August 26, 2026 • 42 mins
Everyone seems to be asking how AI can do a doctor’s job. We go the other way and it changes the whole picture: what if AI’s best role is deleting the admin work that crushes clinicians and slows down care?
I’m joined by Dr James Ries, a board-certified family medicine physician, active duty US Navy provider, and founder behind 20 Mile Medical. James explains how years in telehealth exposed a hard truth about modern healthcare efficiency: the real bottleneck is not clinical decision making, it’s the clutter around it. We dig into where costs actually hide (patient communication, coordination, follow-ups, fragmented systems) and why targeting clinicians first is the wrong optimisation.
James walks us through his AI layer, Telora, built to compile patient intake, highlight red flags, and surface evidence-based protocols in one place before a consultation even starts. We unpack deterministic AI vs generative AI, why patient-facing chatbots can be unsafe at scale, and how reducing cognitive load can lead to more personalised care even in shorter visits. We also explore why more frequent follow-ups can improve continuity and reduce the “three-month pile-up” of questions, plus the broader shift in telehealth towards lifestyle and concierge care.
If you care about AI in healthcare, clinician burnout, telehealth platforms, electronic health records, and safer automation, you’ll take a lot from this conversation. Subscribe, share with a colleague, and leave us a review with your take: where should AI stop and the human clinician begin?
Guest Biography
Dr. James Ries is a board-certified family medicine physician, active-duty US Navy Commander and healthcare entrepreneur who has spent approximately a decade working in telehealth.
He is Co-Founder of Twenty Mile Medical, a telehealth platform created from his experience of trying to improve inefficient healthcare systems from inside them.
After finding existing software poorly aligned with real provider workflows, James began developing his own technology, combining an electronic health record with an AI-enabled layer called Telora. The system is designed to reduce the administrative and cognitive workload surrounding clinical care by organising patient information, supporting evidence-based workflows and streamlining processes such as documentation and communication, while keeping providers responsible for clinical care.
His work focuses on improving healthcare access, provider efficiency and patient experience without treating AI as a replacement for the clinician. Longer term, James hopes that dramatically reducing the cost of delivering telehealth could help make care accessible to uninsured and underserved populations.
- Websites: https://www.twentymilemedical.com and https://www.telora.health
- LinkedIn: https://www.linkedin.com/in/drjamesries/
Subscribe on Substack
Voices in Health and Wellness is published on Substack, as part of The Diagnostician. Every episode lands there with written notes and reflections you won't find in the podcast feed. Free to subscribe.
👉 https://thediagnostician.substack.com/s/voices-in-health-and-wellness
About the podcast
Voices in Health and Wellness features conversations at the intersection of medicine, lifestyle, and human potential, with clinicians, scientists, and thinkers shaping the future of care.
About Dr Andrew Greenland
UK-based medical doctor and founder of Greenland Medical, specialising in Integrative and Functional Medicine. Trained in both conventional and root-cause approaches, Andrew helps people optimise health, performance, and longevity, with a focus on cognitive resilience and healthy ageing.
Listen
Watch
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Dr Andrew Greenland (00:03):
So, welcome back to Voices in Health and
Wellness, where we speak withpractitioners and healthcare
innovators about the challenges,opportunities, and ideas
shaping the future ofhealthcare.
Today we're exploringartificial intelligence, but
from a slightly differentperspective, rather than asking
how AI could replace healthcareprofessionals, what if we used
it to remove the administrativeburden surrounding them, giving
(00:23):
clinicians more time andcapacity to focus on patient
care?
And could doing that not onlyimprove efficiency and access,
but also reduce one of the majorcontributors to provide a
burnout?
My guest today is Dr.
James Ries, a board-certifiedfamily medicine physician,
active duty US Navy provider,and healthcare entrepreneur
behind 20 Mile Medical.
(00:43):
James is developing anAI-enabled healthcare platform
designed to automate much of theadministrative workload
surrounding clinical care whilekeeping the healthcare
professional firmly at thecenter of clinical decision
making.
So with that, James, I'd loveto welcome you to the show and
thank you very much indeed forjoining me this afternoon.
Dr James Ries (01:00):
Yeah, thank you, Andrew, for having me.
This is a great opportunity totalk about this very unique
subject.
So thank you.
Dr Andrew Greenland (01:06):
So, James, before we get into the AI and
the technology itself, can youtake us back a little and what
was your journey into medicineand what originally attracted
you to family medicine?
Dr James Ries (01:16):
Yeah, that's a great question.
I actually initially wanted tobe into pediatrics, but then of
course, you know, as youprogress through medical school
and as you do all your medicalrotations, you realize that, or
at least some people realize,they like a little bit of
everything.
And they and the other thing isuh, you know, telling a patient
that I'm not trained to seethem is something that I just
(01:39):
didn't uh didn't sit right withme.
So family medicine was alogical good fit, you know, from
from cradle to grave, you know,taking care of everyone and and
not turning anyone away wassuch a good motivator.
Dr Andrew Greenland (01:53):
Thank you.
And you've also combinedmedicine with active duty and
with a naval career.
And how has that experienceshaped the way you think about
healthcare delivery andefficiency?
Dr James Ries (02:03):
Yeah, quite a different animal.
Um, I was actually veryinterested in just uh aviation,
and that's what got me into themilitary pipeline.
But since I've been activeduty, I've realized that you
know providing excellent levelof care, no matter where you
are, is uh a different thing.
Sometimes you can't controlyour brick and mortar
(02:26):
environment.
Sometimes you can't controlwhere you're gonna be
administering that care.
And so that um really lendsitself well to telehealth,
certainly, but it lends itselfto all kinds of um unique
problems that you have to solve.
So you become a problem solver,which is a really interesting
(02:48):
way to go about it.
But I'm really happy that Idid.
You know, um, whether you're ona ship in the middle of the
ocean or whether you're in adesert, you know, things change,
but the standard, you know, youwant to maintain it as much as
possible.
So the military career has beenquite rewarding, a lot of fun
experiences, but I have about ayear left and I'm looking
(03:10):
forward to transitioning becauseuh keeping that tempo is quite
difficult with a family andeverything else.
You know, it's it's fun, but itdoesn't last forever, and
that's a good understanding tohave.
Dr Andrew Greenland (03:22):
Thank you.
And your point about problemsolving.
So, at what point did you startmoving from simply recognizing
problems in healthcare tothinking, I actually want to
build something that solvesthem?
Dr James Ries (03:32):
Yeah, that is an excellent question.
I I got into telehealth.
I've been doing it for about adecade, but really it took off
and piqued my interest duringCOVID because people had little
to no access to care.
Sometimes they couldn't, youknow, they weren't even allowed
to, you know.
So um when that all changed, itmade me realize that there were
(03:55):
that there were people outthere that would otherwise not
access medical care unless youtook down some barriers.
Even getting in the car anddriving is a barrier for some
people.
And if you provide somethingthat can still provide them a
good standard of care and goodhigh-level care uh care, no
(04:17):
matter where they're located,you reach a large segment of the
population that would otherwisego untreated.
Um, a really good example ismental health.
You know, psychiatrists askpeople to come to the clinic for
care, and sometimes theirspecific phobia is leaving their
house.
You know, so there's a bigbarrier there.
(04:38):
And there's also people thatare located in the middle of
nowhere, and they just reallydon't know what to do, even with
something as simple as a cold.
And um, and so that really,really piqued my interest uh
when it came to telehealth.
And I saw a very un untreatedlarge pop global population of
(04:59):
people, and that became a newproblem to solve.
Dr Andrew Greenland (05:04):
So you talked about a couple of
problems that you've solved, butwere there any actual
frustrations from the work thatyou've been doing as a
practicing physician that alsodrove you to create what you've
created in the platform?
Dr James Ries (05:15):
Excellent question.
Yeah, I kind of missed that.
Um, so when I started withCOVID, um, it started basically
with you know synchronizedvisits, phone calls, but very
quickly I realized that you knowthe the standard was really um
relied on the provider.
(05:35):
And that variability is whatgets people hurt, you know, and
and the systems themselves werevery incredibly inefficient to
the point where just to get onepatient seen, you would have to
staff that whole program withcustomer service and pharmacy
reps and you know all kinds ofsystems that normally wouldn't
(05:58):
be there, and systems thatbecome very cumbersome that now
have to talk to each other.
And I think that after um,after all those systems are in
place and you and you have itrunning, it becomes expensive.
And so you really have to, froma first principles type of
thinking, start over, notpiggyback on what's been done,
(06:21):
but start over and think, well,how can I do this efficiently so
that we're not having to leanon all that administrative
burden?
You know, uh the providerprobably only gets about 10% of
the fee.
The rest goes to all theadministration, all the all the
(06:43):
uh scheduling and all thecoordination.
And that really is the umthat's the low-hanging fruit.
That's what needs to be cutfirst.
And I think what's happeningnow is people are still relying
on these old systems and they'retrying to think how can we cut
costs, how can we become moreefficient?
(07:03):
And the target seems to go tothe provider first because AI is
very good at writing a soapnote.
