[This Transcript is Unedited]
NATIONAL COMMITTEE AND VITAL HEALTH STATISTICS
Hearing on “Meaningful Use” of Health Information
April 28, 2009
Marriott Wardman Park Hotel
2660 Woodley Road, NW
CASET Associates, Ltd.
Fairfax, Virginia 22030
TABLE OF CONTENTS
- Welcome and Roll Call – Harry Reynolds, Jr., Chairman
- Introductory Remarks: David Blumenthal, M.D., M.P.P, National Coordinator for Health Information Technology
- Panel 1: Vision of Health and Health Care Transformed
- Helen Darling, National Business Group
- Elliott S. Fisher, M.D., M.P.H., Dartmouth University
- Stephen Schoenbaum, M.D., M.P.H., The Commonwealth Fund
- Carolyn Clancy, M.D., Agency for Healthcare Research and Quality (AHRQ)
- Panel 2: Meaningful Use Capacity/Functionality in Electronic Health Records
- David Classen, M.D., M.S., Computer Sciences Corporation (CSC)
- Terry McGeeney, M.D., M.B.A., TransforMED
- Blackford Middleton, M.D., M.P.H., MSc, Partners Healthcare
- John Tooker, M.D., M.B.A., FACP, American College of Physicians
- Panel 3: Meaningful Use Capacity/Functionality in Health Information Exchanges
- William Bernstein, Manatt Phelps Phillips
- Rachel Block, New York E-Health Collaborative
- J. Marc Overhage, M.D., Ph.D., Indiana Health Information Exchange
- Gina Perez, Delaware Health Information Exchange
- Panel 4: Meaningful Use Capacity/Functionality in Quality Reporting
- Kristine Martin Anderson, Booz Allen Hamilton
- Theresa Cullen, M.D., M.S., Indian Health Service
- John D. Halamka, M.D., M.S., US Healthcare Information Technology Standards Panel
- Karen Kmetik, Ph.D., American Medical Association
- Panel 5: The Glide Path to Meaningful Use for 2011 and Beyond for Vendors
- Justin Barnes, Greenway Medical
- Glen Tullman, Allscripts
- Mark Segal, Ph.D., General Electric Co.
- Fred Trotter, Open Source Representative
- Panel 6: The Glide Path to Meaningful Use for 2011 and Beyond for Providers
- Nancy Johnson, Baker Donelson
- Jonathan Perlin, M.D., Ph.D., Hospital Corporation of America
- Farzad Mostashari, M.D., NY City Department of Health & Mental Hygiene
- Judy Murphy, R.N., FACMI, FHIMSS, Aurora
- Open Public Comments
P R O C E E D I N G S (9:00 a.m.)
MR. REYNOLDS: You notice we are starting exactly at 9:00. It is a trend we
will continue throughout the next two days. I will make some more comments on
that in a bit.
Good morning. I would like to call this meeting order. This is the first day
of two days of hearings of the National Committee on Vital and Health
Statistics, Executive Subcommittee on “Meaningful Use” of Health
Information Technology. The National Committee is a main public advisory
committee to HHS on national health information policy.
I’m Harry Reynolds, chair of NCVHS. I want to welcome committee members,
HHS staff, and others here in person, and also welcome those listening in on
the Internet. I’d like to remind everyone to speak clearly and into the
Let’s now have introductions around the table. Different than our normal
process, we will not introduce everyone around the room, which would mean we
would break for lunch right after that. For those on the National Committee, I
would ask if you have any conflicts of interest related to any issues coming
before us today, would you please so publicly indicate during your
MR. J. SCANLON: Good morning, everyone. I’m Jim Scanlon. I am the Executive
Staff Director for the full committee, and I am the Acting Assistant Secretary
for Planning and Evaluation at HHS.
MS. GREENBERG: Good morning. I am Marjorie Greenberg from the National
Center for Health Statistics, CDC, and Executive Secretary to the committee.
MR. BLAIR: I’m Jeff Blair from Loveless Clinic Foundation, a member of the
Executive Subcommittee. I am not aware of any conflicts of interest.
MS. TRUDEL: Karen Trudel, Centers for Medicare and Medicaid Services,
liaison to the full committee.
MR. FRIEDMAN: Good morning. Charles Friedman, Deputy National Coordinator
for Health Information Technology.
DR. WARREN: Judy Warren, University of Kansas School of Nursing, member of
the Executive Subcommittee. No conflicts.
DR. W. SCANLON: Bill Scanlon, Health Policy R&D, member of the
Executive Subcommittee. No conflicts.
DR. TRENKLE: Good morning. Tony Trenkle, CMS.
DR. BLUMENTHAL: Good morning. David Blumenthal, National Coordinator for
Health Information Technology.
DR. OVERHAGE: Good morning. Marc Overhage with the Regenstrief Institute
and the Indiana Health Information Exchange, member of the committee. I have no
DR. MIDDLETON: Good morning. Blackford Middleton, Partners Health Care
System and Brigham and Women’s Hospital, member of the full committee. No
DR. SUAREZ: Good morning. I am Walter Suarez. I am with the Institute for
HIPAA/HIT Education and Research, a member of the committee, and no conflicts.
DR. TANG: Paul Tang, Palo Alto Medical Foundation, member of the Executive
Subcommittee. No conflicts.
DR. CARR: Justice Carr, Caritas Christi Healthcare, member of the
subcommittee and no conflicts.
DR. FITZMAURICE: Michael Fitzmaurice, Senior Science Advisor for
Information Technology, the Agency for Health Care Research and Quality,
liaison to the full committee.
DR. FRANCIS: Leslie Francis, Law and Philosophy Pat the University of Utah,
member of the Executive Subcommittee, and no conflicts.
MS. AMATAYAKAL: Margaret Amatayakal, contract writer to the committee.
MR. HOUSTON: John Houston, University of Pittsburgh Medical Center, member
of the committee. I have no conflicts.
MS. BUENNING: Good morning. Denise Buenning, Senior Advisor, CMS, Office of
E-Health Standards and Services and lead staff to the NCVHS Standards
MR. REYNOLDS: I forgot to say, I am Harry Reynolds with Blue Cross Blue
Shield North Carolina, and no conflicts.
I am going to make a few comments that I would like everybody in the
audience to listen to, since many of you are presenters, so I am going to put
some ground rules down that we have set up for this.
But first, I would like to mention that not to long ago, meaningful use
just seemed like a basic phrase. But it appears now to revolutionalize the next
five years and future of health IT.
I would really like to thank a number of people for the concentrated and
concerted effort on this hearing. As some of you are aware, these two days were
originally scheduled as a standard subcommittee for the group. We worked
closely with Jeff Blair and Judy Warren to change this a little bit. Incredible
work by ONC, CMS, HHS, NCVHS and others in putting this together.
Chuck Friedman, Karen Trudel and Denise Buenning have done a yeoman’s
effort to do this. I would also like to make sure, if you will put your hands
up, all the NCVHS support staff that is in the room that has made a difference
in this, thank you so very much. You have done an incredible and masterful job
to pull this effort together. Those of us who just flew in last night made a
few phone calls and didn’t quite see everything that happened, but it was
magical, mystical, and it is here. So I’d like to thank everybody for that.
Over two days we have 40 presenters. You need to please stay on time. We do
have lights up on the podium as you are presenting. When that light hits
yellow, you have two minutes. When that light hits red, you are done. The
reason is, what is very important to us is that we hear from everyone. The
other thing is that different from some of our other hearings, the committee
will be asking pertinent questions today, information and insight, a little
different approach than we have at times.
Also for those of you in the audience who understood that we did have some
open mike time, we will have a document available at the break for people to
sign up for the 15 minutes each day of open mike
More importantly what I would like to say is that we have three ways of
getting information for this hearing. The first is, obviously we will have some
speakers sitting up at the podium. The second is, we will have some people
stand up at the microphone for the period of time that we have available. But
more importantly, anyone can submit written testimony of one to two pages to
us. We consider all of those equal. We have to have some dialogue to build a
body of work, but all of that input is equally important. So if you sign up, I
would recommend that you please prepare written comments and send them in along
with everyone else within 24 hours of the end of the hearing, so that if you
don’t get to the microphone, you are not left out. So that is available. Please
take advantage of it. Don’t consider if you don’t get to the mike that we
didn’t hear you. We will in fact hear you.
For each panel as we start, I ask you to start on time, and more
importantly respect those that are after you. If we start getting behind or
people start spilling over, then we are going to have a problem. I am actually
going to authorize the next panel to go ahead and stand behind the panel that
is starting to run behind, to remind them if necessary that they are not
fulfilling their requirements. We need all of the information, not just most of
the panels. If we let too many delay, we will do that.
When we are finished with this, we will prepare a meeting and written
comments summary over the next two-plus weeks, send it on to ONC and then post
it on our website, so we do plan to do that.
Without further input from me, I’d like to welcome Dr. David Blumenthal as
the National Coordinator for Health Information Technology. To many in this
room, his work is well known. More importantly, to those of us that didn’t know
him personally, his work ethic has been exciting during the preparation of this
hearing. It makes us even more excited with what he is going to be doing.
So Dr. Blumenthal, welcome. We stand ready to assist you and HHS as you
with others lead the health IT vision of the country. So with that, Dr.
