professor, and I care for adults and children,
and it's patients are not seen in the same order
that they arrive. They are seen often by the sickest
ones first. So thankfully you and your family were not
(21:36):
that sick. There's an emergency, of course, we'll address that.
But trust me, the very very like near death patients
and severe traumas and medical and traumatic resuscitations, they're seen
really fast. It's not sixty minutes. And there are workflow
tools where you can even register a patient without any
information quickly, be it medical or trauma, so that you
(21:59):
can give them medication and treat them and get do
an ultrasound or get some X rays or something at
the bedside to help them. All of that stuff has
its own workflow in American medicine that uses the electronic
medical record and the use of majune technology probably isn't
(22:20):
as relevant there if you look in resource limited scenarios.
So another example, I helped with some of the hurricane
relief efforts in western North Carolina where it was multiple
groups together and some of us were using paper and
some of us were using another technically medical record. That
(22:42):
will also wasn't really designed by a clinician, and it
was clear that there were a lot of opportunities to
improve with that. That's a place where we can inject
and help and where we have the skill set and
the expertise and the domain knowledge to improve in those settings.
And we've had conversations with groups that were interested and
our services from there. So from the over arching perspective
(23:05):
of you know, just emergency medicine care, usually it's going
to be in the low resource or disaster response or
international area where we can improve rather than the traditional
brick and mortar emergency department. But medicine always changes, you know,
if we build something and then it expands and then
(23:28):
it continues to expand. And that's right up both Gray
and my alley of ways that we can improve in
areas that also build on the background expertise that we have.
Speaker 3 (23:38):
That's perfect, and I'll say a slightly different sort of
view of it. I again not sort of the emergent
view in terms of how quick we can get you in,
but I think that some of the stuff we do
can make the time we spend with you more valuable.
And I think that you know, most of the time,
the about of time that we spend with a patient,
most of it is collecting data and very little of
(24:01):
it is actually spent talking to a patient. And so
I think using technology, including some of the June you know,
questionnaire data gathering, structured data capture sort of stuff, a
lot of that could be done beforehand. And so if
you have a fifteen or a twenty minute appointment, or
you know, whatever John has available to be able to
see you in the emergency room before the next trauma,
(24:25):
you know, instead of spending all that time or the
bulk of it trying to gather all your information, make
sure that we have all of your medications correct, that
we know all of your history. If that can all
be shared or gathered more efficiently ahead of time, then
we can sort of review it quickly and then we
can spend more time talking to you. We can spend
more time ensuring that we didn't miss anything that's important.
(24:48):
We can ensure that you understand the plan, and so
the time that we do spend with you can be
actually spent with you and so sort of caring about
what the problem is instead of just asking you a
whole lot of questions. And so I think there are
some ways, even in sort of a traditional setting where
(25:09):
you know, we could be beneficial that. Again, I think
there's lots of opportunities to sort of improve that.
Speaker 1 (25:15):
Yeah, thank you for that.
Speaker 2 (25:15):
I mean the term doctor, go ahead, John, Oh, go ahead.
The term doctor comes from dose air, which means to teach.
It is literally your role is to map somebody's health
literacy and then help explain and work through that. And
the other point from Gray is if you've ever seen
there's a twenty twelve article in Jamma called the Cost
(25:41):
of Technology by Toll and it was a drawing of
a seven year old in their pediatricians clinic where you know,
they're smiling, Mom's smiling, big sisters holding baby child like
baby sibling. Everybody's smiling. Then the doctor is smiling but
has their back turn and a staring at a computer.
And this is over thirteen years years ago when that
(26:01):
was done, and it's still relevant today.
Speaker 3 (26:04):
And again I think that's again to dive back into
the technology. I think that's where Flutter is so nice
and actually part of the reason we chose it as
the health stack. Again, I was looking at sort of
a global perspective and everything global it has to be mobile, right,
you know, it has to be at least available, and
most EHRs or not most of them were never designed
(26:25):
to be and so you know, we're looking at what's flexible,
what can be done, you know, and multiple platforms, multiple devices.
And that's sort of what sold me on Flutter because,
as John said, one of my colleagues said, well, she
had seen this talk about Flutter. Maybe I should look
into it. So I did, and I got on Slack
(26:47):
and I started working with John. And the punchline to
his story was that it was about two years later
when I realized that the speech my colleague had seen
was given by John that then got that, then got
the end of flu. So it was yeah, it's a
full circle.
