Dr. Ruben Amarasingham, Founder and CEO of PCCI

Mark Masse ll i: This is Conversations o n Health Care . I am Mark Masselli.

Margaret Flinter: And I am Marg a ret Flin ter .

Mark Masselli: Well Marg a r et , we have seen some challenges in the startup of
the online insurance marketplaces on the national level . Since the opening of the
st
exchanges October 1 there have been quite a few technical issues that have led
to difficulties across the countr y for those trying to navigate the online
marketplace in 36 states relying on the federal exchanges . And a few go od
notes in terms of some of the individual states but across the bo ard Marg a ret , I
think it’s not been a shining moment.

Margaret Flinter: Well we were warned that ther e would be difficulties Mark , that
we would have hoped for a perfectly smooth rollout . And maybe it was to be
expect ed with such a large system turning into existence and I bet there are lot of
people doing a retrospective on where the various bum p s in decisions were th at
weren’t so good in retrospect . But we do know this that the leader s at the White
House and Health and Human Services are working around the clock to fix the
fl a ws in the system , and that’s going to be a ongoing process.

Mark Masselli: It is . Robert Pe ar of the New York Times had a huge piece the
other day certainly ta lking about the problems , so there is a lot of finger pointing
on t hey have to get this thing fixed . And I think the pol ls show that only 25% of
those who tr ied to navigate the exchanges had a satisfactory experience. So
there is a lot of work to be done and let’s hope they get right to it.

Margaret Flinter: That’s right . And I do want to shine a little ray of optimism on
things , Mark . We no te that this hasn’t been the case in those states that s et up
the i r own exchanges . There were some technical gli tches in the state - base d
marketplaces in New York, California and our State of Connecticut and others ,
but it seems that the problems we re worked out much more quickly. Tens of
thousand s of Californian have already signed up for insurance in the first few
weeks , 40,000 New York residen ts have gained coverage and here in
Connecticut , we have certainly had the experience of seeing people both having
a chance to educate themselves by going to the online exchange and sign up for
coverage. But of course , these are the states that are also showing more
wil ling ness to help facilitate the experience for their residen ts and I do think that
has a positive impact. People do need a l ittle hand - holding in understanding it.

Mark Masselli: Well , hopefully one of the ke y redesign s is allowing people to go
through without having s etup an account, no different like you buy a car or a
house. You go in, you look at it, you take it a round for a spin and you look at the
price and then you come back to it. So that ’s a fundame ntal re design that
hopefully will happen with the federal exchanges . So there is some work to do
with some big decisions on the long term implication as they la y this redesign
out.

Margaret Flinter: Well most experts agree also that the numbers of Americ ans
who will be gai ning insurance on the exchanges is going to pick up greatly in the
next month ahead. We should remind our listeners that they have until
th st
December 15 to sign up in time for their coverage to begin on January 1 . So
st
open enrollment do es continue until March 31 but if you want to be covered on
th
January 1 you have got to complete the process and enroll by December 15 .
So that’s plenty of time to both educate yourself and shop around .

Mark Masselli: Some of the problems have certain ly been mas ked by the
shutdown Congress has had and dealing with the debt limit , dealing with this
year’s budget , dealing with last year’s s equester Marg a ret . All th ose have an
enormous impact. They certainly have a huge impact in the health care arena .
The National Institute of Health normally welcomes 200 new patients per week
into clinical trials , has had to turn most of those patients away including critically
ill children . It’s still too early to a ss ess how much damage ha s been done with
the governm ent shutdown but there is no doubt it’s had an impact on the health
of many Americans .

Margaret Flinter: I couldn’t agree with you more , Mark . And our guest today is
an individual who is working to have a very positive impact on the health of many
Ameri cans , Mark . Dr. Ruben Amarasingham is an Internist , Founder and CEO of
a PCCI , a center for innovation at the Parkland Health & Hospital System in
Dallas, Texas . He talks about the program he has cr e ated that has shown some
very promising results in redu cing costly hospital admissions, using predictive
analysis and health data analytics.