DR. BLUMENTHAL: Thank you, Mr. Chairman, and thank you to your committee
and to the staff for the incredibly hard work that has been done to put this
meeting together on such short notice. As many of you know, I have not been
here long, this is my second week in my role as National Coordinator, but I
have been overwhelmingly impressed with the quality and the dedication of the
staff that have been in place before I got here, and that continue to serve the
public in so many ways.
I want to start with a brief note about my own personal involvement in this
issue, which although a somewhat untraditional role, but maybe suggestive of
where policy and focus is heading in the future.
Unlike so many of the dedicated people who have labored on the topic of
health information technology and its use, I am not an expert on the technology
itself, I am not an informatician. I got interested in health information
technology as a primary care physician about ten years ago when I started using
a record, because I was frankly forced to, to be a member of my health system.
I realized that it was changing my work and changing it for the better.
Medicine was going to change, health care was going to change as electronic
health records became more widespread.
I then put on an academic hat and began studying the implications and
barriers and facilitators to the adoption of health information technology. So
I know it as a user, and I know it from a policy perspective. I think those two
perspectives are perhaps highly relevant to where we are today, focused on the
use of health information technology and focused on the policies that will
promote its effective use.
Over the next few months, the Administration with the help of people like
those who are here today, expert testimony from many stakeholders, the many
interested and supportive members of the public, we will be guided by the high
tech provisions of the American Reinvestment and Recovery Act, and we will
address the challenges of promoting the effective use of health information
technology in accordance with current policy. There is an enormous amount of
work to be done. It is very challenging. This meeting will be extremely
important in informing us in that work.
Today’s topic could not be more relevant, the topic of meaningful use. As
you said, Mr. Reynolds, one couldn’t anticipate the amount of weight that these
two words will carry just a few short months ago. The law, the high tech
provisions of ARRA, promise to reward meaningful use of health information
technology. Our job is to define that term in a way that makes it meaningful
and in a way that will advance the health and welfare of Americans.
The definition will importantly inform everything that the Office of the
National Coordinator and many other offices in the Department of Health and
Human Services do with respect to health information technology. It will inform
how we think about certification. It will inform how we think about the funding
of supports for meaningful use. It will inform how we think about the promotion
of training to support the dissemination and use of health information
This is a complex job, but in some ways the overall goals are quite simple.
By focusing on use rather than on the technologies themselves and on meaningful
use at that, the Congress has clearly communicated that information technology
and all of the valuable work that goes into developing that technology is
ultimately important not for its own sake, but for its ability to improve the
health system and promote the health of Americans.
So in defining meaningful use, we must focus on its impact on the daily
lives of Americans, and the impact of information technology on the daily lives
of Americans and their caretakers, both public and private, in their personal
health system and in the public health system, its actual impact on patients,
on professionals, on institutions and on the families of Americans, as well of
course as on the efficiency with which we provide health care services.
So that is a tall order. It is easier to say than to deliver on. But as I
said, this is the beginning of a very important process. Our ears are open. We
are grateful to the many experts who have made themselves available to help us
with this process of learning.
Thank you again, Mr. Chairman.
MR. REYNOLDS: If the first panel please come on up and take their seats,
I’d appreciate it.
While they are coming up, I’d like to make another comment. If there are
any media inquiries that come forth, if they would please contact Nick Pappas
at HHS, area code (202)690-6343, and he and others will coordinate any requests
and use member of the committee or government officials as needed to do that.
So that would be the process that I would ask that everyone be willing to
Our first panel today is going to be covering the vision of health and
health care transformed. We had them in an order, but due to our early status,
which is good, we are shooting for that on a regular basis in the next two
days. I’ll just briefly mention their names, if each of you would tell us a
little bit about you before you present.
We are going to have Dr. Carolyn Clancy, Helen Darling, Dr. Elliott Fisher
and Dr. Stephen Schoenbaum. If it is not a problem to anyone, Ms. Darling, if
you would go ahead and start first, since you three were gracious enough to be
here. Then we will fit Carolyn in as soon as she joins us.
MS. DARLING: Thank you very much. My name is Helen Darling. I am the
President of the National Business Group on Health, which is a private
nonprofit membership organization of mainly very large employers, including 64
of the Fortune 100. Our members cover about 55 million active employees, their
dependents and retirees.
Shall I go ahead and make my comments now? All right, I just wanted to make
sure I wasn’t just introducing myself.
We had originally thought that I would be after these two men, and I was
going to skip a lot that they would say, because it is so important, I’m sure
they would be much more eloquent. So imagine the context. You have already
heard Elliott Fisher and Steve Schoenbaum.
MR. REYNOLDS: Well, if you believe that would be a better way, they are
sitting right over there with you.
DR. DARLING: No, that’s fine. But I think it is important, because it is
the work of people like Elliott Fisher and Dr. Schoenbaum that have actually
set the stage for much of what we are doing today, and appropriately so.
HIT should be used to insure that we as a nation start getting what we are
paying for, by effectively identifying over use and waste in the current health
care system, increasing the use of evidence based care to improve health care
quality, improving patient safety, implementing pay for performance and
informing consumers of the effectiveness, quality and cost of health care.
Anyone who saw the paper this morning about virtual colonoscopy would have
seen an excellent example of what we need to know and why we need to know it. I
think they said that 40 percent of the people are choosing a virtual
colonoscopy over a traditional one, and have lots of problems with that, as I’m
sure you know.
We have to do all of these things to reduce what the President called
appropriately, wisely and accurately, the crushing burden of health care costs
on families, businesses, employers and the economy. He has said this
repeatedly. That is actually one of the most important things in our judgment
about the national health care reform movement now, is the recognition that if
we do not control health care costs pretty dramatically, quickly, the entire
country is going to be swamped by those costs. We won’t get better care again,
as Dr. Fisher has appropriately identified many times; we will just a lot more
care, some of which could be harmful, all of which will be expensive, and
certainly not all of it will be appropriate to what we need.
We believe that it is critical that we target the unprecedented level of
federal funds in the economic recovery package to physicians who demonstrate
that they will use health information technology to improve the effectiveness
and efficiency of care. HIT has the potential if implementing correctly to
improve providing evidence based care and fundamentally change the nation’s
health care delivery system to save lives, reduce suffering and save billions
We believe that the five areas of performance measures endorsed by the
National Quality Forum, of which I am on the board, to measure the effective
use of HIT include electronic prescribing, interoperability, information
exchange, care management, quality registries and the medical home are
appropriate structural measures, and we urge you to consider process and
outcomes measures where possible, for example, the rate of care according to
guidelines, and reductions in preventable medical errors.
We need to transition from this frankly absurd, outdated, paper based
health care system that is horribly inefficient. We spend twice or more than
any other even advanced rich country in the world on health care, and we
clearly do not get what we should be getting for that. We want a continuous
learning, dynamic 21st century health care system which we do not have, that
will reduce medication errors, duplication of diagnostic tests, help physicians
manage complex cases, particularly those of the chronically ill, improve
coordination of care and reduce preventable readmissions to hospitals. I know
you all have seen the data recently on that.
We need to use the information game to improve providers’ adherence to
evidence based standards of care. Everyone knows well the Beth McGlynn work,
RAND Corporation work on this. Outcomes in health care quality to reduce the
rising costs of health care, a continuous learning health care system, and for
more details you should see the IOM’s Roundtable on Evidence Based Medicine on
that topic, will improve patient care and insure that even after drugs or
treatment are approved for use in the U.S. health care market, that we will
continue to determine if these treatments are reliable and effective, so that
instead of more care, we provide the right care for the right person at the
This modern health care system will avoid unnecessary risk and outdated or
even harmful treatments in the case of new clinical information, similar to the
situation that arose over a popular medication for arthritis a couple of years
ago. As I’m sure you all know, it took 12 years to find out that a medication
that everyone believed was on balance quite positive, quite appropriate, turned
out not to be for some people, and it took 12 years, but it also took the
Kaiser Permanente medical records system to find that out.
If you look at all the research in this country, almost all of the good
research about what is going on in the practice of medicine is coming out of
integrated delivery systems that have advanced HIT, and they have physicians
and nurses and others who know how to use them.
The three barriers to modernizing the nation’s health care system are the
lack of HIT adoption among providers — you have the authors of the recent
study here fortunately — are highly fragmented health care system, and the
lack of an adequate workforce that can use this technology. Clearly we need to
have higher adoption rates to get the is life saving technology into the hands
of our nation’s providers. In addition, we know that health care data are
stored in paper form in silos. Our physicians spend 20 to 30 percent of their
very valuable time wasted searching for information —
MR. REYNOLDS: Could you make sure the mike is turned on, please?
DR. DARLING: Thank you. Insuring the meaningful use of HIT requires more
training on the part of the current workforce to use this technology and
restructuring our nation’s payment and medical education systems to increase
the amount of primary care physicians, physicians assistants, nurses and nurse
practitioners so that supply meets the demand to transition to our nation’s
health care system. We need to restructure our payment system, but I realize
that is not the topic of today, but we need to do both of those things in order
to make sure that we can have meaningful use.
The financial crisis, the meltdown, and we are not even talking about swine
flu yet. The financial meltdown that this country has gone through has shown us
more dramatically and in many ways tragically than ever that we as a nation
have to totally transform the U.S. health care system. We have to if you will
use the overuse and the waste that is in there to finance the care that we need
to have for the 44, 47, whatever the number is, million Americans who do not
have any kind of insurance.