But whether it can administergood health care, that's a
different question.
And there's systems in placenow that are not administering
healthcare.
And so you don't have to worryabout that when it comes to
(07:24):
scheduling and you know,coordination and and
documentation.
You know, these things can bestreamlined and you don't run
the risk of losing that standardof care.
And and I think it's just needsto be kind of re-rethought from
the ground up because we have anew system in place now.
Dr Andrew Greenland (07:44):
Thank you.
So before we sort of dive alittle bit more into the
platform on your work, could youperhaps give us a sense of what
your role is and how it fitsinto the bigger picture of 20
Mile Medical?
Dr James Ries (07:54):
Yeah, of course.
Yeah, 20 Mile Medical.
So I've been working with many,many platforms doing telehealth
care.
And I my uh co-owner is a PA.
He also does this.
And after a while of creatingjust systems for ourselves to
make things more efficient andbetter care, better patient
(08:16):
communication, he had said, youknow, we should probably just do
this ourselves because we'respending a lot of time fixing
other people's systems.
So he encouraged me.
He actually was the catalystfor all this, and we started our
own telehealth platform withall the improvements and all the
patient care that we had beenforcing other platforms to to
(08:40):
fix, and we just did itourselves.
And I'm very uh grateful forthat because we've done a
fantastic job, and everyone isis very appreciative on the
platform.
While we were um while we werecreating the platform, we had to
um use software.
And when we started usingsoftware, we realized after
interviewing a dozen or moredifferent software platforms
(09:05):
that they were all verydeficient, and they were all
developed by um you knowsoftware developers, not
physicians, and they were alldeveloped with the patient uh
advertising and patient facing,um, but not anything to do with
the provider and how the care isactually administered.
And we we landed on one clientand we've been using them for a
(09:30):
while, but man, it's they theymake things very limiting for
the patient-providerinteraction, even as something
as simple as I need to change mycredit card, you know, becomes
like a monumental, you know,this this mountain that we have
to climb up.
And I'm like, we have got tochange.
And so we couldn't find any.
(09:51):
And he told me, look, you needto build this yourself.
And so I started uh gettinginto um coding software, and I
do have you know some backgroundin that just as a hobbyist, but
it's turned into quite apassion project.
And while I was developing ourplatform, I also was developing
(10:11):
the tools that we had been usingourselves and I implemented it,
and that's where that AI layercame in.
So I built the electronichealth record and the AI
software together in unison.
And um, what is verymarketable, scalable, but also
increases patient care, which isthe most important thing, is
(10:34):
that AI software.
So offering that same layer weput on our own EHR to other EHRs
fixes a lot of those problemsthat we've been so frustrated
about trying to reach out topatients, even if it's a, you
know, I send the patient amessage and I and it breaks down
from there, or uh, I need thepatient to know I've made a
(10:56):
change to their chart.
These are all really smalladministrative things that would
be common sense that are justnot in the in the pipeline right
now.
And the AI layering um fillsall these holes so much nicely
and it integrates very easily toany platform.
And so we called the AIsoftware Talora, and that is our
(11:20):
um that is our AI unit that weuse, and it doesn't do any
medical care, all it really doesis it you know formats, it
takes all the patient, you know,intake and history and
everything, and it presents itto the provider in one strategic
(11:40):
location right in front ofthem, so they have all the
information they need beforethey've even clicked on the
patient.
And um, at the same time, whatit does is I've built protocols
in place that areevidence-based, meaning this is
the standard of care.
Don't go outside of this ever.
And so when a patient hassomething that is not safe to
(12:03):
treat or it's contraindicated,the system will notify the
provider so the provider doesn'thave to go search through all
the records, pull up theirup-to-date or their
evidence-based guidelines, it'sall pulled into one picture for
the provider so they canconfidently say that they're
treating that patient at thehighest level of their medical
(12:26):
license.
And that is not a actuallydifficult thing to do, it just
hasn't been done.
And so the patient, you canspend the entire visit with the
patient.
You don't have to spend it inthe chart, you don't have to
spend it gathering data, lookingup things yourself.
You spend the entire visit withthe patient, and that is very
rewarding.
(12:46):
It's very rewarding.
Dr Andrew Greenland (12:48):
Amazing.
So, in this new phase of yourcareer, what does a typical day
look like for you now, balancingmedicine with actually building
the technology?
Dr James Ries (12:56):
Yes, well, a lot of after hours coding.
I'm a father of three littleones.
I, you know, I'm my I have someclinic days, I have some
administrative days, and so myclinic days are very much
reserved for just that when I'mon base and I'm treating those
patients.
But um, you know, there is alot of fill time, and uh and I
(13:20):
have the AI coding pretty muchalways pulled up somewhere, and
there's a list of tasks thatneed to be completed that day.
And when the kiddos are asleepin bed and and I know that the
you know the dog's been walkedand everything is quiet, then I
get on the computer and I startdoing as much coding as I can
that day.
(13:40):
That's typically my work day,which has been, you know, if you
build it right from the groundup, it's gonna be a really rough
couple months.
And then afterwards, if you'vebuilt it correctly, the main the
maintenance on that is is notso robust, you know, and you can
confidently say that care isbeing administered correctly.
(14:01):
And so I'm uh I'm now at thatfinish line.
It feels very good.
Dr Andrew Gree (14:06):
Congratulations, amazing.
Dr James Ries (14:08):
Yes, yes, no more late night coding sessions.
Now I'm um, you know, in theprocess of talking to clients
about why this is good for boththem but also their patients and
why it's very inexpensive touse, you know, and that's a
unique conversation I enjoyhaving with clients.
Is this not something that ishas to be um, you know, breaks
(14:31):
the platform's, you know, piggybank.
It's in fact, most of the timethey're saving quite a bit of
money because they can resolveall those crazy administrative
burdens they've been hacking onto telehealth over the past
decade for no good reason.
Dr Andrew Greenland (14:48):
So a lot of the healthcare conversation in
AI seems to start with what canAI do instead of the doctor?
I know you kind of come at thisfrom an opposite position.
Why do you think replacing theprovider is the wrong problem to
solve?
I know you don't hold thatview, but clearly it's out
there.
Dr James Ries (15:03):
It's the only view I really see,
unfortunately, and it's uheveryone is uh applauds it.
So there's a few problems withthat.
I'm glad you bring it up,Andrew.
So um, you know, AI is designedto be user-friendly, which in
most cases is great.
However, if you have an AIchatbot that is forward-facing
(15:26):
to a patient, sometimes patientsdon't need the treatment
they're seeking, or sometimesthe treatment they're seeking is
harmful.
There is a provider layer thatneeds to be with the AI at some
point that um makes sure thatthey're getting good standard of
care, you know, even if it'snot exactly what they wanted.
(15:50):
We can educate as to why thismight be a better treatment plan
for them.
The AI tool will um make theuser happy, however, that is,
and that's a dangerous thing tohave, right?
So, you know, you have somebodythat, you know, a lot of
telehealth does drug and alcoholabuse, you know, and if you
(16:13):
have someone coming in thatsays, Well, I'm out of my pain
medicine, it's reasonable forme, right?
Chatbot?
Yes, it is, absolutely.
And you know, I've heardstories um about AI chatbots and
users uh asking it otherquestions outside of that
clinical console, that then itjust kind of goes off the rails.
So it's a dangerous tool, andunfortunately, it's not
(16:36):
scalable, you know.
So if you use an AI chat toolor an AI generating tool, it may
get it right a dozen times, mayget it right a hundred times,
but a hundred thousand timesit's gonna get it wrong, and
that one time it gets wrong isis when somebody gets hurt.
And so the two systems at playright now in the in the world
(17:01):
are generative AI anddeterministic AI.
Generative AI is somethingwhere you just let the tool do
what it needs to do and it cangenerate based on that.
Deterministic is you haveprotocols in place that um uh
medically say, don't everrecommend anything outside of
(17:21):
this protocol, or this patientis being seen for this clinical
visit, unless an emergencysymptom comes up, please only go
to this.
So deterministic means it's notactually a lot of AI, it's
actually a lot of clinicalauthoring in the background that
you're um telling the AI touse.
And so it's determining atreatment plan based on your uh
(17:46):
clinical authorship.
And that's a different toolaltogether.
You're not replacing theprovider, you're actually making
the provider smarter and moreefficient and more, you know,
yeah, uh, their time is morewith the patient.
And the providers aren't theones um soaking up all the cost.
(18:08):
So you really don't need toreplace them, you need to make
them better and less exhausted.
That's what needs to reallyhappen uh first.
If we ever get to a day wherethe costs have been cut so much
by using AI and you really wantto replace the provider, I think
that's an entirely differentconversation at that time.
(18:29):
But we're not anywhere close toit.
We're not anywhere close tomaking that leap.
Um, and so we really have toresist uh you know that urge to
do that.
Dr Andrew Greenland (18:40):
So on that point, if provider cost isn't
itself necessarily thefundamental problem, where are
the real inefficiencies andcosts coming from that people
perhaps don't know about ordon't think about?
Dr James Ries (18:49):
Yeah, and this might change per country, but
you know, advertising cost ishuge.