Mark Masselli: You know , there are hundreds of hospitals across the country
who are looking into the system he has developed so some interesting
innovations out there M arg a ret . W e w ill also get a visit from L o r i Robertson ,
Managing Editor of Fact C heck.org . She is looking at false claims about the
health care law.

Margaret Flinter: And no ma tter what the topic , you can hear all of our shows by
Googling CHC Radio , and as always , if you have comments , email us a t
www.chcradio.com or find us on Facebook or Twitter , we love hearing from you.
We will get to our interview with Dr. Amarasingham in just a moment.
Mark Masselli: But f irst , here is our producer Maria n n e O’ Hare with this week ’ s
Headlines News.
(Music) Maria n n e O’ Hare : I am Maria n n e O’ Hare with the se Healthcare Headlines .
Obamacare is taking it on the chin in the opening weeks of open enrollment. A
quick check of the federal online insurance exchanges ha s shown only roughly a
quarter of those who have attempted to log on to the federal exchange have

r ated the experience as favorable . The system has gone down more than it’s
been up and folks are having difficulty getti ng past the point where they can
create an account. Essentially , there are unpredictable results among users
across the 36 states being covered by the federal portal , Health C are.gov , but
there is progress being made in getting to the root of the technical problems.
According to a published report in the New York Times , Obama Administration
official s feel the system is about 70% on it s way to being fully functional . But
there were widely varying views on when that remaining 30% of the problems
with the in surance marketplaces would be fixed .

Confidential progress reports from the Health and Human Services Department
showed senior offic ial s repeatedly expressed doubts t hat the computer systems
for the federal exchange would be ready on time. The largest co ntractor was
awarded its contract in December 2011 but apparently , the government was so
slow to issuing specifications , they didn’t start writing code until this spring , and
changes were being made to specification s up to the last minute. While
customers are having an easier time within the state - base d exchanges , the re are
still complain t s , chief among them being the cost of the plans themselves which
do v a ry widely from state to state. Another big problem across states and federal
exchanges is many in t he low er economic brackets lack an email address which
is required for the online signups.

Medica id expansion is advancing in several states . Ohio and New Hampshire
are taking up that expansion as a way of getting more vulnerable residen ts the
coverage t hey need and qualifying for federal dollars to underwrite the expansion
as well. 27 million Americans might be seeking help in the health care a rena for
knee problems . The estimated number of those Americans grappling with knee
troubles, glucosamine and chondroitin has long been thought to have a beneficial
effect , tested first in horse s and then in dogs before being mass market ed to
humans as a natural solution. But according to a study published in the New
England Journal of Medicine, it turns out ther e is not much difference in knee
function improvement or pain reduction from glucosamine than from placebo
pills . What’s the best source of relief according to this study? Losing weight and
exercising. They had a more lasting effect on knee improvement. I am Marianne
O’Hare, with these Healthcare Headlines .

(Music)

Mark Masselli: We are speaking today with Dr. Ruben Amarasingham , Founder ,
President and CEO of PCCI , a non - profit research and development corporation
in Dallas , Texas that specializes in real time health predictive analytics. Dr.
Amarasingham is an intern ist and is the Director of Biomedical Informatics for the
University of Texas Southwestern , NIH Center for Translational Science . He is
also an Associate Professor in the Departments of G eneral Internal Medicine and
Clinical Sciences at the University of Texas Southwestern Medical Center . Dr.
Amarasingham was a Robert W ood Johnson Clinical Scholar and was recently

given one of the Robert W ood Johnson Foundation Young Leader Awards for his
exceptional contribution for improving health care. Dr. Amarasingham , welcome
to Conversation s o n Health Care .

Dr. Ruben Amarasingham : Thank you. It ’ s a real pleasure to be here and feel
free to call me Ruben.

Mark Masselli: Ruben , great perfect. Y ou have created a ne w tool for reducing
hospital readmissions that’s gaining quite a bit of national attention , and in fact ,
several hundred hospitals have made inquiries about your program , and many
hopefully are in the process of adopting it . And you ha ve developed a system in
clinical setting that respon d s to the needs of large vulnerable populations . And
w hile you are also an intern ist, serving largely poor and underserved population
in Dallas , you are also a skilled coder and data analyst who s aw rea l opportunity
in thi s area of creating a solution. So can you describe for us that clinical
environment you we re working in before that led to the development of your
predictive analytic system , aimed at reducing those costly hospital admissions ?