Speaker 1 (27:03):
Yeah, full circle indeed. And I mean, thanks for sharing
the procedure about how the emergency rooms work, right, I mean,
I know everybody comes in there and there's nothing more
frustrating than waiting, right, so all your anger levels go
up and you see a lot of people walking up
(27:24):
to the poor receptionists and saying, I've been here for
ninety minutes. You know why is nobody seeing me? So
that part is super helpful. So shifting years a little bit,
John and Greg, obviously you started May June with a vision, right,
how much of that have you achieved? And what are
the next couple of years looking like for you? Guys
with May June, either of you can take it.
Speaker 2 (27:47):
So I'll say that our vision always shifts forward. We're
always working to build new things, better things. Everything we've
built in May June has been contracted work for hire
to satisfy a specific problem that needed some extra usability.
(28:07):
Flare that was cross platform by design and frequently had
one group or one solution that didn't match the initial expectations.
So a good example of the first thing we built
was for any MS agency that had protocols to help
in rural areas that the initial the content was excellent,
(28:30):
that was something they built and curated, but the delivery
needed help, and so we helped and you know, we
continue to help and support them for years. We're always building,
and we're always doing this in a way that can
scale into other solutions and can apply some of the
(28:50):
backgrounds and experiences that we have as as practicing clinicians
and informatis and we have other informatics clinicians that are
a part of MADE as well, and in addition to
more tech first people as well. I think for us,
we're always trying to find that best opportunity where we
(29:11):
can have the biggest impact on the broadest scale in
a way that pushes things forward in a good way.
And I think that's always where we've strived. And so
are we there yet? I'll say no, Are we always
going down that path?
Speaker 1 (29:25):
Yes?
Speaker 2 (29:26):
And you know, it's one opportunity after the next and
one solution after the next as we learn and make
our code better and as we provide that service of
Firefly for the global community, in addition to building upon
that and harnessing it with the tools that we that
we then deploy got it.
Speaker 3 (29:47):
Yeah, Well, I think that a lot of what we've
done so far was still sort of foundation building to
an extent. It was sort of learning and building out
our skills building out. I mean, some of it was
certainly sort of technology building out, the tech stack and
that sort of thing, and some of it was also
sort of finding where we wanted to be. You know,
(30:07):
We've done you know a number of different projects in
a number of different places, sort of in and sort
of slightly out of healthcare, always intertwined, you know, And
so I think we're we've sort of found our niche
and sort of where we want to be. And I
think that's again it's sort of exciting, and a lot
of times it's places that other you know, groups don't
(30:29):
want to go. And again, sometimes there's a lot of
times there's a monetary barrier to it, you know, whether
it's sort of rural areas or low resource settings or
disaster relief one of the ones that that again John
brought up earlier. You know, these are all really important
to have good health tack tech stacks for if you're
going to actually collect data and be able to show
(30:50):
what you're doing, you know, who you're treating. You know,
it's important to be able to record all that. It's
really hard, though, but I think you know a lot
of that fit what our interests and what our passions
are and our knowledge and our expertise, and so I
think we're sort of excited and looking forward to being
able to use you know, who we are and what
(31:12):
we want to do it, you know, even more effectively
as we move forward.
Speaker 1 (31:17):
That is great to hear, especially the part that you're
always kind of looking forward, right. And I try to
bring up AI in most of our conversations because I mean,
it seems to be affecting everything I mean, and especially
I mean, I've heard in healthcare there's a lot of
use cases and John, you started off with scribing. I've
heard a number of you hospitals using those livescribers sort
(31:41):
of thing. But with your vision for me and kind
of looking forward and the impact that AI can make.
Or the other trend I've been hearing about is you know,
taking the hospital to the home, right, I mean, if
there's critically ill patients, I mean, you're moving a lot
of equipment into the patient's house and still collecting data
(32:03):
and things like that. Are there's a lot of you know,
devices that the patients can carry home, but as a doctor,
you still want to collect all that information. So looking
at some of these new trends, I mean, what are
your thoughts on you know, how May June will evolve.
Obviously you would have given that thought. I mean I second,
maybe a second part of the question is, I mean,
(32:25):
how much of AI have you actually tried to bring
into the whole tech stack and try to do something
with it.
Speaker 2 (32:31):
There's a few things from that. So the first one
is AI from the perspective of the clinician and the
second is AI from the perspective of the developer, and
they're totally different buckets. So let's take the first bucket. So,
first one, I've given local talks, given talks at Amya
about this, and I've tested this and I also co
(32:52):
chair our service line Innovation and Informatics group with an Atrium,
and we've tested this from the very beginning and with frequency,
there is a difference between So as a scribe, you
mentioned this, there's a difference between a good scribe and
a bad scribe, and a good scribe speeds you up.