Um, medication cost is reallybig.
Um, you know, and AI tools canhelp because they can um, you
know, they can search and theycan uh market these things by
finding pharmacies that are alittle bit more cost effective
(19:10):
or you know, those kind ofthings.
Um the overall system ofpatient communication is very
poor.
Uh, and because it's so poor,you have to uh hire, you know,
staff to do these things just tojust to notify a provider that
a message was sent.
Whether, you know, whether ornot that's relevant or not, it
(19:32):
doesn't matter, but you have tohire a whole staff of people.
And uh AI can easily notify aprovider that there's a message
there, you know.
So that's that's one example.
Um, who needs follow-ups forwhat reason and at what time?
Uh, you know, right now there'sno real AI thought process.
It's I'll I'll just hire anurse and a medical assistant to
(19:54):
staff that full time.
Um, and you know, there's a lotof processes that don't involve
having to call the patient.
And those processes are greatfor AI because um you don't need
to use a human voice for thatbecause they're not contacting
the patient.
And that's such as remindersfor follow-up appointments, or
(20:14):
if a lab gets put into thesystem, what happens with it?
You know, do we need to reallyhire someone to click through
buttons to put that in the rightplace?
Probably not.
So um the list goes on and onfor administrative things
happening in the background thatdon't even need a human voice
to talk to the to the patient.
Dr Andrew Greenland (20:36):
And what do you think um people
misunderstand about where aclinician actually creates
value?
Because there seems to be thiskind of thing where everybody's
looking, well, you know, doctorclinician versus AI.
Where are people missing thepoint here?
Dr James Ries (20:50):
Yeah, I think people are missing the point of
the, you know, the tone of thevisit, the tone of the patient.
Um, you know, patient historythat might come up at a a later
time.
The overall relationship fromhuman to human is is different,
and it always will be because umthere's certain limitations
(21:14):
that you can only sense as a asa human, talking to another
human, or even just typing toanother human.
A lot of context is is uh is istaken out of context, you know.
I think everyone can I thinkeveryone can uh empathize with a
text message that you send to afriend that they take the wrong
tone with, and now they thinkthey're mad, you're mad at them
(21:37):
when really it was a sarcasticjoke, you know.
So there's lots of things thatrequire a human touch.
Also, um, you if you're gonnause AI to replace the provider,
you're going to remove theexperience that the provider has
by treating human patients formany, many years.
(21:58):
Um when I When I say that, Imean what is the patient's life
like?
Where do they live?
Are they in the middle ofnowhere?
Do they live with theirparents?
Do they live with somebodyelse?
Are they going through adivorce?
You know, these these are smalllittle trinkets that um need a
human context.
And if an AI tool is is um isjust focusing on one algorithm
(22:23):
that they're programmed to do,you're gonna miss those contexts
that might change the way youtreat that patient.
Dr Andrew Greenland (22:31):
Thank you.
So let's make this tangible.
Um imagine I'm a patiententering your system for the
first time.
Talk us through what happens umfrom a patient perspective so
we can get an understanding ofhow the platform kind of fits
together.
Dr James Ries (22:44):
Yeah, of course.
Yeah, so if you come into 20Mile, we'll um we'll use like an
urgent care as an example.
So if you come into the um 20Mile Medical with an urgent care
need, you will uh click aseries of service lines that are
relevant to your chiefcomplaint or the reason you're
(23:05):
going there.
Anything from men's health tomental health to urgent care.
And when you click the visit toopen, it will then uh triage
what type of urgent care thatyou need help with.
And then the um after yousubmit from the drop-down menu,
you'll be presented with anintake.
Uh and the intake is veryrobust.
(23:28):
First, it presents a safetyscreen should you be here at all
as telehealth, and it's thingslike suicidal ideation.
Am I short of breath?
Am I having chest pain?
Those things instantly andimmediately get routed to an
emergency room or a 911 visit,um, which is also great because
people who are waiting to go tothe emergency room for their
(23:50):
symptoms don't they get triagedfaster, right?
So that's one benefit thattelehealth does, and it has
nothing to do with profits.
It's people that need to get tothe emergency room but you know
basically don't want to go forvarious reasons.
They'll use telehealth first,but it's good to get told
immediately that that's asymptom that need that in fact
(24:13):
does need to go, and they'rethey're more likely to seek
emergency care if they gettriage properly.
So you go through a series ofsafety checks, um, and then if
it seems appropriate at thatpoint, they then start filling
out the details of theirsymptoms.
We use uh the intakes are umevidence-based.
(24:33):
So the questions we're askinghave real impact on as to what
treatment type that they'regonna get, whether it's an
antibiotic or whether it'sconservative care or whether
it's a um you know a three-dayfollow-up or something to that
degree.
So they're asking uh so they'reanswering a series of
questions, and then they'll alsoput in the medical history,
(24:56):
their medications, and and alltheir details.
Once they hit submit on that,then the AI system will um
compile that data.
And while it's compiling thedata, it's also looking for red
flags.
So it's looking for things thatuh the provider needs to not
miss.
Uh, and again, that'sevidence-based items.
(25:17):
And of course, if it you know,based on the um clinical
authorship, if it you know, itwill present whether the patient
uh you know should be offeredthis form of treatment versus
another.
It's a recommendation to theprovider.
Once that's done, then the umit reaches a an awaiting room
(25:40):
for us, and we click on thevisit when it when the AI is
done with uh doing what it needsto do.
And when we click the visit,there's option for um you know
intake uh driven, which isasynchronous, which means if the
intake is robust and thepatient has filled out the
appropriate um you knowinformation and there's no real
(26:02):
risks, then we can communicatewith the patient at that point.
We can reach out to them andsay, this is what I've uh found
out and this is what I'mrecommending, this is why.
Um the AI will also bring upevidence-based citations uh such
as you know, based on thisclinical guideline that we use
(26:22):
here at 20 Mile, you wouldqualify for an antibiotic course
because it puts you at lowerrisk.
Um we can then submit theprescription directly to the
pharmacy within minutes and theycan go directly to the
pharmacy.
Other uh you know, treatmentswe offer men's health, women's
(26:42):
health, we do directly shiptheir medication to their home.
And so that's something that isnice.
If they need labs for whateverreason, we can also order from
there.
If they need to see a videovisit, so to get a proper
screen, we can click the videovisit and it sends them a link
to open and we can see them liveright there on the spot.
(27:04):
So we have a lot of tools, andthat's just kind of an overall
summary on how we do like abasic type urgent care um uh
visit.
But it's very quick, it's veryefficient, and um, because those
things are taken um off thetable, we can spend as much time
as we want chatting with thepatient, which is really nice.
(27:26):
You know, we can go back andforth uh, you know, until we're
both you know very uh happy andsatisfied and uh on board with
the treatment plan.
We send those treatment plansbefore we send the medication so
we can say, this is what Ithink.
Well, you know, what do youthink?
Are you okay with this?
Do you have a good follow-upplan?
And then we can send themedication that way.
(27:48):
Um, so it's a very nice system.
My providers are very happy.
Yeah.
Dr Andrew Greenland (27:53):
Brilliant.
Well, this um ties in verynicely to my next question
because on our previousconversation, you were talking
about mental or cognitive loadfor clinicians.
And obviously, you've designedsomething which is very good at
alleviating that.
Can you speak a bit more tothat and and how you've managed
to enable clinicians to see morepatients just by reducing the
cognitive load of all of thatadministration?
Dr James Ries (28:15):
Yeah, there's a couple, there's a couple strings
to pull on that.
And one, you know, the firstone is providers in general, you
know, we get some instantgratification from helping
others.
And the more times that youcould trigger that, uh happier
we become, you know.
So at first, when you ask aprovider to see 20 patients an
(28:37):
hour, they immediately go, oh mygosh, I'm gonna be exhausted.
But if you design a systemwhere they're just getting the
instant, you know, feedback andthe instant gratification and
none of the cumbersome steps,uh, they actually prefer that.
They actually feel veryproductive and useful and they
feel like they've treated a lotof people well.
(28:58):
Um, it all started with tryingto uh make other systems more
efficient.
And I found that when I didthat by using some of these
tools, the providers were askingme personally for more patients
and more hours, you know,because typically if they have
(29:18):
to write the entire soap note orfind the information or look
things up, a couple hours oftelehealth is enough to really
wear them out.
Uh, and what I found when Iimplemented these tools is they
would go four, five hours andnot be exhausted because all the
information was presented infront of them and they were able
(29:39):
to just talk to the patient.
And the exhausting part is nottalking to the patient, it's
getting all the information.
Uh, and so that's how we'vetaken that burden.
I didn't really anticipate thewhole cognitive load until I saw
it, until I saw it live.
I just thought it was going tobe uh safer and more uh
efficient and better for thepatient to have this stuff taken
(30:02):
care of ahead of time.
But what I didn't anticipate isthat providers were not worn
out, seeing you know, 20 or morean hour.
They were actually very happywith it.
Dr Andrew Greenland (30:12):
So brilliant.
And there is a kind of a bit ofa counterintuitive thing here.