Dr. Rub en Amarasingham: During my fellowship at John Hopkins Medical
Institution , I had a chance to work with some re nown ed individuals and focus on
how you use Information Technology to improve the care of the patient and
specifically how Information Technology ma y give rise to predictive modeling or
the ability to use data to identify what may happen to a patient or health system
in advance of any adverse outcome . Then when I returned to Dallas , and to ok a
position at Parkland Health and Hospital System at UT Southwestern Medical
Center, one of my first assignment s as a medical director was to review the
charts of patients who were readmitted to the hospital and it’s just sort of a
standard practice . And in my review of these patients’ charts and looking at wh at
the f actors were th at might have cause d a readmission , some of the things that
really jumped out at me obviously was first ther e are enormous s ocio economic
determina nts of health , where a patient lives, what kind of access to food , their
education and l iteracy status, as well as clinical options that are generated for the
patient . And it became clear to me that this would be a great example of the use
of electronic predictive modeling because there are factors that you can identify
that seem ed to be con sistent among patients that were getting readmitted . And
from that came this idea of can we model the social determin ants of health
directly from information on the electronic medical record and can we also
understand and model the physiolog ic or clinical factors that affect a patient and
use that to say which patient s might be at a high er risk for readmissions . And so
that was really the genesis of my work with respect to developing this software
and this modeling.

Margaret Flinter: Ruben , I am very in trigued by what you just said and also
intrigued by the experiences that set people out on their respective paths in life .
And your experience as a forth year medical student with Parkland ’s Home less
Medical Outreach Service was another one of those kind of pivotal experiences

that shaped your future work , study ing the ho m eless population in Dallas and
finding th at half the homeless population you we re serving were women and
children. So when you look at those socioeconomic factors , I would like for our
l isteners to help them get a real sense of how this works , and t ell us how those
things factor in not for people who have been in the hospital but maybe for
people who are at the community level , how do you collect those socioeconomic ,
social determin ants o f health from the community and primary care and use them
to prevent those hospitalizations from happening in the first place . Has that been
part of your approach?

Dr. Ruben Amarasingham: That is , and actually , there ha s just been great work
across the United States trying to solve these problems , recognizing just as you
d escribed that socioeconomic determina nts, the environment you live in, your
family status, your housing situation for children and then obviously the same
kinds of factors for adults. All of those factors play a huge role in whether or not
you can maintain a healthy environment for yourself and preven t needing to go to
the hospital . And so people are trying to understand well can we measure that,
how can we use that data , try to improv e communit ies, improve patient s’ health,
and there is a lot of very exciting initiative s going on.

In our community , one of th e things that we are developing is called the Dallas
Information Exchange Portal , and the goal there is to create a network of
in formation systems that connect data from the hospital system s , from individual
primary care clinics and then from social organizations like (12:44 inaudible) and
the Salvation Army, and data from individuals ’ phones and other data that if they
want to part icipate in the se system s to try to help manage their care, that they
c an be plugged into this larger network . What that would allow is for us to really
have a precise understanding of what’s available to an individual in the
community to help social and c ommunity organizations work together with health
care providers, to try to improve all the things that are traditionally outside of the
health care sector but there is no question that it has a big impact on health . And
I think if we are able to link up t he school systems with the health systems and
with these community - based organizations that are providing social service s,
food and nutri tion and adequate housing , when we combine all of these great
effort s that people are making in a combined community ef fort , then I think we
can make a real difference for our patients . And we have a major effort going on
in Dallas . I think there are other efforts in other cities , and as these develop, I
think there is incredible potential for these technologies to help these vulnerable
populations.