We have a fiscal crisis, not just a health crisis. The combination of all
of those things together mean that we absolutely have to act. My final word
would be that we have to act urgently. Urgency, fiscal prudence, quality health
care system, a safe health care system, a learning health care system, those
are all the things that we have to have.
Thank you very much.
MR. REYNOLDS: Thank you. Would either of you have a preference as to who
goes next? We will go then with Dr. Fisher.
DR. FISHER: I am Elliott Fisher. I am professor of medicine at Dartmouth
Medical School, former internist, or an internist who stopped practicing about
five years ago, at the VA where we had an electronic health system. It is
really a great honor to be here, and I am happy to join you.
Variations in per capital health care spending are pretty well recognized,
but less well known is that the growth in spending has also varied dramatically
across the United States. The differences may appear small, but compounding
makes a different. If all of the U.S. regions grew at the rate of San Francisco
over the last 15 years for the next 15 years, the Medicare program would save
about $1.4 trillion under current growth rates.
What explains the higher spending rates? Almost all of the differences that
we need in our work at Dartmouth points to differences in what we call supply
sensitive services, the kinds of things that are shown on this slide. Medicare
beneficiaries in Miami are hospitalized much more frequently for conditions
that would be treated outside the hospital. They see physicians more
frequently, they receive many, many more imaging services are referred to
specialists more often, have a smaller proportion of their visits to primary
care, and have many more physicians involved in their care.
And more care isn’t always better for patients. We have got extensive
research that documents that. I won’t spend much time on it.
What is going on, and what needs to be done? That is what should guide your
thinking as you are thinking about what meaningful use is. Our recent insights
are that most clinical decisions require judgment. Only a small minority can be
specified through firm clinical guidelines. Otherwise, why would you bother to
send them to medical school? Gray area decisions, discretionary decisions about
when to refer to a specialist or when to order a test, are the primary driver
of differences in spending across regions. We think it is because in part we
have a payment system that rewards growth, overuse, fragmentation, and insures
that whatever capacity the society generally provides to the health care
system, it will be fully utilized. Physicians adapt their prices to new and
existing capacity. Income pressures and the price cutting that most health
plans are involved in motivates the purpose of new technology, recruitment of
more specialists, high margin treatments, referral and admission of more
complicated patients, and we believe that the poor quality that we see in the
higher spending regions is a direct consequence of the fragmentation we see
across health systems.
If we are going to achieve a reformed health care system, we have to
address what we believe are the underlying causes of poor performance. I will
focus primarily on the principles. We have lack of support now for improvement,
care management and care coordination. We need to develop virtual or real
integrated systems that can support clinicians in their practice. We fail to
recognize the role of the local delivery system — how many hospitals, how many
primary care physicians, how many specialists we have as a determinant of both
quality and cost. We need organizational accountability for total costs and for
the capacity of the local delivery system.
We have a public that believes that more is better in the face of evidence
to the contrary, and that tends to equate less care with rationing. If that is
the case, comparative efficacy research is extremely important, as are
comprehensive performance measures of the quality of the systems within which
patients are receiving their care.
Finally, have a payment system that rewards more care, increased capacity,
and increasingly entrepreneurial behavior. If that is the case, we want to
foster accountability for overall spending and the quality and costs of care
through either comprehensive care management fees or global shared savings
Let me spend a minute on each of the three ideas, organizational
accountability, measurement and payment reform, because all three depend upon
an effective HIT infrastructure to achieve their greatest success.
We have been promoting the notion of an accountable care organization.
These would be local, integrated, real or virtually integrated delivery systems
that would bring together physicians and their local hospital or hospitals, and
that could be accountable for the full spectrum of patients’ care, from the
onset of illness through a full episode of care, and secondly, they are large
enough to be able to reliably measure their performance on important attributes
There are numerous potential accountable care organizations, physician
hospital, organizations, independent practice associations, integrated delivery
systems or regional collaboratives such as are present now in many parts of
North Carolina and Indianapolis.
A key insight from our recent work is that these are feasible to establish
in almost every market throughout the United States. Most physicians practice
within natural referral networks of the specialists and primary care physicians
around a given, or perhaps two, hospitals. If you group those together
empirically as we have done, what you find is that for the Medicare population,
70 percent of their care is provided by that group of physicians. That is
sufficient to support the kinds of changes in payment and changes in
performance measure that we need to achieve.
Why is organizational support important? We will hear more from others
about organized delivery systems and what they have been able to achieve, but
here is one example in terms of the implementation of evidence based practice
at Intermountain Health Care. They have an electronic health record that
allowed them to define the protocols which would be used to treat diabetic
patients, and updated them monthly in terms of advances in their professional
knowledge. These protocols help guide physicians’ practice where there was good
strong evidence about how to treat a particular patient. Over the seven year
period shown on this slide, they achieved marked improvements in the proportion
of patients who had well controlled blood sugars, as is shown on that slide.
Equally important, to show an example of probably — I don’t know if Dr.
Blumenthal is one of the physicians displayed on this screen, but I use these
slides from Tom Lee, who used them at the Federal Trade Commission. EHRs with
data warehouses could also be used to feed back information to clinicians for
those gray area decisions that are the major driver of waste and overuse in the
current delivery system.
What this slide shows, for those of you who can’t read the small print —
and we will talk about confidentiality later; you may have a question about how
these came to me with all of your names behind them on a spreadsheet, I
carefully got rid of those — what this shows is rates of advanced imaging
across primary care physicians adjusting for case mix within the Partners
Health Care System. We see fivefold differences in ordering of advanced imaging
services. The only way to address this kind of variation where judgment is
required is to feed that information back to the clinicians and start to have
the conversation about why they are practicing differently and what is the
right way to make decisions in such uncertainty areas.
Better information. I talked about performance measurement. Better
information is critical both for comparative effectiveness research and for
performance measurement of health systems. Evaluation of treatment outcomes and
provider performance, we now well recognize, requires a longitudinal framework.
This will require registries with clinical and patient reported risks and
baseline health status, the ability to capture the specific treatments that are
provided for patient follow-up, so that we can determine what their health
outcomes are, and then equally important, an analytical and organizational
capacity that allows feedback and improvement within the organization and
comparisons across organizations.
There is an essential infrastructure that you will be struggling with,
which are shown on this slide.
Payment reform is the third critical element that HIT can support. Our
current payment system has two primary effects. It fosters unprofessional
behavior in some physicians and some health systems, but most importantly it
presents a barrier to aligning care with better value for patients and
Payment reform should support high value care, whether through episode
based payments, which have the potential to improve care and lower costs, or
only with the kinds of measures that are going to be available through HIT. You
must be able to examine the outcome of those treatments. You must be able to
determine whether there are more episodes being delivered, and put them in the
context of a population based accountability for overall costs.
Similarly, global shared savings as we promoted or prospective global
payments are much more effectively done in the context of population-based
accountability for overall quality and the overall costs of care. Our
perspective is that successful implementation will require comprehensive
outcome and cost tracking.
Let me wrap up. I have a minute to spare, so I am not going to quite wrap
up. I think if our charge as we think about meaningful use is to establish a
vision of what we want to see in the U.S. health care system, I would nominate
three ideas: integration, integrated delivery systems widely accepted as the
way we need to be at some point, accountability for overall costs and quality,
and a shift to value based payments. If that is the vision of where we need to
go, then our charge is to figure out what are the policy steps we take over the
next five to ten years that get us there. HIT has a tremendous opportunity,
whether it is through requiring advancing standards to support the key
functions, making subsidies contingent upon the degree of local network
participation so that you can have all the patients get all of their care being
seen by the same providers, performance measure or payment reform. All of those
different policy pieces can come together to support improved care, and I
believe improved efficiency of the U.S. health care system.
Thank you very much.
MR. REYNOLDS: Thank you. Dr. Schoenbaum, please.
DR. SCHOENBAUM: Thank you very much. I appreciate the invitation to speak
with you this morning. I am Steve Schoenbaum, a physician who is Executive Vice
President for Programs at the Commonwealth Fund, and also Executive Director of
our Commission on a High Performance Health System.
This Commission has stated that the objective of health care and a health
care system is to lead to longer, healthier and more productive lives for the
population and the individuals in it. For health care to make this
contribution, everyone must have access to it, and the care must be of
excellent quality, efficient and equitable.
The Commission has recommended five key strategies to achieve these
outcomes. They are shown in the exhibit on the screen, which I will not read.
In addition, my colleagues, Karen Davis, the President the Fund, Cathy Shayne
and I back in 2000 set out a 2020 vision for a patient centered health system
that the committee’s first question to us as a panel about the critical
characteristics and enablers of a safe, patient centered, high quality
oversampling that optimizes patient outcomes. That vision included the elements
that are shown on this next exhibit. One of those, you will note, is having
clinical information systems that support high quality care, practice based and
What is the current reality? In a recent survey, three out of four adults
in this country reported having one or more access problems, such as getting an
appointment with a doctor the same or next day when sick without going to the
emergency room, getting advice from their doctor by phone during regular office
hours, or getting care on nights, weekends of holidays without going to the ER.
In addition, almost half reported one or more problems with coordination of
care, such as failure to provide important information about their medical
history or test results to other doctors or nurses they think should have it,
not having test results or medical records available at the time of a scheduled
appointment, or their physicians not receiving a report back from a specialist
they had seen.