Because I think when we spokebefore, you said that making
that interaction more efficientcould actually allow the patient
to receive more care ratherthan less.
Can you unpick that a littlebit for us?
Because it does feel a littlebit counterintuitive if the
patient's only getting threeminutes with the clinician.
Dr James Ries (30:30):
Yeah, yeah, exactly.
Um, well, if we are so focusedon gathering the information,
looking up guidelines, or doingthese things, uh, unfortunately,
our mental load is going tomiss a lot of other aspects.
If you present all thatinformation at the same time in
front of the provider, then itallows them the breathing room
(30:54):
to think outside the box.
And so they get a little bitmore creative, but in a good
way.
Things like added supplementsor, you know, what is your
living arrangement?
You know, where are you um, youknow, are you impoverished?
Can you afford this?
These questions come up becausethe rest of the care is taken
(31:17):
care of quickly.
Um, and so the patient gets amuch higher level of care, more
uh customized care.
So by using the AI layeringtool, you're really um stepping
outside of the traditional, youknow, uh the patient fits inside
this box, and this is themedicine I'm gonna give them.
(31:38):
If you remove that, then theprovider thinks, well, what else
can I do?
How what other things can I do?
Because I got to thisconclusion very quickly.
And so they tend to offer morethings to the patient because
they feel like, well, that wasthat was easy.
That was an easy visit.
What other things can I ask?
(31:58):
And so the patient ends upgetting much, much higher level
of care, a lot, a more a lotmore personal, you know.
We'll ask them some extraquestions that we honestly, just
due to exhaustion, probablynever would have asked them.
You know, if we spent all thattime looking all those things
up, we think to ourselves, okay,I need to finish this visit now
so I can get to the nextpatient.
(32:18):
But if all that's done for youalready, uh, gives you a little
breathing room.
And even in even in three toyou know, three to five minutes,
you can ask a lot of personalquestions in those few minutes
if you don't have to do anythingelse.
Dr Andrew Greenland (32:32):
Brilliant.
I think you also gave me theexample of potentially moving
from quarterly follow-ups tomonthly follow-ups.
What sort of impact could thatincreased frequency have, in
your opinion?
Dr James Ries (32:43):
Yes, that's a good, that's a really good
question.
We um, you know, at first wewere doing three-month
follow-ups because that seems tobe the standard, but when you
have a system that is veryefficient, it allows you to
perform those follow-ups a lotsooner.
Um, between your care andbetween your three-month visit,
(33:03):
patients have a lot of questionsthat tend to build up slowly.
And then during theirthree-month visit, they tend to
overload the physician with allof those questions that they've
had.
If you provide them a moreregular tool, it it gives them
the ability to have thoseanswered on a more regular
basis.
Um, and the follow-ups are uheven more efficient.
(33:27):
And so if you're doing this ona on a scale, it's quite easy to
pull that follow-up in, andthen the patient um perceives
that as a much higher level ofof care, but like actually
caring for them, not justmedical care.
Um, and so when we pulled the30-day follow-ups, we felt the
(33:48):
physicians and the providersfelt like they were having a lot
less questions, when reallythey were probably just less
questions all at once.
You know, and and so theproviders are happier because by
doing a 30-day follow-up,they're only getting a few
questions.
But if they're doingthree-month follow-ups, they're
getting inundated with a lot.
(34:10):
And so uh higher frequency offollow-ups is a higher frequency
of visits, but less questionsper visit.
And providers typically arehappy with that.
Dr Andrew Greenland (34:20):
Thank you.
So, pulling back from your ownplatform for a moment, what
major shifts are you seeing inhealthcare and telehealth more
widely at the moment?
Dr James Ries (34:30):
I'm sorry, repeat that question.
What makes me think telehealthshould be used more widely?
Dr Andrew Greenland (34:33):
No, no, I think saying um pulling back
from your own platform, what areyou seeing in terms of major
shifts in healthcare andtelehealth specifically right
now?
Elsewhere.
Dr James Ries (34:42):
Yeah.
Yeah.
So I think you know, telehealthhas had its run when it comes
to urgent care.
I think COVID solved thatproblem a lot.
It made people realize that umthey could administer care uh
safely from a distance.
Uh, I think people, you know,bought off on that idea.
(35:03):
And now I think um it's moreconcierge type model.
I think people are starting tothink to themselves, well, if I
could get seen for a cold, well,you know, what else can I get
seen for?
Do I really need to go to theprovider in person for annual
labs?
Probably not, you know, and sopeople are expanding their minds
(35:24):
a little bit as to what can beaccomplished over a video or a
phone visit.
And they really should.
I mean, really, they should becritical of what's necessary and
what's not.
You don't want unnecessarythings.
That's what drives healthcarecosts up.
Um, and so now I'm seeingpeople use it for annual
wellness visits, I'm uh a lot oflifestyle things, right?
(35:45):
So people who uh would normallynot be comfortable talking to a
provider in person about, youknow, uh sexual dysfunction are
now more opt to fix that.
Uh, and they are happy to doso.
Uh, and so we're getting a lotmore into lifestyle, right?
And so we're getting intoweight loss, we're getting into,
(36:06):
you know, even you know, thechanging platform and
environment when it comes topeptide treatment.
So people are starting to uhconnect the dots as to how else
can I improve my life using thisbecause it's a lot easier,
which I encourage.
I think if people are seekingout ways to improve their life
and they feel like they'rehealthy enough to not have to go
(36:28):
to a doctor um, you know, everythree months, then we can offer
them, you know, we can kind ofuh dial in their health, you
know, this is how we getstronger, this is how we lose
weight, this is how and so Ithink it's more a lot more
lifestyle and concierge, butthat's just people taking
ownership of their health.
Dr Andrew Greenland (36:47):
Thank you.
Obviously, building what youbuilt is an amazing achievement.
I can't believe it's all beenstraightforward.
What have been some of thechallenges and bottlenecks along
the road that made it kind ofum tricky?
Dr James Ries (36:58):
Okay, I guess coding software, my goodness.
Um, getting systems to talk toone another is a real is a real
chore.
You just have to do it once.
If you do it once the rightway, then it'll work out.
But you know, overall systemsthey're not designed to to talk
to each other.
Uh, and so getting that all onone, getting everyone, every
(37:21):
system that you code on the samepage is very very difficult.
Um, also just learning theprocess of becoming a coder
from, you know, that's somethingthat was definitely unique, and
I've learned a ton.
Uh, I really enjoy doing it,which is also something I
learned.
(37:42):
I did it out of necessity and Iended up really enjoying it.
And so it's uh it's somethingthat I plan to continue doing
because I get a lot ofsatisfaction from it.
Um, seeing something, you know,seeing something that you
create and then seeing it liveon um a patient or a website or
a portal or an intake is veryrewarding to me because you you
(38:06):
did all that little hard worklate at night and now you can
see it actually happen, likesomething actually happened from
all those little letters I putin, you know.
So um, I'm trying to think ofother things.
Really just learning thesystem.
I will say it's a lot easier todo now, you know.
With AI systems, I don't thinkeven just a year or two ago I
would be able to do what I did.
(38:27):
I think uh systems now code foryou and they help guide that.
You give them what youroverarching goal is, your
medical experience, and what youwant out of the product, and it
will help you to code for that.
And so um, I think a lot ofmore people should be doing this
(38:47):
from a physician standpoint,because the the gap between
software developer and physicianuh is closing.
And I think that the quickerthat closes, the better it is
for patients.
Thank you.
Dr Andrew Greenland (39:02):
And from a kind of founder-developer
perspective, what are themetrics you're most interested
in tracking with your platform?
Dr James Ries (39:10):
Yeah, so a few metrics.
The nice thing about the waythat we go about it and using
these AI tools in the backgroundis you're uh you're privy to a
lot of data that you otherwisewouldn't track because it's too
cumbersome, like patientoutcomes.
Um, you know, even if it's assomething as like this is the
(39:31):
number of pounds lost, or umthese this percent of people had
side effects on my um on myplatform.
You can gather data on almostany input that you want.
And that is a unique experiencebecause normally that would be
quite cumbersome to try andgather.
(39:51):
And now it can be a few minutesof coding and you can have real
live feedback on processimprovement, whether your
programs are actually working ornot, and whether the patients
are are coming back, you know,and so um that's really unique.
I really am I'm gonna enjoyseeing that uh blossom over
time, you know, have actuallyhaving the data we can use.
(40:14):
I think that there's gonna be alot of um research that needs
to be done on peptides andweight loss that now we have
access to as long as people areusing these systems correctly.
Dr Andrew Greenland (40:26):
So, what about the future?
What where do you want to be insort of 12 months' time with 20
Mile Medical?
What does the future hold?
Dr James Ries (40:32):
Yes, so 20 Mile Medical, we want to be, you
know, we want to be in theemployer space, uh, which is a
unique space.
I feel like the the employersare kind of uh left to deal with
whatever insurance, you know,contractors that they have to
deal with.
I feel like uh we could takecare of a population of people
(40:56):
very well by having this dataand seeing the outcomes and
seeing how each individualunique you know employer um can
benefit their uh their smallpopulation of workers.