A bad scribe has things interjected that gives you almost
(33:13):
twice the amount of time that you would be using otherwise.
Tech always changes. Even a year ago, there were a
lot of bad scribe ambient dictation tools out that a
lot of it's improved even in the last eight months.
And if you want something to look at, there was
a AI tech sprint from the VA, the Veterans Administration
that was hosted at the start of this year and
(33:34):
then announced early Q two, and if you just search
for the AI tech sprint within the VA, you can
see some of the companies that were well embedded in
that perspective. I think ambient dictation is going to have
a high opportunity to help clinicians caring for patients, albeit
in limited resource settings. That's going to be a challenge still,
(33:55):
just because all of these need a lot more cloud
services of a level to them, So that's going to
be your mileage made very depending on where you practice,
but there's definitely opportunity. The other aspect for decision support
from the clinician's perspective, I think is great. What used
to be called clinical decision support is now being rebranded
(34:17):
by the ONC as a decision support Instrument, which includes
AI and also requires you to have to get past
that black box of where did this algorithm come from?
Is this actually a hallucination?
Speaker 1 (34:29):
Or am I like?
Speaker 2 (34:31):
Where am I getting this recommendation? And I think that
guardrail is just as important as the solution itself, because
medical hallucinations of artificial intelligence and machine learning programs can
cause those same issues that I talked about where the
fifteen year old is swallowing almost forty pills. We don't
want that, and so there's that. The other perspective of
(34:53):
building within AI tools as technology specialists and developers, using
a model to help you as you're writing code is
essentially your own personalized pair programmer. I think is really
useful if you have at least a basic understanding of
the tech that you're trying to build in. It's not
going to supplant your knowledge and information, and you have
(35:15):
to work through some of the suggestions that are high
confidence and low accuracy, otherwise your code will suffer from that.
But there's really a lot of opportunities as you're building
to help speed that up, and time has shown us
that as well as from the perspective of the clinician
as you're caring for patients of the bedside.
Speaker 3 (35:36):
And I again, and I think there's some ways to
use it that have not been are not sort of
classic when you think about it, that could end up
being really useful. We know, for instance, that most research
has to be published in English, it has to be
written in a certain form of English, and much of
the world doesn't speak English as their first language. And
so if you can't write a research paper in the
(35:57):
proper pros, your research is not going to be a
to it, just even if it's high quality and so,
but having AI help you write a research paper, you
know where your research and your methods are sound, but
because you may struggle with English, you can't get it published.
That's a barrier to entry that shouldn't be there, and
AI might be really helpful for, you know, to help
(36:18):
you with that. One of the ones we were actually
looking at was sort of what John was talking about,
where you teach people some basic data science and then
use AI to help you know, analyze your data. You know,
there's lots of places where they don't have people that
can do it themselves, and sort of AI can help
with that. And so I certainly think AI has a
(36:39):
good role to play. I think you have to be
really careful with it. I think that most AI, especially
in a healthcare setting, is untested. I think implementing a technology,
you should study it. We do with all the rest
of the technology, and yet AI I don't think is
undergoing the same sort of you know, scrutiny, which again
I think is going to have some difficulty and some
(37:02):
problems that it's going to introduce. It has to be
trained on something. Lots of the old healthcare data is
not great and a lot of it's biased, and I
think that's the other thing to really be cautious about
is that if you're going to train on it, we
know AI tends to amplify sort of biases and data,
and so you have to be really cautious. I think
(37:23):
using AI for especially for decision making.
Speaker 2 (37:26):
The classic example for that was the likelihood of a
skin lesion being cancer was increased from the AI got
algorithms if you had a ruler right next to it,
because the images they were trained on whenever you're measuring
the malignant lesion, they all had rulers, and so the
ruler was a predictive variable.
Speaker 1 (37:47):
That's not right. Yeah, I did AI course at Stanford
this summer and that's the exact same story that professor
told us about. So I do understand. I mean, you're
going to be super castle, but there are some merits, right,
I mean, if you test it to grace point and
put it through the rigor of going through all that
(38:09):
unbiased data and things like that. I mean, there is
some benefit to it. So switching gears a little bit.