Interestingly, 90 percent or more of U.S. adults report that it is
important or very important to them to have one place or doctor responsible for
primary care and coordinating care, having a place to go on nights and weekends
besides the emergency room, having access to their own medical records, and
having all their doctors have access to their medical records when needed.
Three things are central to both our personal vision or the Commission’s
vision of achieving excellent health outcomes in health care for all Americans.
They include having a health system that emphasizes primary care and that
enables each person who wants a primary care clinician to have one, having
incentives that encourage coordination of care, and having robust programs and
systems including appropriate exchange of health information, that supports
better coordination of care and a variety of other services such as after hour
The Commission on a high performance health system recommends, quote, that
patient and provider incentives should be aligned to encourage use of effective
evidence based health services, avoid use of unproven or ineffective care,
avoid misuse of services, and avoid over utility, duplication and waste.
In a recent set of papers, my colleagues and I have described how
incentives, particularly payment incentives, interact with the way that care is
organized and the degree to which caregivers become accountable for care. We
have also argued that new innovative Medicare payment policies, including
progressively sophisticated ways of paying for bundles of care or episodes of
care and paying for performance, could lead the way in transforming the U.S.
With respect to information technology, the Commission has stated, quote,
sufficient funding and leadership should be committed to achieve universal
implementation of interoperable electronic information systems within five
years. The Commission recognizes that in practice, health information
technology used meaningfully can be instrumental in providing evidence of
effectiveness, just in time through decision support, avoiding duplication and
waste, with legible and accessible documentation of patient specific
information, and achieving better coordination of care by transfer of
appropriate information among providers.
My personal history in health care as a physician, manager and patient has
led me to believe that meaningful use of information technology can best be
realized through implementing systems that have basic functionalities that
facilitate the care of patients by physicians, nurses and other providers.
These systems need not necessarily be extremely complicated.
In 1981, I began working as a physician and manger at Harvard Community
Health Plan in Boston, a staff model HMO that had begun to use a so-called
automatic medical record in 1972. In the era prior to the widespread use of
work stations, when inputters entered information from encounter forms, and the
output was a dot matrix printout, it was nonetheless possible to have a medical
record with availability in multiple sites and to improve quality of care with
appropriately deployed reminders and prompts.
Indeed, it was possible beginning in the 1980s for a patient to get care
from a primary care clinician in one Harvard Community Health Plan location,
eventually there were 14, specialty care in another, urgent care at night and
on weekends in yet another, and emergency care if needed in an affiliated
Boston area hospital. The entire record would be available to facilitate the
patients care in each location and document information needed by the next
provider. In addition, beginning in the mid 1980s, we developed a system of
reminders and prompts to improve the provision of necessary age specific
preventive care for all patients, and also specific care for diabetic patients.
We also developed registries of patients with chronic conditions, and provided
the information in them to clinicians to enable appropriate follow-up. By
implementing a tracking system for positive test results, it was possible to
insure that follow-up of abnormal test results such as positive cervical
cytologies or Pap smears, were virtually 100 percent. By programming our
laboratory result reporting system, we could provide clinicians with
information about the appropriate next step for following up different types of
abnormal Pap smears.
Furthermore, although the original COSAR system has now been replaced by an
EPIC system, it has been possible to preserve and enhance the database. Today,
if I got to the Harvard Vanguard Medical Associates website, choose the My
Chart function and provide suitable login information, I have access to
portions of my medical record, including for example my immunization history,
over the past 25 years. I can also request appointments and prescription
Many years ago, the medical leadership of Harvard Community Health Plan
began to establish an annual set of clinical goals, objectives for improvement
of clinical care and outcomes that we wanted to achieve. These goals helped
drive the meaningful use of available information systems, indeed, helped
harness the power of these systems for achieving better outcomes efficiently.
The vision of the Commission, the Commonwealth Fund and my own personal
vision for achieving better health outcomes for the population of this country
require having a national leadership that sets performance goals and
facilitates the deployment of systems and incentives that will achieve the
goal. Failure to have this sort of national leadership has been I believe a
central barrier to having a more effective and efficient health system in this
The Commission on a High Performance Health System understands that a
largely private health system is an important feature of the U.S. health
system, a feature that has been associated with an enviable track record of
information, only one of which I have been giving you an example of. It
understands that not only do we have the most expensive health system of any
developed country, but also that performance across our country is
significantly poorer than that achieved in several others, and that overall
performance averages about two-thirds of what can be achieved, and varies
enormously by where one lives.
In virtually all over major developed countries, there is national
leadership and balancing of interests of patients with those of providers and
insurers. Despite our spending 17 percent of GDP on health care, and despite
government being the single largest payer in this country, national leadership
of the health system has been lacking. Starting with setting of performance
goals, assuring accountable leadership, we are much more likely to use our vast
resources meaningfully, effectively and efficiently, and I congratulate you for
being involved in that process.
MR. REYNOLDS: Thank you. Our next speaker is Carolyn Clancy. Carolyn, we
started a little early, so you weren’t late. We were a little more prompt than
we thought we would be. So welcome.
DR. CLANCY: Good morning, everyone. I must say, I am thrilled to be here. I
do believe this is the first time I have had the opportunity to testify before
one of my colleagues, Mike Fitzmaurice. So let me just say, if you disagree
with anything I say, he had nothing to do with this.
My written testimony, including two recently published articles with
several colleagues, has been submitted to you.
What I want to talk about is, do we have the right stuff, and by the right
stuff, I mean the capacity in health care delivery to make sure that we rapidly
translate scientific advances to the benefit of patient care. I think it is
well known frankly that our track record to date would suggest that we don’t.
The question is, how does meaningful use in the context of other scientific
infrastructure supports for health care delivery make that contribution.
It is interesting that the Wright Brothers’ plane first took off the ground
right around the same era, a few years before the Flexner Report in American
medicine, when there was a huge amount of interest in making sure that there
was much more of a scientific basis for the delivery of medical care.
They didn’t have systems for safety back then, but as of World War II,
obviously aviation and medicine pursued somewhat different paths. But even by
World War II, despite the advent and use of Doppler technology and so forth,
bombs were not incredibly accurate. And of course today, aviation is now used
as a model for how health care could work. I think the crash in the Hudson
River and the really dramatic and miraculous save of every single passenger is
a particular example. I am always struck by the fact that airlines strive to
continue to do better things.
So where are we in medicine and health care then? I think we passed the
Wright Brothers stage, although I think if we were honest, we would acknowledge
that we crash more often than we need to. We certainly have not reached a
systems based approach where the right thing to do is the easy thing to do. Too
often, for clinicians the right thing to do requires heroic effort, and we
clearly need information infrastructure and support to achieve this goal.
If you could imagine in less than five years, what an encounter with a
patient with chronic illness could look like, supported by both a personal
health record and an electronic health record, and all the transactions that we
do on paper today are done automatically and seamlessly and there is constant
checking. Very importantly, that electronic backbone is linked to a social
networking function where clinicians and others can share best practices. They
know that they are getting the latest evidence that is kept up to date, and all
of this is supported by payments. I think that is the vision you have heard
from two of my preceding colleagues. I had to miss Helen’s, but I am guessing
she would agree as well.
The question is, how do we make progress? If you buy this vision of
transforming health care, then what we are talking about are a number of
components. I think I may mention one or two that are a little bit different or
different areas of emphasis, no surprise in any of these on this slide, I don’t
think, for any of you.
Clearly health IT can enable quality improvement, but to quote someone
testifying in this very hotel a number of years ago around patient safety, he
said to err is human, but to really screw up, you have to have a computer. So I
think we know that if we automate what we are doing now, we are not going to
get to the kind of health care that we want. Clearly that means that
implementation and meaningful use has to take very careful consideration of
patient experience and work flow.
Comparative effectiveness is something that AHRQ is amazingly, incredibly
excited and sometimes exhausted about, but I do think that that is going to
give us the power to tailor information in the same way that Amazon.com does
today, when you log on or they send you emails to say you might be interested
in something. We know technologically how to deliver that kind of tailored
information; what we need is the infrastructure to actually have the
information to deliver.
I think this has big implications for those who are providing care at the
front lines. The guiding quote I always think about is Don Browick, who once
observed years ago that only those who provide care can improve that care. It
is fine for all of us and exiting to imagine what all of this information flow
could mean, how it would change what is done today, but it will also require
some serious thinking about the confidence and skills of those who are
providing care now, who have been trained to see patients one at a time, do the
very best you can, and move on. There is a gap there between the kinds of
skills that they will need to be practicing medicine in an information age. I
would say this goes right down to all front line health workers.
Quality measurement is a big part of this. Usually I can amuse people who
spend their lives thinking about health IT by sharing with them the little
tidbit that when physicians and other organizations develop guidelines and
measures, they never ever once think about a data source to assess their
efforts. Over time of course, we are going to have to increase the
interdependence between those developing measures and those developing
electronic health records and other applications.
I know that you have already heard enough from Elliott and Steve that
payment reform will be a huge enabler here. Increasingly, we would like to
imagine that these measures are linked to episodes, that is to say, they are
not linked to silos of care, because we have made some progress there, but I
think we can do a lot better.