So I think that that'sinteresting um dynamic.
We're just getting into thatspace a little bit.
Um but the overarching goal andwhy I did all this was to
(41:20):
eventually, and this sounds kindof crazy, but what I'm
realizing is that the AI toolsand an efficient platform and an
efficient way to reach thepatient, if you can do this
correctly, the cost of actuallyreaching care to people becomes
(41:40):
so minimal that I think that itcan almost be free.
And if I can scale Talora and20 mile, I would like to
actually start a purely freehealth care for telehealth that
is um donation-based.
(42:00):
Yeah, I think I can I think Ican get the cost down enough,
but it would be almost like anonprofit for you know homeless
or uninsured.
Dr Andrew Greenland (42:11):
Wow, what an amazing um aim to be able to
have.
That's fantastic.
With that, James, I'd love tothank you for joining me this
afternoon.
It's been such an interestingconversation.
But I think it really reframeswhat we mean to when we talk
about AI making healthcare moreefficient, because efficiency
doesn't necessarily have to meanless care, less time, or
removing the clinician.
If the technology can take awaythe grind, the documentation,
(42:33):
the administrative workload, andthe cognitive burden
surrounding the consultation,perhaps the result is actually a
provider who can provide morecare, have more capacity to do
so in the first place.
So I'm very grateful for theconversation and thank you again
for joining us.
Dr James Ries (42:48):
Yeah, thank you, Andrew.
It was fun.
So, welcome back to Voices in Health and
Wellness, where we speak withpractitioners and healthcare
innovators about the challenges,opportunities, and ideas
shaping the future ofhealthcare.
Today we're exploringartificial intelligence, but
from a slightly differentperspective, rather than asking
how AI could replace healthcareprofessionals, what if we used
it to remove the administrativeburden surrounding them, giving
(00:23):
clinicians more time andcapacity to focus on patient
care?
And could doing that not onlyimprove efficiency and access,
but also reduce one of the majorcontributors to provide a
burnout?
My guest today is Dr.
James Ries, a board-certifiedfamily medicine physician,
active duty US Navy provider,and healthcare entrepreneur
behind 20 Mile Medical.
(00:43):
James is developing anAI-enabled healthcare platform
designed to automate much of theadministrative workload
surrounding clinical care whilekeeping the healthcare
professional firmly at thecenter of clinical decision
making.
So with that, James, I'd loveto welcome you to the show and
thank you very much indeed forjoining me this afternoon.
Dr James Ries (01:00):
Yeah, thank you, Andrew, for having me.
This is a great opportunity totalk about this very unique
subject.
So thank you.
Dr Andrew Greenland (01:06):
So, James, before we get into the AI and
the technology itself, can youtake us back a little and what
was your journey into medicineand what originally attracted
you to family medicine?
Dr James Ries (01:16):
Yeah, that's a great question.
I actually initially wanted tobe into pediatrics, but then of
course, you know, as youprogress through medical school
and as you do all your medicalrotations, you realize that, or
at least some people realize,they like a little bit of
everything.
And they and the other thing isuh, you know, telling a patient
that I'm not trained to seethem is something that I just
(01:39):
didn't uh didn't sit right withme.
So family medicine was alogical good fit, you know, from
from cradle to grave, you know,taking care of everyone and and
not turning anyone away wassuch a good motivator.
Dr Andrew Greenland (01:53):
Thank you.
And you've also combinedmedicine with active duty and
with a naval career.
And how has that experienceshaped the way you think about
healthcare delivery andefficiency?
Dr James Ries (02:03):
Yeah, quite a different animal.
Um, I was actually veryinterested in just uh aviation,
and that's what got me into themilitary pipeline.
But since I've been activeduty, I've realized that you
know providing excellent levelof care, no matter where you
are, is uh a different thing.
Sometimes you can't controlyour brick and mortar
(02:26):
environment.
Sometimes you can't controlwhere you're gonna be
administering that care.
And so that um really lendsitself well to telehealth,
certainly, but it lends itselfto all kinds of um unique
problems that you have to solve.
So you become a problem solver,which is a really interesting
(02:48):
way to go about it.
But I'm really happy that Idid.
You know, um, whether you're ona ship in the middle of the
ocean or whether you're in adesert, you know, things change,
but the standard, you know, youwant to maintain it as much as
possible.
So the military career has beenquite rewarding, a lot of fun
experiences, but I have about ayear left and I'm looking
(03:10):
forward to transitioning becauseuh keeping that tempo is quite
difficult with a family andeverything else.
You know, it's it's fun, but itdoesn't last forever, and
that's a good understanding tohave.
Dr Andrew Greenland (03:22):
Thank you.
And your point about problemsolving.
So, at what point did you startmoving from simply recognizing
problems in healthcare tothinking, I actually want to
build something that solvesthem?
Dr James Ries (03:32):
Yeah, that is an excellent question.
I I got into telehealth.
I've been doing it for about adecade, but really it took off
and piqued my interest duringCOVID because people had little
to no access to care.
Sometimes they couldn't, youknow, they weren't even allowed
to, you know.
So um when that all changed, itmade me realize that there were
(03:55):
that there were people outthere that would otherwise not
access medical care unless youtook down some barriers.
Even getting in the car anddriving is a barrier for some
people.
And if you provide somethingthat can still provide them a
good standard of care and goodhigh-level care uh care, no
(04:17):
matter where they're located,you reach a large segment of the
population that would otherwisego untreated.
Um, a really good example ismental health.
You know, psychiatrists askpeople to come to the clinic for
care, and sometimes theirspecific phobia is leaving their
house.
You know, so there's a bigbarrier there.
(04:38):
And there's also people thatare located in the middle of
nowhere, and they just reallydon't know what to do, even with
something as simple as a cold.
And um, and so that really,really piqued my interest uh
when it came to telehealth.
And I saw a very un untreatedlarge pop global population of
(04:59):
people, and that became a newproblem to solve.
Dr Andrew Greenland (05:04):
So you talked about a couple of
problems that you've solved, butwere there any actual
frustrations from the work thatyou've been doing as a
practicing physician that alsodrove you to create what you've
created in the platform?
Dr James Ries (05:15):
Excellent question.
Yeah, I kind of missed that.
Um, so when I started withCOVID, um, it started basically
with you know synchronizedvisits, phone calls, but very
quickly I realized that you knowthe the standard was really um
relied on the provider.
(05:35):
And that variability is whatgets people hurt, you know, and
and the systems themselves werevery incredibly inefficient to
the point where just to get onepatient seen, you would have to
staff that whole program withcustomer service and pharmacy
reps and you know all kinds ofsystems that normally wouldn't
(05:58):
be there, and systems thatbecome very cumbersome that now
have to talk to each other.
And I think that after um,after all those systems are in
place and you and you have itrunning, it becomes expensive.
And so you really have to, froma first principles type of
thinking, start over, notpiggyback on what's been done,
(06:21):
but start over and think, well,how can I do this efficiently so
that we're not having to leanon all that administrative
burden?
You know, uh the providerprobably only gets about 10% of
the fee.
The rest goes to all theadministration, all the all the
(06:43):
uh scheduling and all thecoordination.
And that really is the umthat's the low-hanging fruit.
That's what needs to be cutfirst.
And I think what's happeningnow is people are still relying
on these old systems and they'retrying to think how can we cut
costs, how can we become moreefficient?
(07:03):
And the target seems to go tothe provider first because AI is
very good at writing a soapnote.
But whether it can administergood health care, that's a
different question.
And there's systems in placenow that are not administering
healthcare.
And so you don't have to worryabout that when it comes to
(07:24):
scheduling and you know,coordination and and
documentation.
You know, these things can bestreamlined and you don't run
the risk of losing that standardof care.
And and I think it's just needsto be kind of re-rethought from
the ground up because we have anew system in place now.
Dr Andrew Greenland (07:44):
Thank you.
So before we sort of dive alittle bit more into the
platform on your work, could youperhaps give us a sense of what
your role is and how it fitsinto the bigger picture of 20
Mile Medical?
Dr James Ries (07:54):
Yeah, of course.
Yeah, 20 Mile Medical.
So I've been working with many,many platforms doing telehealth
care.
And I my uh co-owner is a PA.
He also does this.
And after a while of creatingjust systems for ourselves to
make things more efficient andbetter care, better patient
(08:16):
communication, he had said, youknow, we should probably just do
this ourselves because we'respending a lot of time fixing
other people's systems.
So he encouraged me.
He actually was the catalystfor all this, and we started our
own telehealth platform withall the improvements and all the
patient care that we had beenforcing other platforms to to
(08:40):
fix, and we just did itourselves.
And I'm very uh grateful forthat because we've done a
fantastic job, and everyone isis very appreciative on the
platform.
While we were um while we werecreating the platform, we had to
um use software.
And when we started usingsoftware, we realized after
interviewing a dozen or moredifferent software platforms
(09:05):
that they were all verydeficient, and they were all
developed by um you knowsoftware developers, not
physicians, and they were alldeveloped with the patient uh
advertising and patient facing,um, but not anything to do with
the provider and how the care isactually administered.