How much of adoption have you seen here? And I
know for a fact that, I mean, you've tried to
take this outside of the Americas as well to some
of the countries. Who is using whether it's Firefly or
(38:32):
the overall stack in general, who is using it today?
Speaker 3 (38:35):
So last week we checked again there are certain developers
that were implementing it in Brazil, Colombia, Mexico, Germany, Italy, India,
and Nigeria, Unity Candidas.
Speaker 1 (38:53):
It's great.
Speaker 2 (38:54):
I'll say that's good adoption from what we've developed over
the time. I always think that's.
Speaker 1 (39:00):
And speaking of adoption, John and Gray, I mean, I'm
assuming all of this is open source as well, and
everything you do, you put it out there for people
to contribute or anything people should know about everything.
Speaker 2 (39:12):
Everything that is the baseline package that we want to
help everybody is open source. So Firefly is open source.
The EMS project that I talked about, that's open source.
Some of the things that we've built were closed, some
have different permissive licenses that allow other people to learn from,
but not necessarily directly compete with, so like functional source
(39:35):
if you've heard about those. But realistically, our goal is
we always want to move the needle forward. We're not
trying to train up other people that then immediately directly
compete with us. But for Firefly, we want this to
be useful for everyone. That is a core package that
(39:55):
everybody needs to have and that will always be open source,
you know, through the duration that we're supporting, and it
needs to be like that.
Speaker 3 (40:06):
Just definitely, and we and again we want it to
be We want Google's Open Health Stack we brought up before,
and the Android SDK is used in a number of standards.
Now it's actually the WHO is creating smart guidelines that
it's going to be used sort of around the world,
which is based on Open Health Stack and Android SDK,
(40:26):
And again we're not competing with them where there's no
reason too, there's lots of work, but we want sort
of a parallel option because again we see the value
in Flutter and being able to create something that could
then be used and implemented again sort of based on
WHO guidelines. And that's again the sort of thing where
we're not We don't want to keep it to ourselves.
(40:48):
We want to share it. We want anyone who's interested
or available or wants to try it or or you know,
work with it, develop, you know with it. We we
want to encourage that.
Speaker 1 (40:57):
Perfect Thank you so much. Any anything else that I
have not asked about Me June that you want to
get out there in the community to not say, If.
Speaker 2 (41:05):
Anybody is interested for us, certainly go to Majune dot com.
If you're interested in Firefly, there's Firefly dot dev. There's
a slack that people can join where you can collaborate
with others across the globe that are building solutions here.
We're pretty easily approachable. We're always happy to collaborate with
others that are deeply invested and interested in the space
(41:27):
and hopefully making things better for patients and families and clinicians.
Just as we go from you know, one month to
the next, one year to the next. We're always happy
to work work together and to collaborate.
Speaker 1 (41:43):
I agreed, well, said John, So that Gray and John
always a pleasure speaking to you both. I'm always sharing
for you guys and talking about me June to whom
I think is relevant. So wish you guys the best.
We appreciate it, and uh you know anything you know.
I'll be sure to publish all the details about me
(42:05):
j you and all your repos and your slack and
everything on the YouTube channel, so I'm sure there'll be
a bunch of questions I'm sure you can answer. But
thanks so much for taking the time amidst your busy schedules,
lovely to speak to both of you, and I learned
a lot. Hopefully everyone listening in and tuning in learns
(42:26):
a bunch as well. Thank you so much.
Speaker 3 (42:27):
Thank you for having us pleasure. It was fun.
Speaker 1 (42:29):
Thank you.
Popular Podcasts
My Favorite Murder with Karen Kilgariff and Georgia Hardstark
My Favorite Murder is a true crime comedy podcast hosted by Karen Kilgariff and Georgia Hardstark. Each week, Karen and Georgia share compelling true crimes and hometown stories from friends and listeners. Since MFM launched in January of 2016, Karen and Georgia have shared their lifelong interest in true crime and have covered stories of infamous serial killers like the Night Stalker, mysterious cold cases, captivating cults, incredible survivor stories and important events from history like the Tulsa race massacre of 1921. My Favorite Murder is part of the Exactly Right podcast network that provides a platform for bold, creative voices to bring to life provocative, entertaining and relatable stories for audiences everywhere. The Exactly Right roster of podcasts covers a variety of topics including historic true crime, comedic interviews and news, science, pop culture and more. Podcasts on the network include Buried Bones with Kate Winkler Dawson and Paul Holes, That's Messed Up: An SVU Podcast, This Podcast Will Kill You, Bananas and more.
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com