What we haven’t talked about explicitly is, what about the 160,000 or so
primary care physicians — and I know you will hear from John Tooker about this
later today — who work in very small practices? One question I have is, how
might they be helped to be part of ongoing, what I am going to call quality
improvement networks? This is a concept, not an obvious existing prototype. If
the question is, is there some way this electronic infrastructure and the
concept of meaningful use can be linked to a common electronic infrastructure
that changes how we think about what is scalable for a small mom and pop
practice, instead of having every single very small unit replicating the same
functions of finding test results, finding patients and so forth. Clearly if
that were the case, then we could imagine that these practices could become
part of a learning network, and also a mechanism for spreading effective best
Safety. I think everyone understands that we have to move beyond just the
oops approach, where when something bad happens we have a review, and then we
educate everyone in sight or who was anywhere within miles of the event like
crazy. I think that health IT is going to be a much bigger part of our
anticipating hazards and harms in patient care.
My former boss and colleague and friend John Eisenberg used to say that
chlorophyll was not the only green catalyst. Clearly we are going to need
measures and metrics that allow us to pay for high value, high quality care.
Where we are now with pay for performance today, I think we have probably maxed
out in terms of our current capacity. So there is an interesting chicken and
egg phenomenon, where we need health IT and an infrastructure to make it
possible to do more.
I want to close by commenting on a transformation framework. The question
is, this is clearly a huge part of the economy, how do we build a road map to
This very busy slide talks about translational blocks and getting from
science to routine systematic application and practice. My colleagues and I
believe that it is not just a matter of two translational blocks, which has
previously been described, but also it is a matter of aligning testing
effective practices or efficacious practices to those policy interventions and
business interventions such as measurement and accountability, health care
system design and scaling and spread to make this systematic.
But very importantly, that same infrastructure could offer important
feedback signals to the very earliest stages of the biomedical enterprise, so
that they understand when their interventions don’t work as understood. That is
a kind of link that we don’t have in health care today.
Obviously this is not just about moving electrons. We are going to need a
transformation in culture. I won’t belabor this except to say that we have seen
in some of the work that we have funded, including the Keystone Project in
Michigan, its success begets a lot of culture change, and people get incredibly
excited by seeing dramatic achievements that they had not thought were
possible. I left Dr. House off this slide deliberatively.
Clearly at a time when it is very easy to figure out where you are in any
city, we owe it to Americans to figure out how to have a GPS and navigational
tool for health care. I think it is fair to say that we are the number one
health care system in the world in terms of achievability or breakthroughs
sometimes, but where we need to get to is to be number one in terms of
reliability, getting it right every time.
So the bottom line here is that we have to engage our success in meaningful
use in terms of improved outcomes, quality and value. I think that emphasizing
and incentivizing data sharing as an essential step to care coordination is
going to be a very critical piece here.
We are going to need to be very strategic in terms of tailoring investments
to a stage of adoption of a particular community or set of practices. I will
also say that if there are two areas where we should focus, it would be
improving chronic illness care, because that is where quality measurement and
public reporting has brought us only modest success. It is going to require
connectivity between clinical care and community based resources and in
reducing disparities. We know all about those gaps. The high tech field makes
it possible to do that, and frankly it is going to make it possible for us to
more systematically collect that.
I will just close by saying that meaningful use is a destination and a
journey. Thank you.
MR. REYNOLDS: Thank you to everyone. A very prompt panel, thank you. One
thing good is when you start off with a vision, it is nice to have a little
extra time to chat about it. We are ahead.
First I would like to pose to the panel itself, if there are any things
that one of the other members said that you would like to reinforce. I know
that was mentioned by some of the earlier speakers. If you have anything just
briefly that you would want to reinforce that somebody else said, it would be
helpful. Then I would like to move on to questions from our committee.
DR. DARLING: I would just like to bring to your attention some work that
was done recently by something called the National Priorities Partners, 28
organizations that got together and spent over a year trying to identify
I mention it only because it is fairly specific, what they came up with. A
group headed by Dr. Bernie Rosoff, many of you may know him, identified those
areas of overuse. All of us have talked about that in one way or another. They
include inappropriate medication use, unnecessary laboratory tests, unwarranted
maternity care interventions, unwarranted diagnostic procedures, inappropriate
non-palliative services at the end of life, unwarranted procedures and
consultations, preventable emergency department visits and hospitalization, and
potentially harmful preventive services with no benefit.
It was a very concrete list worked out by a huge number of extremely
knowledgeable people, building on the work of especially the three people here
and some of the work they cited. But if we could have some of those things
resolved pretty dramatically, the opportunity for not just improving care and
quality of life, but saving immense amounts of money is there.
DR. FISHER: I would love to build on what Carolyn was talking about. I
think her notion of the quality improvement networks is a concept that deserves
serious thought, about the definition that we use. That is, to invest in
electronic health records in practices that are not connected to the same
electronic health record seems to me a risky investment, whereas putting in a
requirement that says within some number of years you will be part of a quality
improvement network or a payment model which would be an accountable care
organization, are quite similar concept in the more distributed virtual network
I think it really deserves serious thought as a foundational element of
HIT. I was imagining, what is the question that a patient must be asked; all
the physicians in the last three months had access to my electronic health
records. There is some measure that the patients could report on their
experience. The quality improvement network idea would build toward that kind
of shared evaluation record.
MR. REYNOLDS: No other comments? John Houston.
MR. HOUSTON: Thank you very much. I know Dr. Fisher spoke of a value based
payment system, and Carolyn Clancy also spoke of payments based on high value,
high quality care. I guess the problem that I see often is that the contract
between a member and a payer is often only one year in time.
The question I have is, how do you derive value where high quality care
often involves long term investment by payers? Again, a payer can put a lot of
money into a patient, yet a patient has the freedom often to simply shift
payers each year. So how do we get past that hurdle?
DR. FISHER: Thoughtfully. I think if you look at the opportunities, most
patients don’t change plans every year, but many do, especially as patients get
older they tend to move less.
I think the challenges are, with better performance measures, and we are
hoping to work with Dr. Clancy and AHRQ in trying to implement systems that let
you understand patient risks once a year, a good health risk appraisal, you
could measure changes in health risks even in as short a period as a one year
time frame, that would start to create some accountability at the system level
for helping me stop smoking, lose weight, exercise, as well as get my
cholesterol, blood pressure and other factors under control.
So I think it is a challenge. I don’t think it is impossible. The second
challenge which the health care system faces particularly is how do we align
our efforts within the delivery system with community based efforts to include
the health of a resident population. Again, that is the second course of
leverage, trying to get a set of performance measures which will come from the
systems that you require us all to have, that let us understand community
health and use all of the providers’ information in that community to
understand which neighborhoods have ill health, where they are not walking, and
then you start to get the kinds of synergies that I would hope you would see
from an information system that informed us about public and clinical health
opportunities for improvement.
DR. SCHOENBAUM: It is a very interesting question. I think there are
multiple ways of trying to address it. Assuming that we are going to continue
to have a multi-payer system, which I take as a given, then one has to ask, how
does one get some conformity; is it among them.
It is a question that we faced in the HMO business many years ago, because
we were committed to doing things that maybe had payback decades later, but we
felt they were part of our mission. So having these common goals that leads to
the provision of those elements of care that you as the payor may not
immediately benefit from is still a sensible thing, because it fits in with the
broader set of goals that you have, and perhaps performance measures that you
are being held to.
Another is to think about changing the way that we align the incentives,
and who is doing the paying for those particular elements. Think for a minute
about the fact that one of the expenses that is shouldered by our Medicare
program is end stage renal disease. One of the expenses that is not shouldered
by our Medicare program is prevention of end stage renal disease, and going out
and trying to find people who are diabetic and might become at risk for end
stage renal disease, making sure that they are treated appropriately.
It may not be easy, but one could conceive how one could change the payment
system so that in fact, the payment was aligned with the party that was going
to benefit in the long run.
MR. REYNOLDS: Were you going to make a comment?
DR. DARLING: Yes. First, even when plans change frequently, people don’t
change their positions. But the key is the performance of those positions,
among other things.
Second, if we have the kind of data systems that we are talking about, the
amount of misuse and overuse could go away. So for example, one of the things
that we see happen in the private sector, if someone changes a plan and that
means they change their physician, then basically that physician today has to
start over and do everything wastefully. There are tests that don’t need to be
done every year. There are some that have three to five years. There is almost
no way an average citizen can get the information, so when they do change
physicians by choice, they can’t say here is what has been done.
Once we have the ability for someone, whether it is in their electronic
personal health record which is portable, or it is through systems that allow
the physicians and other parts of the system to share data, then a tremendous
amount of problems and waste go away. I think we could then hold the physicians
and the patients in this instance accountable for what they are doing or not
When somebody has the ability to say, for the most part there are providers
who are doing a much better job, we can’t answer those questions now. If we
could, you would have a different flow, it seems to me, of individual choices
for their patients and for their physicians.
DR. CLANCY: Just to amplify on my colleagues a bit, people coming at this
question from the payment side would start to say, we are just going to have to
try some things. So we are going to have to design it and build it as we go. We
need some pilots or something like a reinvented idea of demonstrations.
What I can say right now is, there is a lot of debate ongoing right now
about whether pay for performance is just penny ante bean counting, or a step
on a path to get us somewhere better. Where I think it stops is at the current
capacity of most parts of the delivery system today to do anything much more
than make minor tweaks.