And we we landed on one clientand we've been using them for a
(09:30):
while, but man, it's they theymake things very limiting for
the patient-providerinteraction, even as something
as simple as I need to change mycredit card, you know, becomes
like a monumental, you know,this this mountain that we have
to climb up.
And I'm like, we have got tochange.
And so we couldn't find any.
(09:51):
And he told me, look, you needto build this yourself.
And so I started uh gettinginto um coding software, and I
do have you know some backgroundin that just as a hobbyist, but
it's turned into quite apassion project.
And while I was developing ourplatform, I also was developing
(10:11):
the tools that we had been usingourselves and I implemented it,
and that's where that AI layercame in.
So I built the electronichealth record and the AI
software together in unison.
And um, what is verymarketable, scalable, but also
increases patient care, which isthe most important thing, is
(10:34):
that AI software.
So offering that same layer weput on our own EHR to other EHRs
fixes a lot of those problemsthat we've been so frustrated
about trying to reach out topatients, even if it's a, you
know, I send the patient amessage and I and it breaks down
from there, or uh, I need thepatient to know I've made a
(10:56):
change to their chart.
These are all really smalladministrative things that would
be common sense that are justnot in the in the pipeline right
now.
And the AI layering um fillsall these holes so much nicely
and it integrates very easily toany platform.
And so we called the AIsoftware Talora, and that is our
(11:20):
um that is our AI unit that weuse, and it doesn't do any
medical care, all it really doesis it you know formats, it
takes all the patient, you know,intake and history and
everything, and it presents itto the provider in one strategic
(11:40):
location right in front ofthem, so they have all the
information they need beforethey've even clicked on the
patient.
And um, at the same time, whatit does is I've built protocols
in place that areevidence-based, meaning this is
the standard of care.
Don't go outside of this ever.
And so when a patient hassomething that is not safe to
(12:03):
treat or it's contraindicated,the system will notify the
provider so the provider doesn'thave to go search through all
the records, pull up theirup-to-date or their
evidence-based guidelines, it'sall pulled into one picture for
the provider so they canconfidently say that they're
treating that patient at thehighest level of their medical
(12:26):
license.
And that is not a actuallydifficult thing to do, it just
hasn't been done.
And so the patient, you canspend the entire visit with the
patient.
You don't have to spend it inthe chart, you don't have to
spend it gathering data, lookingup things yourself.
You spend the entire visit withthe patient, and that is very
rewarding.
(12:46):
It's very rewarding.
Dr Andrew Greenland (12:48):
Amazing.
So, in this new phase of yourcareer, what does a typical day
look like for you now, balancingmedicine with actually building
the technology?
Dr James Ries (12:56):
Yes, well, a lot of after hours coding.
I'm a father of three littleones.
I, you know, I'm my I have someclinic days, I have some
administrative days, and so myclinic days are very much
reserved for just that when I'mon base and I'm treating those
patients.
But um, you know, there is alot of fill time, and uh and I
(13:20):
have the AI coding pretty muchalways pulled up somewhere, and
there's a list of tasks thatneed to be completed that day.
And when the kiddos are asleepin bed and and I know that the
you know the dog's been walkedand everything is quiet, then I
get on the computer and I startdoing as much coding as I can
that day.
(13:40):
That's typically my work day,which has been, you know, if you
build it right from the groundup, it's gonna be a really rough
couple months.
And then afterwards, if you'vebuilt it correctly, the main the
maintenance on that is is notso robust, you know, and you can
confidently say that care isbeing administered correctly.
(14:01):
And so I'm uh I'm now at thatfinish line.
It feels very good.
Dr Andrew Gree (14:06):
Congratulations, amazing.
Dr James Ries (14:08):
Yes, yes, no more late night coding sessions.
Now I'm um, you know, in theprocess of talking to clients
about why this is good for boththem but also their patients and
why it's very inexpensive touse, you know, and that's a
unique conversation I enjoyhaving with clients.
Is this not something that ishas to be um, you know, breaks
(14:31):
the platform's, you know, piggybank.
It's in fact, most of the timethey're saving quite a bit of
money because they can resolveall those crazy administrative
burdens they've been hacking onto telehealth over the past
decade for no good reason.
Dr Andrew Greenland (14:48):
So a lot of the healthcare conversation in
AI seems to start with what canAI do instead of the doctor?
I know you kind of come at thisfrom an opposite position.
Why do you think replacing theprovider is the wrong problem to
solve?
I know you don't hold thatview, but clearly it's out
there.
Dr James Ries (15:03):
It's the only view I really see,
unfortunately, and it's uheveryone is uh applauds it.
So there's a few problems withthat.
I'm glad you bring it up,Andrew.
So um, you know, AI is designedto be user-friendly, which in
most cases is great.
However, if you have an AIchatbot that is forward-facing
(15:26):
to a patient, sometimes patientsdon't need the treatment
they're seeking, or sometimesthe treatment they're seeking is
harmful.
There is a provider layer thatneeds to be with the AI at some
point that um makes sure thatthey're getting good standard of
care, you know, even if it'snot exactly what they wanted.
(15:50):
We can educate as to why thismight be a better treatment plan
for them.
The AI tool will um make theuser happy, however, that is,
and that's a dangerous thing tohave, right?
So, you know, you have somebodythat, you know, a lot of
telehealth does drug and alcoholabuse, you know, and if you
(16:13):
have someone coming in thatsays, Well, I'm out of my pain
medicine, it's reasonable forme, right?
Chatbot?
Yes, it is, absolutely.
And you know, I've heardstories um about AI chatbots and
users uh asking it otherquestions outside of that
clinical console, that then itjust kind of goes off the rails.
So it's a dangerous tool, andunfortunately, it's not
(16:36):
scalable, you know.
So if you use an AI chat toolor an AI generating tool, it may
get it right a dozen times, mayget it right a hundred times,
but a hundred thousand timesit's gonna get it wrong, and
that one time it gets wrong isis when somebody gets hurt.
And so the two systems at playright now in the in the world
(17:01):
are generative AI anddeterministic AI.
Generative AI is somethingwhere you just let the tool do
what it needs to do and it cangenerate based on that.
Deterministic is you haveprotocols in place that um uh
medically say, don't everrecommend anything outside of
(17:21):
this protocol, or this patientis being seen for this clinical
visit, unless an emergencysymptom comes up, please only go
to this.
So deterministic means it's notactually a lot of AI, it's
actually a lot of clinicalauthoring in the background that
you're um telling the AI touse.
And so it's determining atreatment plan based on your uh
(17:46):
clinical authorship.
And that's a different toolaltogether.
You're not replacing theprovider, you're actually making
the provider smarter and moreefficient and more, you know,
yeah, uh, their time is morewith the patient.
And the providers aren't theones um soaking up all the cost.
(18:08):
So you really don't need toreplace them, you need to make
them better and less exhausted.
That's what needs to reallyhappen uh first.
If we ever get to a day wherethe costs have been cut so much
by using AI and you really wantto replace the provider, I think
that's an entirely differentconversation at that time.
(18:29):
But we're not anywhere close toit.
We're not anywhere close tomaking that leap.
Um, and so we really have toresist uh you know that urge to
do that.
Dr Andrew Greenland (18:40):
So on that point, if provider cost isn't
itself necessarily thefundamental problem, where are
the real inefficiencies andcosts coming from that people
perhaps don't know about ordon't think about?
Dr James Ries (18:49):
Yeah, and this might change per country, but
you know, advertising cost ishuge.
Um, medication cost is reallybig.
Um, you know, and AI tools canhelp because they can um, you
know, they can search and theycan uh market these things by
finding pharmacies that are alittle bit more cost effective
(19:10):
or you know, those kind ofthings.
Um the overall system ofpatient communication is very
poor.
Uh, and because it's so poor,you have to uh hire, you know,
staff to do these things just tojust to notify a provider that
a message was sent.
Whether, you know, whether ornot that's relevant or not, it
(19:32):
doesn't matter, but you have tohire a whole staff of people.
And uh AI can easily notify aprovider that there's a message
there, you know.
So that's that's one example.
Um, who needs follow-ups forwhat reason and at what time?
Uh, you know, right now there'sno real AI thought process.
It's I'll I'll just hire anurse and a medical assistant to
(19:54):
staff that full time.
Um, and you know, there's a lotof processes that don't involve
having to call the patient.
And those processes are greatfor AI because um you don't need
to use a human voice for thatbecause they're not contacting
the patient.
And that's such as remindersfor follow-up appointments, or
(20:14):
if a lab gets put into thesystem, what happens with it?
You know, do we need to reallyhire someone to click through
buttons to put that in the rightplace?
Probably not.
So um the list goes on and onfor administrative things
happening in the background thatdon't even need a human voice
to talk to the to the patient.
Dr Andrew Greenland (20:36):
And what do you think um people
misunderstand about where aclinician actually creates
value?
Because there seems to be thiskind of thing where everybody's
looking, well, you know, doctorclinician versus AI.