Mark Masselli: Well there is lots of potential but there is also lots of hard work . It
seems that you have your head in the clouds and I s ay that with all due respect
that you are working in the area of big data which is so rt of new to the health care
field , it certainly ha s been used in a lot of other areas . And then you are in the
weeds if you will on technology and the health community has not been q uick to
adopt health technology , certainly not like the banking industry , so there is a lot

of issues and you have talked about them in Dallas making sure that the system s
talk to each other , they are interoperable and responsive and they can do things
in real time . So your software system seems to be designed to sort of addr ess
some of these . It’s called PCs f orecast adverse clinical events , in particular , the
likelihood of hospital readmissions using real time data from electronic medical
records . But there is also a surveillance component to your system . So how
does that work exactly and also tell us about some of the innovations that were
applied in developing it.

Dr. Ruben Amarasingham: You know (14:51 inaudible) is a software application
that s its on top of an electronic medical record. So I think the idea or the an alogy
might be for those people that have smartphones . There’s usually a n app store
and there are apps that you can purchase that kind of sit o n top of your phone
and allow you to do all these ne at functions . And so in this case , we have
developed someth ing similar what we call sort of a grand application that can sit
on top of any electronic medic record. It takes the data in real time from this
record , interprets and uses the data to understand what might happen to a
patient before it happen s . So obvi ously , one aspect that we focused on is can
w e predict with a high level of accuracy who has a very high chance of having a
readmission, and then once w e identify that , we can direct resources that really
help with r e ducing readmission of that patient.

Bu t there are lot of other approaches like trying to prevent septic shock mortality,
trying to reduce the likelihood of having a cardio pulmonary arrest in the hospita l.
We have been working on programs to reduce the inciden ts of chronic kidney
disease and d iabetes complications w hen patients go home . So it’s been
fascinating to work on bringing all of this vast rich data that’s collected as part of
the normal health care experience and trying to use that data , interpret it through
the computers to say , hey we can really make a difference on this patient here
that has th ese particular problems and here are potential recommendations.
Some of the innovations that we a re using th at ha ve come from other industries
and from the ma thematical and computer s cience w orlds are Natural Language
Processing , which is the ability for computers and a lgorithms to read the notes of
doctors and nurses and understand what the y are concerned about with the
patient . We use different advance d mathematical modeling techniques and
machine l earning to kind of interpret and define huge amounts of data in split
seconds . And so I think it’s a really exciting time to be doing this work because I
think the infrastructure of electronic medical record s is being adopted across the
country a nd I think technologies like this really allow you to take advantage of this
infrastructure that’s be ing developed.

Margaret Flinter: Well I was just sitting here wondering how many years will it be
before we stop thinking of e lectronic health record s as a big deal . I think what
people really care about is what doe s this mean for m e and how does it matter to
me in terms of the safety and the quality of the care I get. And I wonder if you
could comment, I know that you completed a very significant study for the

Archives of Internal Medicine or th at was published in the Archives of Internal
Medicine, l ooking a t data from a 170,000 Texas patient s and you found a very
significant impact of the u se of electronic health records on patient outcomes . I
wonder i f you c ould tell us about those findings.

Dr. Ruben Amarasingham: E ssentially, we looked at 41 hospitals that had
different stages of these electronic medical record a d option and we looked at
some key components of electronic medical records , how are tes t result s
document ed. Our physician s and nurses we re able to put their n otes in
electronic form as opposed to written from, w ere they able to get help with
d ecisions what’s com monly called decision support , then we looked at what was
the death rate and wh at was the complication rate, what was th e sort of cost that
w as expected, and generally across the bo ard we found that i f the technology
systems were well - constructed and had those features , there was a dramatic and
impressive reduction in important outco mes like death . And also th at features
like decision support , computers helping the physicians make better decision s or
at least interpret data, vast amounts of data or give them signals when something
might be going wrong, reduce d complications for patie nts . And one of the most
interesting findings we found was that the greater the automation of the health
system , the greater that the information was electronic and easily accessible and
easy to use from the physician ’s perspective, actually lower cost as well . And so
I think that some of what we hope is that as electronic m edical record s are
distributed across the United States and if they are constructed well , if they are
built to be highly usable and enhance the physicians ’ and nurses ’ power to help
pa tients , I think we will see , at least on the basis of this study , we c an see
dramatic results I think overtime in the patient s’ quality of care.