So what you are seeing in response, there are some insurers frankly seeking
out partners in the form of physician practices and hospitals to try to say,
look, we want to create a leading edge here. We will put a lot of money on the
table, and we get that you couldn’t even fantasize about getting there now. We
will work with you, but we want that commitment.
I know from some work with Helen that the Boeing Corporation is doing this
in Seattle. They are saying to doctor groups, these are our sickest patients.
We will work with you, but we need to actually change how this is all
So I think you will see a lot of that happening. The question is, in policy
terms, can we create a space for that, or does meaningful use have to be one
size fits all, and how do we link that in such a way that we get smarter and
better over time, or the glide path that I guess we will hear about later.
DR. W. SCANLON: You all mentioned the idea of payment reform as a key part
of improving the health care system. Three years ago, the committee had a
hearing on administrative data. It very much focused on the limitations of
administrative data versus clinical data.
So the question I would have is, in thinking about meaningful use, first of
all what priority do we need to assign to improving what we will call
administrative data, but really can be clinical data that is transmitted to
payors and analysts?
Secondly, the issue of, as we are doing this, and we are thinking about our
world in which there are very different payment structures, how do we define
the information that we need. If we are thinking about an episode as the unit
of payment, simply knowing that there is an episode is not going to be adequate
The last question is, is there any recommendations you can make in terms of
how we move efficiently forward in this effort. We inevitably are going to be
developing new measures, new ideas, and there is also going to be this issue of
retrofitting. We would like to avoid that to the greatest extent possible,
because it has major implications for all of those who are going to be
providing the data.
DR. FISHER: In one of my other roles which I did not mention, I serve on
the National Quality Forum committee that has been worrying about episode based
measurement and how we move forward, and to really understand value in health
It is pretty clear from those committee meetings and from our work that
this will require being able to identify a subset of the population that has a
particular clinical indication, whether it is eligible for knee surgery or a
diabetic patient. So that is essentially a registry characteristic. We will
need to be able to use registries to identify patients at the onset of an
episode, hospitalization for a heart attack, first visit for diabetes in a
given year, whatever, and be able to follow those patients over time.
I would be thinking about — and there are 20 episodes by NQF and the High
Value Health Care Collaborative that Mark McClone is running, supported by RWJ,
where there will be episodes that allow us to capture the costs of care for
those clinical episodes, ranging from diabetes to hip replacement.
The challenge though is that unless you can measure, identify all the
patients, and be able to adequately address the baseline risk characteristics,
and look at actual functional outcomes, we risk under any kind of value based
payment rewarding stinting on care, as we did in the 1990s.
So I would be working hard to think about putting in place the
infrastructure over the next three to give years in the number of projects with
IT vendors to say, we need to build registry capabilities in order to support
eventual payment reform.
The second parallel pathway would be working with the National Quality
Forum and AHRQ to say what are the measures that we would want to use as we
roll out these kinds of episode measures and episode performance measure
systems, what is the common set of measures we would then want built onto this
IT infrastructure that allows us to track patients over time and look at their
I think starting with two or three or 20 conditions, you would be a long
way down the road in three to five years in terms of being able to support
adequate performance measurement, and then the payment reform that we need. If
you had diabetic patients, heart disease patients, and a few of the acute
conditions, you would have a pretty good snapshot of how effective a delivery
system was to capture overall costs and get at the kinds of constraints that
were raised earlier about health. That would be my thinking about how to move
DR. CLANCY: I guess in the short run, we are probably likely to get more
mileage out of enhancing administrative data, because billing data are
ubiquitous, and we need certain things from billing data with respect to timing
of services and so forth.
What we know from work at AHRQ and also work going on at the National
Quality Forum to the health IT expert panel is, when you look at all available
quality measures and put any kind of prioritization on there, it is a
relatively short list of data types that you want records to be able to
Another recommendation that comes out of that health IT expert panel that
Paul Tang chairs is the notion of having a problem list, which frankly in terms
of identifying patients for various purposes would make life really, really
I haven’t quite gotten my own head around how hard it would be or not to
get clinicians to some consensus on that. As far as I can tell, every chart has
a problem list. The problem is, they don’t agree. So we will just leave it at
that. But I think that is worth exploring.
I think for awhile we are going to have a hybrid strategy where we build
out electronic health records to be able to retrieve data more easily for some
purposes, and to retrieve selected data elements that can be linked with
administrative data for efficient quality measure purposes.
How do you get there efficiently? That is hard. I think the overarching
challenge there is, you don’t want to foreclose any options. You don’t want to
be marching down a path where we will look back and say, oh, how did that
happen. So I think it might be an opportunity for some very serious folks to
build some different scenarios about how we get to the future, and guess on
timing, and guess on what that road map looks like, so that we could then step
back and say, there are a few things that we want to maintain the potential of
doing, even if we are not sure if that is two years or eight years.
DR. SCHOENBAUM: A couple of thoughts. My first one, to go along just the
lines of administrative data and the fact that there are certain things that
make it sensible to think about in different ways.
I was in Denmark a few months ago. There, they have partial capitation and
partial fee for service payments for primary care physicians. They pay for
units such as office visits, telephone encounters, and e-mail encounters, but
these are all tallied through the electronic records that the physicians use,
submitted periodically to the payor, who turns out to be government, and then
paid back. So they find that administratively simple.
We are just beginning to realize how much cost physician practices shoulder
as a result of dealing with all of the various payers and all of the various
administrative roles in this country.
So anything that would simplify that should have a significant effect, even
though it doesn’t have much to do with clinical care itself. I think that as
one gets into clinical care, clearly one wants to get increasingly into paying
not just for simple process, but for outcomes. That will be a lot easier to do
if that information is extractable from clinical information that is placed in
So I think one can look at the entire spectrum and say that it makes sense
to think about meaningful use in all of the pieces, or segments.
DR. DARLING: I would just add, when deciding about the priorities that
paying some attention to what information will allow us now to make better
decisions. So for example that might argue for registries or research that will
I say that because as you know we average about $2.4 trillion a year now,
so every year we delay we are layering on more costs. If there is a way that as
we are building the information systems that we need and the clinical systems,
if we can say, if right away we can make a decision about virtual colonoscopy
versus traditional colonoscopy, that is expense right now big time, and it is
making a difference. Anything that we can do to leverage the work of the
information system, the infrastructure, to help bend the curve, the better.
DR. STEINWACHS: Traditionally, medical records capture disease and disease
outcomes well, but when you talk about health outcomes, they aren’t there
usually. Either when you think about an episode, and you ask the question, the
patient didn’t come back, are they really doing better or not, but mainly in
chronic disease management.
So as part of the vision for the electronic health record, do you see a
partition or a part of this electronic health record in which patients could or
might have incentives for entering health information as different from — many
times we talk about the PHR, the personal health record, as a place where these
things happen, but I don’t see those coming together unless it is actually part
of that medical record.
DR. CLANCY: I guess what I would envision is that you could see an
electronic health record in a clinician’s office, and then when clinicians and
patients get together or when they are doing some sort of e-health extremes,
they periodically sink.
This gets into some interesting issues in terms of, if a patient sees their
record and decides something is wrong, can they correct it. I gather that
legally, the answer to that question is no. But you can imagine the kinds of
But I do think one great opportunity is going to be for us to eventually be
able to know over time how we are doing, particularly with chronic illness
care, because right now we are guessing at what are the important snapshots.
For people with multiple chronic illnesses we are not even doing a very good
job at all.
DR. FISHER: We at Dartmouth are slowly rolling out exactly that notion,
where patients will complete surveys prior to seeing their clinician, which
helps the clinician tremendously, but includes data about health risks, health
status, uses the John Wares SF-12s so you can see if someone is depressed, more
substantive depression screens.
I think there are tremendous opportunities to take patient reported
information, make it useful to clinicians so they would want to have it in
front of them when they see their patients, so it is helpful, not a hindrance.
Otherwise it will never be adopted. But that same infrastructure can be used to
report on outcomes six months later and have that roll in appropriately to the
medical record. There are issues of handoffs and how does secure data get
handled, but I don’t see us being able to do any of the things we are talking
about with the vision of episode based payment without knowing whether the
patient was able to walk three months after their knee replacement.
DR. CLANCY: Could I also just add, in terms of short term practical steps?
I know that my colleagues at CMS will be thinking, this sounds like a decade
out and so forth. We are right now testing what we are calling a health IT
supplement to the Cap survey for clinicians and groups, which might be one way
of bringing in patients’ experiences. I’m sure there are dimensions that we
will discuss today that won’t necessarily be part of it.
Elliott, I don’t know if it was you or Steve who mentioned this before, but
the point is, that might be something that could be moved in a relatively short
period of time, and might be a way to get started on this journey if patients
are entering their own information and so forth.
I think when we talk to chronic disease organizations, the National Health
Council and so forth, patients actually want that functionality in their
personal record. They want to be able to link to researchers and others at
their choosing, but nonetheless that is not something that is built in now.
DR. CARR: I would like to go back to something that Carolyn said, where you
mentioned that success changes culture. I would like to reflect that right now,
we seem to have two populations. We have a large group of 60-something
physicians, many of whom don’t have e-mail, and certainly don’t have EHRs. Then
we have all of our kids’ generation that do everything electronically.