Where are people missing thepoint here?
Dr James Ries (20:50):
Yeah, I think people are missing the point of
the, you know, the tone of thevisit, the tone of the patient.
Um, you know, patient historythat might come up at a a later
time.
The overall relationship fromhuman to human is is different,
and it always will be because umthere's certain limitations
(21:14):
that you can only sense as a asa human, talking to another
human, or even just typing toanother human.
A lot of context is is uh is istaken out of context, you know.
I think everyone can I thinkeveryone can uh empathize with a
text message that you send to afriend that they take the wrong
tone with, and now they thinkthey're mad, you're mad at them
(21:37):
when really it was a sarcasticjoke, you know.
So there's lots of things thatrequire a human touch.
Also, um, you if you're gonnause AI to replace the provider,
you're going to remove theexperience that the provider has
by treating human patients formany, many years.
(21:58):
Um when I When I say that, Imean what is the patient's life
like?
Where do they live?
Are they in the middle ofnowhere?
Do they live with theirparents?
Do they live with somebodyelse?
Are they going through adivorce?
You know, these these are smalllittle trinkets that um need a
human context.
And if an AI tool is is um isjust focusing on one algorithm
(22:23):
that they're programmed to do,you're gonna miss those contexts
that might change the way youtreat that patient.
Dr Andrew Greenland (22:31):
Thank you.
So let's make this tangible.
Um imagine I'm a patiententering your system for the
first time.
Talk us through what happens umfrom a patient perspective so
we can get an understanding ofhow the platform kind of fits
together.
Dr James Ries (22:44):
Yeah, of course.
Yeah, so if you come into 20Mile, we'll um we'll use like an
urgent care as an example.
So if you come into the um 20Mile Medical with an urgent care
need, you will uh click aseries of service lines that are
relevant to your chiefcomplaint or the reason you're
(23:05):
going there.
Anything from men's health tomental health to urgent care.
And when you click the visit toopen, it will then uh triage
what type of urgent care thatyou need help with.
And then the um after yousubmit from the drop-down menu,
you'll be presented with anintake.
Uh and the intake is veryrobust.
(23:28):
First, it presents a safetyscreen should you be here at all
as telehealth, and it's thingslike suicidal ideation.
Am I short of breath?
Am I having chest pain?
Those things instantly andimmediately get routed to an
emergency room or a 911 visit,um, which is also great because
people who are waiting to go tothe emergency room for their
(23:50):
symptoms don't they get triagedfaster, right?
So that's one benefit thattelehealth does, and it has
nothing to do with profits.
It's people that need to get tothe emergency room but you know
basically don't want to go forvarious reasons.
They'll use telehealth first,but it's good to get told
immediately that that's asymptom that need that in fact
(24:13):
does need to go, and they'rethey're more likely to seek
emergency care if they gettriage properly.
So you go through a series ofsafety checks, um, and then if
it seems appropriate at thatpoint, they then start filling
out the details of theirsymptoms.
We use uh the intakes are umevidence-based.
(24:33):
So the questions we're askinghave real impact on as to what
treatment type that they'regonna get, whether it's an
antibiotic or whether it'sconservative care or whether
it's a um you know a three-dayfollow-up or something to that
degree.
So they're asking uh so they'reanswering a series of
questions, and then they'll alsoput in the medical history,
(24:56):
their medications, and and alltheir details.
Once they hit submit on that,then the AI system will um
compile that data.
And while it's compiling thedata, it's also looking for red
flags.
So it's looking for things thatuh the provider needs to not
miss.
Uh, and again, that'sevidence-based items.
(25:17):
And of course, if it you know,based on the um clinical
authorship, if it you know, itwill present whether the patient
uh you know should be offeredthis form of treatment versus
another.
It's a recommendation to theprovider.
Once that's done, then the umit reaches a an awaiting room
(25:40):
for us, and we click on thevisit when it when the AI is
done with uh doing what it needsto do.
And when we click the visit,there's option for um you know
intake uh driven, which isasynchronous, which means if the
intake is robust and thepatient has filled out the
appropriate um you knowinformation and there's no real
(26:02):
risks, then we can communicatewith the patient at that point.
We can reach out to them andsay, this is what I've uh found
out and this is what I'mrecommending, this is why.
Um the AI will also bring upevidence-based citations uh such
as you know, based on thisclinical guideline that we use
(26:22):
here at 20 Mile, you wouldqualify for an antibiotic course
because it puts you at lowerrisk.
Um we can then submit theprescription directly to the
pharmacy within minutes and theycan go directly to the
pharmacy.
Other uh you know, treatmentswe offer men's health, women's
(26:42):
health, we do directly shiptheir medication to their home.
And so that's something that isnice.
If they need labs for whateverreason, we can also order from
there.
If they need to see a videovisit, so to get a proper
screen, we can click the videovisit and it sends them a link
to open and we can see them liveright there on the spot.
(27:04):
So we have a lot of tools, andthat's just kind of an overall
summary on how we do like abasic type urgent care um uh
visit.
But it's very quick, it's veryefficient, and um, because those
things are taken um off thetable, we can spend as much time
as we want chatting with thepatient, which is really nice.
(27:26):
You know, we can go back andforth uh, you know, until we're
both you know very uh happy andsatisfied and uh on board with
the treatment plan.
We send those treatment plansbefore we send the medication so
we can say, this is what Ithink.
Well, you know, what do youthink?
Are you okay with this?
Do you have a good follow-upplan?
And then we can send themedication that way.
(27:48):
Um, so it's a very nice system.
My providers are very happy.
Yeah.
Dr Andrew Greenland (27:53):
Brilliant.
Well, this um ties in verynicely to my next question
because on our previousconversation, you were talking
about mental or cognitive loadfor clinicians.
And obviously, you've designedsomething which is very good at
alleviating that.
Can you speak a bit more tothat and and how you've managed
to enable clinicians to see morepatients just by reducing the
cognitive load of all of thatadministration?
Dr James Ries (28:15):
Yeah, there's a couple, there's a couple strings
to pull on that.
And one, you know, the firstone is providers in general, you
know, we get some instantgratification from helping
others.
And the more times that youcould trigger that, uh happier
we become, you know.
So at first, when you ask aprovider to see 20 patients an
(28:37):
hour, they immediately go, oh mygosh, I'm gonna be exhausted.
But if you design a systemwhere they're just getting the
instant, you know, feedback andthe instant gratification and
none of the cumbersome steps,uh, they actually prefer that.
They actually feel veryproductive and useful and they
feel like they've treated a lotof people well.
(28:58):
Um, it all started with tryingto uh make other systems more
efficient.
And I found that when I didthat by using some of these
tools, the providers were askingme personally for more patients
and more hours, you know,because typically if they have
(29:18):
to write the entire soap note orfind the information or look
things up, a couple hours oftelehealth is enough to really
wear them out.
Uh, and what I found when Iimplemented these tools is they
would go four, five hours andnot be exhausted because all the
information was presented infront of them and they were able
(29:39):
to just talk to the patient.
And the exhausting part is nottalking to the patient, it's
getting all the information.
Uh, and so that's how we'vetaken that burden.
I didn't really anticipate thewhole cognitive load until I saw
it, until I saw it live.
I just thought it was going tobe uh safer and more uh
efficient and better for thepatient to have this stuff taken
(30:02):
care of ahead of time.
But what I didn't anticipate isthat providers were not worn
out, seeing you know, 20 or morean hour.
They were actually very happywith it.
Dr Andrew Greenland (30:12):
So brilliant.
And there is a kind of a bit ofa counterintuitive thing here.
Because I think when we spokebefore, you said that making
that interaction more efficientcould actually allow the patient
to receive more care ratherthan less.
Can you unpick that a littlebit for us?
Because it does feel a littlebit counterintuitive if the
patient's only getting threeminutes with the clinician.
Dr James Ries (30:30):
Yeah, yeah, exactly.
Um, well, if we are so focusedon gathering the information,
looking up guidelines, or doingthese things, uh, unfortunately,
our mental load is going tomiss a lot of other aspects.
If you present all thatinformation at the same time in
front of the provider, then itallows them the breathing room
(30:54):
to think outside the box.
And so they get a little bitmore creative, but in a good
way.
Things like added supplementsor, you know, what is your
living arrangement?
You know, where are you um, youknow, are you impoverished?
Can you afford this?
These questions come up becausethe rest of the care is taken
(31:17):
care of quickly.
Um, and so the patient gets amuch higher level of care, more
uh customized care.
So by using the AI layeringtool, you're really um stepping
outside of the traditional, youknow, uh the patient fits inside
this box, and this is themedicine I'm gonna give them.
(31:38):
If you remove that, then theprovider thinks, well, what else
can I do?
How what other things can I do?
Because I got to thisconclusion very quickly.
And so they tend to offer morethings to the patient because
they feel like, well, that wasthat was easy.
That was an easy visit.
What other things can I ask?
(31:58):
And so the patient ends upgetting much, much higher level
of care, a lot, a more a lotmore personal, you know.
We'll ask them some extraquestions that we honestly, just
due to exhaustion, probablynever would have asked them.