Mark Masselli: We are speaking to day with Dr. Rub e n Amarasingham , Founder ,
President and CEO of the Parkland Center for Clinical Innovation , a nonprofit
research and development corporation in Dallas that specializes in real time
health predictive analytics and surveillance strategies to improve health
outcomes. Dr. Amarasingham is an internist , also Director of the Biomedical
Informatics for the University of Texas Southwestern NIH Center for Translational
Science . Rub e n , you participated in a conference in Washington sponsored by
the Gordon and Betty Moore Foundation with a number of other health system
innova tors delving more deeply into the emerging area of health predictive
analytics. Now we are geeks here, tech geeks, so one, I want to know who was
there, and it sounds like a really interesting sort of thought process that was
going on about s ort of explor ing the big challenges that you are facing in this
discipline and some of the hurdles that you face . So wa lk us through , if you will ,
some high level , I think you talked a little bit about some of them before but can
you delve into that and where you see potential partnerships.

Dr. Ruben Amarasingham: Well it was a wonderful conference and we had
people from – luminaries really from around the United States and i nternationally ,
some statisticians from Europe that have been really working in this area . A nd

we d el ved deeply into several topics th ere related to the predictive modeling ,
what needs to happen to move the field forward. With respect to data barriers,
there were several proble ms that were kind of identified . One was you need
systems that are i nteroperable, technology systems that are interoperable so that
data could flow freely from one system to another to allow for the best
communication and interpretation to allow for the best and highest use of
predictiv e models. In so many industries we h ave just incredible data capture
through the system s that have been developed in defense , in travel, in the
banking industr ies and we kind of consider what are some of the ways that we
can capture more data in a way that would help patients , and some of th e ideas
were home monitoring technologies and data from phones if patient s c onsented
to that. And then I think there was also a recognition that we need to be able to
enable data for developers . Health data, clinical data is very unique ,
idiosyncratic al most, and for us to unleash the in novation that might be possible
with advanced predictive models and even just software that can help patients ,
p eople need a sandbox of de - identifie d or anonymized emerged data to build
th e s e applications. So those w e re s ome of the problems that were kind of
discussed and how we could solve them.

Margaret Flinter: Well Rub e n, I have a question for you as you take a futurist
view . You have painted a very exciting future for the tech folks and the data folks
who have this new sandbox to play in , but I am sure in your hospital institution
and in your work you are on a daily basis seeing the next generation of health
care professionals coming through the doors and we certainly see here th ey are
all digital natives and I have no trouble learning a n electronic health record. But
how is the training of health professionals, how do you see a changing in terms
of people being trained to a model or educate d to a model of looking at things
like predictive scores and predictive mode ling and incorporating these social
determina nts of health, but what you see changing in the whole education and
training p aradigm ?

Dr. Ruben Amarasingham: Well I think clearly, it’s going to demand and require
l ot of change within the whole medical educ ation and nursing and professional
education (22:54 inaudible). There are some early changes to introduce
statistics in epidemiology earlier on in the training process so that people will
understand how these algorithms will work. I think that we are mov ing into an
era where machines and machine - supported recommendations and the
understa nding of statistical methodology as it applies to patient care and decision
making will be paramount. And probably I think there are widespread
discussion s now in the med ical education community about what types of
classes in the past both in the undergraduate premedical program and nursing
program as well as in the graduate s to a medical education paradigm could be
replaced , is some what archaic and needs to be replaced wi th a bet ter
understanding of statistics, epidemiology, data science and computer science .
And I think we will start to see probably the undergraduate medical , premedical

curriculum and the medical school curriculum change in response to the
emergenc e of p redictive models.

Mark Masselli: We are speaking today with Dr. Rub e n Amarasingham , Founder ,
President and CEO of PCCI , a nonprofit research and development corporation
in Dallas , Texas that specializes in real time health predictive analytics . Dr .
Amar asingham is a intern ist and is a Director of Biomedical Informatics for the
University of Texas Southwestern NIH Center for Translational Science . Thank
you so much for joining us on Conversations on Health Care Today .

Dr. Ruben Amarasingham: Thank you.

Mark Masselli: At Conversations on Health Care, we want our audience to be
truly in the know when it comes to the facts about health care reform and policy.
Lori Robertson is an award - winning journalist and managing editor of
FactCheck.org, a nonpartis an, nonprofit consumer advocate for voters that aim to
reduce the level of deception in US politics. Lori, what have you got for us this
week?