So as we look at our immediate near term goals, how do we decide where to
put the greatest resources? To bring the 60-somethings, no e-mail docs to what
Harvard Pilgrim ultimately became is one set of resources. I think where you
are one system it goes quickly, but if you’re scattered small practices around,
how much can we do to get them up to speed, versus building a system that in
five years, all the people will be of a generation that is more electronically
DR. CLANCY: I guess the clearest information I have comes from
Intermountain and Brent James, where they have a pretty much train the trainer
approach to, this is how you make it a part of your practice. He says that
works very well for them.
I do take your issue with the generational divide. My own nieces and
nephews embarrass me on many occasions. They were shocked that I wasn’t on
Facebook, for example. But more to the point, the AMA has done surveys showing
that docs use computers and many other parts of life, like 90 percent and up,
to book travel and so on and so forth. So I think there has got to be some
movement on making these products easier to use, as well as figuring out how do
you teach more senior dogs new tricks?
DR. SCHOENBAUM: One of the reasons I emphasize basic functionality is that
I think ideally — I just hope as an old dog that you don’t give up on old
guys. In a sense, if the incentive is right, and by this, I don’t think it has
to be financial, I think it has to be that it is usable, and I can see what the
reward is for using it, then some of us actually still are capable of learning
some new tricks.
There is so much advantage to having everyone have some basic functions
rather than having a handful of people have very sophisticated ones, that if I
had to choose, and I’m not sure one has to always choose as starkly as that,
then I would opt for the basic ones, and trying to design them in such a way
that they were attractive to the users.
DR. DARLING: I would just add also, Carolyn’s comment about the evidence,
the evidence is the old dogs are using them for managing their portfolios and
things like that. So they are amazingly sophisticated when it comes to
something on the personal side.
But if there is a challenge, it seems to me that there ought to be money
within the existing funds to create something like small SWAT teams that come
out of organizations like the American College of Physicians, American Academy
of Family Physicians, those who have had a lot of experience with this, and
send them out to help the physicians who are slower to change or resistant
because maybe they are out there alone, and they are already tired, the last
thing they need is something else that is going to screw up their practice for
six months, and things like that, but literally send — this is a job
opportunity, too — send people out in carefully created SWAT teams, and then
use the personal stories from the successes of those stories, and there will be
many, about how much of a difference this makes in their lives, and publicize
those as widely as possible.
That would not cost nearly as much as we will spend on lots of other things
in the next year, and it could really transform the way the physician community
thinks about this, because they are not seeing it as something that is
distracting them, wasting their time by their standard, and even making them
angry and want to withdraw from the systems we are setting up. Make it a
positive experience and tell a positive story.
MR. REYNOLDS: Being that the Chair is also over 60, Justine Carr will not
get any more questions. Next is Walter Suarez.
DR. SUAREZ: Thank you. This has been a terrific panel to get us started, so
thank you for your thoughtful comments and points.
I think there is health information technology and then there is health
information technology. There is EHRs and then there are EHRs. In other words,
there is a myriad of functionality and applicability and capabilities and
usability of different types of technologies and EHRs.
I think one of the most exciting new areas in the Recovery Act is the whole
field and the funding that has been provided to support the whole area of
comparative effectiveness. My question to the panel is, to what extent do you
see there is an opportunity to apply the same principles of comparative
effectiveness that traditionally go into clinical care and finding more
effective ways of clinical care, into evaluating health information technology,
and looking at comparative effectiveness of health IT and comparative
effectiveness of EHRs, and the functionality and the capability. Is there
something there to be done?
DR. CLANCY: Yes. Let me just say, in terms of the investments in the
Recovery Act for comparative effectiveness, we are still working through
setting priorities in particular for the $400 million that will be allocated at
the discretion of the Secretary of Health and Human Services. So I don’t want
to go too far in terms of saying where we are, because we are not there yet,
and because we have been very deliberative in terms of trying to create a
transparent collaborative process.
We had the first of three listening sessions with stakeholders and so
forth. What I can tell you at AHRQ is that one of our centers for education and
research on therapeutics considers health information technology as
therapeutics, and that is led by David Bacon Nelson. So that might be a nice
model for it.
Thinking about it as related to comparative effectiveness, it becomes more
like the challenge of evaluating devices, where you are constantly seeing
upgrades rather than medications or other interventions which are more suitable
when necessary for clinical trials. So I think that it will require many of the
same methods as well. So I do think there will be cross fertilization no matter
how you think about it.
DR. TANG: Thanks to the panel for sharing your vision on the transformed
health care system. I appreciate your acknowledgement of the role of the
patient and their use of the data in these electronic recordkeeping systems.
I have a direct question though as it applies to meaningful use. Do you
think that patients’ access to and use of information contained in the
electronic health records to be a part of the definition and qualification for
getting the incentive?
DR. SCHOENBAUM: I disclosed some of my roles, I didn’t disclose all of
them. I am a member of the board of the Picker Institute, which promotes
patient centered care. I think the answer is unequivocally yes.
About seven or eight years ago, we funded — we not being Picker, but the
Commonwealth Fund — a project with the University of Colorado Health Sciences
Center, which was one of the first looking at patient access to electronic
medical records. Despite the fears of physicians that patients were going to
get confused, that they weren’t going to like it, that the physicians were then
going to be overwhelmed with all sorts of unnecessary requests, essentially
none of that happened. Patients did like having access to their records.
I think that is now becoming a universal experience. There are issues that
have to be dealt with, but they are dealable with. By and large, when one has
an informed patient, one has better health care. So for me, it is a slam dunk
DR. DARLING: I would just add, thinking about your question, right now at
least in the private sector and particularly for employers and their employees,
there is a modest amount of development going on for personal health records of
When you look at what is going on, we might need a set of principles. Under
no circumstances for example would anyone, an employer, a government or
anybody, spend money on something that can’t be portable. It sounds very basic,
but a lot of the health records that were developed in the last ten years, they
are fine if you stay with that company or that plan. But you can’t even take
your own records with you, even though under the law you are supposed to be
allowed to, but you just literally can’t.
So I would hope that a group of knowledgeable people on this subject would
develop a set of guiding principles that would say under no circumstances
should anyone put money, government money or private money, if they get the
guiding principles they will know what not to do, that don’t meet these
criteria, don’t fall within these principles. So we can at least stop spending
money and entrepreneur’s spending money on things that frankly aren’t as useful
as they should be. Then at some point they are going to say, now we have made
this big investment, we have to somehow not have this strong stand on
something. So we should do that as soon as possible.
DR. CLANCY: The one point I would add, because Steve said it very well, is,
I think that meaningful use also needs to include sharing of patients’
information. We are very, very effective now at having people sign HIPAA forms
and all that kind of stuff. It could become a part of routine practice that
practices are routinely asked, with whom do you want this information shared;
is that your manila folder, is it your primary care medical home, is it another
specialist and so forth.
It is not exactly what you asked, but I think it becomes a way to engage
patients in this. Obviously some patients wouldn’t necessarily — having the
option is important, but actually making sure the information is used for care
coordination is very important as well.
DR. FISHER: My silence should only be taken as agreement.
DR. BLUMENTHAL: I want to thank all the panel members for very, very
stimulating and helpful remarks. There have been a lot of very useful
Obviously as we go forward thinking about meaningful use, we are going to
be thinking about phasing things. We are also going to be thinking about
feasibility of meaningful use and the burden of meaningful use.
If you will forgive me to put you on the spot, and you can come back and
change your views at a later time, but if each of you would say, if there was
one thing, one capability, one performance dimension that you would put at the
top of your list for meaningful use, out of the box, what would it be?
DR. CLANCY: I would say the registry function keyed to specific
applications that you would find out from fulltime practicing docs.
When I think about docs who cross the divide, they have interesting
stories, like what really made it for our practice was when Vioxx was taken off
the market, and for the first time we didn’t have to telephone scripts in case
people call in. We could proactively reach out to everyone to say, please come
in, we are changing your prescription or whatever.
I don’t think you would know that unless you asked people. Academically and
intellectually, I can tell you why registries are really important in a lot of
areas for many applications. But I think figuring out some specific
applications like that that are seamless would be very helpful.
DR. DARLING: I would agree with that. Also, I want to get back to the point
about leveraging the information skill. We have a crisis. We need to move
quickly on some things. So overuse, the ability to be able to identify and
track and put to bed eventually, to be able to answer the questions of, is this
effective, is it working. I think Carolyn’s idea about the registry, that is
one way to find out some of these things.
We happen as you know to have this new possibility of having evidence
development with coverage. So we are going to soon have a number of treatments
or procedures covered under this coverage with evidence development. The aim is
to answer questions quickly about what works and what doesn’t, so we can stop
doing what doesn’t work and energetically go after those things that do work.
DR. FISHER: I was going to say registries, too, absolutely. So I can add
I think the capacity to link and track patients over time, whether it is
linkage to administrative data which will make the analysis of cost and overuse
— I don’t believe that we are ever going to be able to specifically identify
most of the overuse, frankly. I think it is going to come from benchmarking. It
is going to come from the kind of slide I showed you from the outset that says
that the patients in Salem or San Francisco are somehow getting half as much
care as those in Miami and East Long Island. Those comparisons will illuminate
the opportunities for improved efficiency.