You know, if we spent all thattime looking all those things
up, we think to ourselves, okay,I need to finish this visit now
so I can get to the nextpatient.
(32:18):
But if all that's done for youalready, uh, gives you a little
breathing room.
And even in even in three toyou know, three to five minutes,
you can ask a lot of personalquestions in those few minutes
if you don't have to do anythingelse.
Dr Andrew Greenland (32:32):
Brilliant.
I think you also gave me theexample of potentially moving
from quarterly follow-ups tomonthly follow-ups.
What sort of impact could thatincreased frequency have, in
your opinion?
Dr James Ries (32:43):
Yes, that's a good, that's a really good
question.
We um, you know, at first wewere doing three-month
follow-ups because that seems tobe the standard, but when you
have a system that is veryefficient, it allows you to
perform those follow-ups a lotsooner.
Um, between your care andbetween your three-month visit,
(33:03):
patients have a lot of questionsthat tend to build up slowly.
And then during theirthree-month visit, they tend to
overload the physician with allof those questions that they've
had.
If you provide them a moreregular tool, it it gives them
the ability to have thoseanswered on a more regular
basis.
Um, and the follow-ups are uheven more efficient.
(33:27):
And so if you're doing this ona on a scale, it's quite easy to
pull that follow-up in, andthen the patient um perceives
that as a much higher level ofof care, but like actually
caring for them, not justmedical care.
Um, and so when we pulled the30-day follow-ups, we felt the
(33:48):
physicians and the providersfelt like they were having a lot
less questions, when reallythey were probably just less
questions all at once.
You know, and and so theproviders are happier because by
doing a 30-day follow-up,they're only getting a few
questions.
But if they're doingthree-month follow-ups, they're
getting inundated with a lot.
(34:10):
And so uh higher frequency offollow-ups is a higher frequency
of visits, but less questionsper visit.
And providers typically arehappy with that.
Dr Andrew Greenland (34:20):
Thank you.
So, pulling back from your ownplatform for a moment, what
major shifts are you seeing inhealthcare and telehealth more
widely at the moment?
Dr James Ries (34:30):
I'm sorry, repeat that question.
What makes me think telehealthshould be used more widely?
Dr Andrew Greenland (34:33):
No, no, I think saying um pulling back
from your own platform, what areyou seeing in terms of major
shifts in healthcare andtelehealth specifically right
now?
Elsewhere.
Dr James Ries (34:42):
Yeah.
Yeah.
So I think you know, telehealthhas had its run when it comes
to urgent care.
I think COVID solved thatproblem a lot.
It made people realize that umthey could administer care uh
safely from a distance.
Uh, I think people, you know,bought off on that idea.
(35:03):
And now I think um it's moreconcierge type model.
I think people are starting tothink to themselves, well, if I
could get seen for a cold, well,you know, what else can I get
seen for?
Do I really need to go to theprovider in person for annual
labs?
Probably not, you know, and sopeople are expanding their minds
(35:24):
a little bit as to what can beaccomplished over a video or a
phone visit.
And they really should.
I mean, really, they should becritical of what's necessary and
what's not.
You don't want unnecessarythings.
That's what drives healthcarecosts up.
Um, and so now I'm seeingpeople use it for annual
wellness visits, I'm uh a lot oflifestyle things, right?
(35:45):
So people who uh would normallynot be comfortable talking to a
provider in person about, youknow, uh sexual dysfunction are
now more opt to fix that.
Uh, and they are happy to doso.
Uh, and so we're getting a lotmore into lifestyle, right?
And so we're getting intoweight loss, we're getting into,
(36:06):
you know, even you know, thechanging platform and
environment when it comes topeptide treatment.
So people are starting to uhconnect the dots as to how else
can I improve my life using thisbecause it's a lot easier,
which I encourage.
I think if people are seekingout ways to improve their life
and they feel like they'rehealthy enough to not have to go
(36:28):
to a doctor um, you know, everythree months, then we can offer
them, you know, we can kind ofuh dial in their health, you
know, this is how we getstronger, this is how we lose
weight, this is how and so Ithink it's more a lot more
lifestyle and concierge, butthat's just people taking
ownership of their health.
Dr Andrew Greenland (36:47):
Thank you.
Obviously, building what youbuilt is an amazing achievement.
I can't believe it's all beenstraightforward.
What have been some of thechallenges and bottlenecks along
the road that made it kind ofum tricky?
Dr James Ries (36:58):
Okay, I guess coding software, my goodness.
Um, getting systems to talk toone another is a real is a real
chore.
You just have to do it once.
If you do it once the rightway, then it'll work out.
But you know, overall systemsthey're not designed to to talk
to each other.
Uh, and so getting that all onone, getting everyone, every
(37:21):
system that you code on the samepage is very very difficult.
Um, also just learning theprocess of becoming a coder
from, you know, that's somethingthat was definitely unique, and
I've learned a ton.
Uh, I really enjoy doing it,which is also something I
learned.
(37:42):
I did it out of necessity and Iended up really enjoying it.
And so it's uh it's somethingthat I plan to continue doing
because I get a lot ofsatisfaction from it.
Um, seeing something, you know,seeing something that you
create and then seeing it liveon um a patient or a website or
a portal or an intake is veryrewarding to me because you you
(38:06):
did all that little hard worklate at night and now you can
see it actually happen, likesomething actually happened from
all those little letters I putin, you know.
So um, I'm trying to think ofother things.
Really just learning thesystem.
I will say it's a lot easier todo now, you know.
With AI systems, I don't thinkeven just a year or two ago I
would be able to do what I did.
(38:27):
I think uh systems now code foryou and they help guide that.
You give them what youroverarching goal is, your
medical experience, and what youwant out of the product, and it
will help you to code for that.
And so um, I think a lot ofmore people should be doing this
(38:47):
from a physician standpoint,because the the gap between
software developer and physicianuh is closing.
And I think that the quickerthat closes, the better it is
for patients.
Thank you.
Dr Andrew Greenland (39:02):
And from a kind of founder-developer
perspective, what are themetrics you're most interested
in tracking with your platform?
Dr James Ries (39:10):
Yeah, so a few metrics.
The nice thing about the waythat we go about it and using
these AI tools in the backgroundis you're uh you're privy to a
lot of data that you otherwisewouldn't track because it's too
cumbersome, like patientoutcomes.
Um, you know, even if it's assomething as like this is the
(39:31):
number of pounds lost, or umthese this percent of people had
side effects on my um on myplatform.
You can gather data on almostany input that you want.
And that is a unique experiencebecause normally that would be
quite cumbersome to try andgather.
(39:51):
And now it can be a few minutesof coding and you can have real
live feedback on processimprovement, whether your
programs are actually working ornot, and whether the patients
are are coming back, you know,and so um that's really unique.
I really am I'm gonna enjoyseeing that uh blossom over
time, you know, have actuallyhaving the data we can use.
(40:14):
I think that there's gonna be alot of um research that needs
to be done on peptides andweight loss that now we have
access to as long as people areusing these systems correctly.
Dr Andrew Greenland (40:26):
So, what about the future?
What where do you want to be insort of 12 months' time with 20
Mile Medical?
What does the future hold?
Dr James Ries (40:32):
Yes, so 20 Mile Medical, we want to be, you
know, we want to be in theemployer space, uh, which is a
unique space.
I feel like the the employersare kind of uh left to deal with
whatever insurance, you know,contractors that they have to
deal with.
I feel like uh we could takecare of a population of people
(40:56):
very well by having this dataand seeing the outcomes and
seeing how each individualunique you know employer um can
benefit their uh their smallpopulation of workers.
So I think that that'sinteresting um dynamic.
We're just getting into thatspace a little bit.
Um but the overarching goal andwhy I did all this was to
(41:20):
eventually, and this sounds kindof crazy, but what I'm
realizing is that the AI toolsand an efficient platform and an
efficient way to reach thepatient, if you can do this
correctly, the cost of actuallyreaching care to people becomes
(41:40):
so minimal that I think that itcan almost be free.
And if I can scale Talora and20 mile, I would like to
actually start a purely freehealth care for telehealth that
is um donation-based.
(42:00):
Yeah, I think I can I think Ican get the cost down enough,
but it would be almost like anonprofit for you know homeless
or uninsured.
Dr Andrew Greenland (42:11):
Wow, what an amazing um aim to be able to
have.
That's fantastic.
With that, James, I'd love tothank you for joining me this
afternoon.
It's been such an interestingconversation.
But I think it really reframeswhat we mean to when we talk
about AI making healthcare moreefficient, because efficiency
doesn't necessarily have to meanless care, less time, or
removing the clinician.
If the technology can take awaythe grind, the documentation,
(42:33):
the administrative workload, andthe cognitive burden
surrounding the consultation,perhaps the result is actually a
provider who can provide morecare, have more capacity to do
so in the first place.
So I'm very grateful for theconversation and thank you again
for joining us.
Dr James Ries (42:48):
Yeah, thank you, Andrew.
It was fun.
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
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
The Joe Rogan Experience
The official podcast of comedian Joe Rogan.