Lori Robertson: Well we have seen both side s in this debate over Obamacare
dist orting the facts about premiums . We caution our readers to be vary of claims
about big premium rate decreases or increases , and it’s important to ask
compared with what when politician s make such claims . For instance, days
st
before the health care exchanges launched on October 1 , Pres ident Obama
said that average premium s for the Illinois exchange were 25% lower than what
individu als were able to get previously buying insurance on their own. The
Illinois officials used that figure in comparing exchange rate with what the federal
gover nment had predicted premiums would be . It wasn’t a comparison with
individual market pricing. Obama made a similar comparison with California ,
saying that exchange premium s in that state were about 33% lower. But
California official s said premiums were u p to 29% lower compared with rates for
small employer plans not individual market plans . Official s said they gave that
comparison because both market s wouldn’t deny applicants based on preexisting
condition s, but the President didn’t explain that . Obama also mentioned New
York , and he was right in that comparison . The Governor’s Office said exchange
premiums were at least 50% lower on average than 2013 individual market rates .
And that’s my fact check for this week. I am Lori Robertson , managing editor of
FactCheck.org .

Margaret Flinter: FactCheck.org is committed to factual accuracy from the
country's major political players and is a project of the Annenberg Public Policy
Center at the University of Pennsylvania. If you have a fact, that you would l ike
checked, email us at www.chcradio.com . We will have FactCheck.org's Lori
Robertson check it out for you here on Conversations on Health Care.
(Music)

Mark Masselli: Each week Conversations highlight s a brigh t idea about how to
make wel lness a part of ou r communities and everyday lives. Much emphasis
ha s been paid of la te to the dangers of distracted driving. A number of states
have enacted laws banning texting and driving which has led to a number of
traffi c de aths across the country. But what abut distracted walking ? A recent in -
depth study conducted by S afe K ids.org showed some pretty s tartling statics.
Older teens now account for over half of all p edestrian deaths of children and
one of the main culprit s, distracted walking .

Kate Carr: We saw a 25% increase in teen fatalities within the last five years and
that’s what alarmed us and we did an observational study where we collected
data from over 34,000 observations of middle school er s and high school er s while
they were walki ng in a school zone and crossing the street. We sa w that one in
five high school students were distracted by using their mobile device , they were
either texting , using headphones , or talking on their cell phone . And we saw one
in ei ght middle schoolers also doing the same thing.

Mark Masselli: Kate Carr is President and CEO of SafeKids .org , whos e mission
is to find the best ways to keep kids safe . Of f the tens of thousands of kids they
observe d walking to school in various differe nt neighborhoods , almost 40% were
seen crossing the street while texting, talking on their phone, list ening to music
with headphones or playing some gaming device. She decided there needed to
be a campaign to promote better awareness to reduce the trend a nd they created
a Moment of Silence .

Kate Carr : This campaign is a reminder to especially teens , there are 50% o f the
fatalities in kids under the age of 19 , but for everybody who is distracted while
walking . Put that device down , create a moment of sil ence when you are
crossing a street or you are on a s idewalk or in a parking lot around cars, put
your device down, take your headphone s or your ear buds out.

Mark Masselli: She realize s that kids , especially tee ns, will not be separated
from their mobil e devices but if they could just put them down while walking
across the street or through intersections , engage in a moment of silence , the
number of teen pedestrian deaths would be greatly reduced. She is urging
parents and pediatricians a like to access their site Safe K ids .org for more details
about a Moment of Silence Campaign and bring that message home to kids. A
simple slogan - based campaign to raise awareness about the dangers of
distracted walking that has potential to save the lives of child pedest r ians and
adults for that matter, now that’s a bring idea.

Margaret Flinter: This is Conversations on Health Care . I am Marg a ret Flin t er .

Mark Masselli: And I am Mark Masselli , peace and health.

Conversations on Health Care, broadcast from the campus of WESU at
Wesleyan University, streaming live at www.wesufm.org and brought to you by
the Community Health Center.