So I would add to the registry the capability that it can be linked to
administrative data for patient follow-up and evaluation of health system
DR. SCHOENBAUM: First, I object to having to pick only one. I am going to
be slightly deviant in my choice. I would choose, if stuck with only one,
decision support. I think providing information to people that will lead to
better care is extremely important, and that there are data to support that it
leads to better outcomes. We now have it in the hospital with Amarassingham
study, and we have it from things like CPOE. So I think that is the one, if
stuck with one, that I would stick with.
MR. REYNOLDS: To play off that question, if you are sitting there at the
same time in 2014, what looks good to you, as to how we look? When you are
sitting up there at the same time in the same room, what looks good to us?
If you were sitting at the same desk in 2014, what does it look like that
would make you feel that we have made a big change?
DR. DARLING: I’ll start. First, that we would have an infinitely safer
health system. That we would have a rate of cost growth that is sustainable, so
tied to other economic indicators in a way that made it sustainable. That we
had more health, not just more health care, and that we produced all of this
much more efficiently.
All that seems sort of obviously, but we don’t have any of those things
going on right now, and we have been working at it a long time. So I think
those are the goals I would see.
DR. CLANCY: I guess I would say that we have seen a dent in chronic illness
care. I share all of Helen’s aspirations, but I must say, producing an annual
report on quality and disparities has made me acutely aware of just how stable
the system and performance is. We see improvements every year, but they are
So what I would love to see is a much sharper increase in that curve that
you can track back to at least one component of meaningful use. I don’t see how
you make a dent in chronic illness care without being able to target those
people with particular conditions who are having the most trouble.
DR. FISHER: I will take Helen’s better health and better care and lower
cost. It is the easiest way I can remember what we are aiming for, and it is
Don Berwick’s triple aim, essentially. That is where we are going.
Then I think it is an information system that really supports our knowing
that we are going there, so that we are doing it measurably, and that patients
are getting better care measurably, health is improving measurably.
I think that is the piece that this architecture can provide, is the
measurable piece of lower cost, better care, better health.
DR. SCHOENBAUM: I would say that our Commission has been putting out a
scorecard. It has something like 37 indicators. I would settle by 2014 for some
kind of significant, substantial improvement in any five of them. I don’t care
My guess is that improving any five would require that we now have some
national leadership in the health system, that we have done something
significant in the organization incentives and the infrastructure such as the
information technology to support it.
DR. CLANCY: I am going to add one minimalist one, too. I talk to fewer
doctors, hospitals, other health professionals who complain about sending data
to a black hole. In other words, docs and others are getting more feedback on
how they are doing in a much more timely way. We know that is critical to
making those dramatic improvements in quality.
So even if we got there in 2014, I think that would be a giant leap forward
from where we are now.
MR. REYNOLDS: Leslie, you had a question. I am going to hold the questions
and the answers to five minutes, and we will break promptly at 10:45.
DR. FRANCIS: Dr. Clancy touched on this question in her response to Paul
Tang’s question. Although it is clear from the data that most patients do want
interoperability, there are also questions about whether patients trust
interoperability and whether patients want sensitive information in particular
I wonder if any of the panelists would like to comment on the role of
privacy protections in meaningful use.
DR. DARLING: I’ll start with that, because this is a big issue among
employers, because we obviously want to make certain that our employees
understand independence, that no one can have access to their information.
But I would say first that it is extremely important, we know that the only
way what we are talking about is going to work is if everyone has confidence
that their information is protected.
But the second requirement that I think is really important is that we give
people the chance to opt out of anything they want to opt out of, basically.
They can be assured, if they don’t want their information shared with anybody
or anything under any circumstances, they have that choice. That has to be
enforced by groups or people that are trusted, the whole concept of trust and
If you look at the public opinion data on who is trusted and who isn’t,
there are not a lot of people who are trusted, frankly, but those who are, it
is mainly physicians, by the way, so if it is physicians groups saying that you
can trust this, we have done what we need to do to insure that trust. But in
all instances, every patient should have the option of opting out of any
ability for anyone to look at their data. I think that is what is going to be
required if you look at the public opinion polls.
DR. CLANCY: I would agree with that, and just add that most of our
experience and our mindset and frankly many of our laws about privacy all
revolve around a point to point set of transactional basis.
The real potential here is to change that a great deal, but ultimately
everyone needs to be part of that conversation. So I think that we are going to
have to have a lot of transparency about use, not just of personally
identifiable information, but even de-identified, and what that really means
and so forth.
My sense is that CMS is going to have to implement meaningful use before we
actually get to that playing field. But I think we have got to continue the
process that was started over the past few years with dialogues in multiple
states and so forth, for people to start to get their head around what does it
mean that data collected once can be used in multiple ways and so forth.
DR. SCHOENBAUM: First, I am delighted that there is now public discussion
of the privacy issue. I wrote an op-ed on the subject in 1996. In 1988, one of
the Presidential candidates was a member of Harvard Community Health Plan, and
we had to face this issue squarely, not that we had ignored it before then.
But in fact, I think that one ought to have opt-outs rather than opt-ins,
that one ought to have as up front guarantees of security as one thinks is
reasonable for the other meaningful uses of the data, and that then one has to
have very severe penalties for breaches.
We know that all systems can in some way be played with or hacked or
whatever, but if we are very serious about what those penalties are, then we
ought to be able to deal with the issue and deal with it effectively.
MR. REYNOLDS: With that, we all thank you very, very much. Outstanding job,
we really appreciate it. We will start promptly at 11. I would like the
speakers in their seats before that. We do have a signup sheet for public
comment. Thank you to this panel.
MR. REYNOLDS: Our next panel is on meaningful use capacity and
functionality in electronic health records. The thing that I would like to
remind each of the panelists again is that you do have the lights in front of
you. At eight minutes the yellow light goes on, and then you get two minutes
We had a luxury with the first panel in that we were early. Dr. Blumenthal
and I didn’t take as much time as we thought. We won’t have quite as much time
for interaction with each of you, but after you finish we will open it for
whatever amount of questioning we can have.
So with that, we have you in the order of Dr. Classen, Dr. McGeeney, Dr.
Middleton and Dr. Tooker. Is there any reason that anybody would want that
changed? Seeing all four of you there together, we will go by the agenda. That
would be Dr. Classen, if you would please introduce yourself and then make a
DR. CLASSEN: Thank you very much. It is a pleasure to be here to address
the panel on the issue of meaningful use in electronic health record systems.
My name is David Classen. I am a practicing infectious disease physician at
the University of Utah and associate professor of medicine at the University of
Utah, and the Chief Medical Officer at CSC. Today, what I am going to go
through is an overall model of meaningful use of electronic health record
systems very briefly, because I know others on the panel will go through it in
more detail. Then I am going to share with you results of a study we have been
doing over several years of commercial electronic health record systems after
implementation that might be helpful in informing the panel’s decisions on how
we look at meaningful use. Unfortunately there hasn’t been a whole lot of study
of commercial systems after implementation, so we think this might be helpful
as we go through this.
So if we can bring the slides up on the screen, just to start with what an
overall model of what an electronic health record system meaningful use might
mean. It starts potentially with the right product. That could be a qualified
or a certified product, then the right product implemented in the right fashion
with the essential functionality, and then adopted by the care providers in
such a way to enhance the quality and safety of care, and that hopefully leads
to the right outcome.
The right product might be a qualified or certified EHR product that meets
these criteria. Clearly they should be very, very relevant to the care process,
but also should address issues of interoperability, privacy and functionality.
But to be effective, that product has to be implemented with effective
capabilities. I have outlined here some things that others on the panel will
talk about, but clearly there are some differences between the effective
implementation of EHRs in the inpatient setting versus the ambulatory setting.
I have outlined some of those differences here.
Certainly there is overlap potentially for quality reporting from EHR and
potentially for health information exchange, but there are some things that may
be particular to each site. So as we think about meaningful use, it may vary
between the inpatient and the outpatient setting. Indeed, that is one of the
things we have learned from our ongoing study of commercial EHRs in actual use.
But also, it probably needs to involve the right adoption. Having a very
good system that very few clinicians used is probably not terribly effective in
improving the overall quality and safety of care, and this is a problem. There
are a number of organizations that have implemented these systems with a not
very high level of adoption.
So having these systems that effectively support the work flow and thought
flow, that enable role based use by clinicians, and then also achieve a level
of adoption on the inpatient and outpatient setting by clinicians, could be
part of any meaningful use definition, especially given the track record so far
that many of these systems are not used after implementation.
Then the final part of this model is the right outcome, that these systems
after implementation are advancing the delivery of high quality, safe and
effective care. This does need to be demonstrated. This cannot be assumed from
all the work that we have done, as you will see in a second, and it could be
done in several ways. It could be done through the billing process as an
additional code. It could be done through the existing accreditation process,
or it could be done using an existing quality reporting process. It could be
done by submission of quality reporting measurement and outcomes, or it could
be done through existing public reporting or pay for performance programs.
We have been conducting a study of EHR systems after implementation through
one of these public reporting programs that I will talk about in a second. We
have already over the last several years been collecting information about
electronic health record use in both the inpatient and the ambulatory arena
through an existing program.
That program is the Leapfrog program standard for computerized physician
order entry within electronic health record systems. This is a standard that
was initially developed by Leapfrog, but was harmonized with the National
Quality Forum standard in 2006. We have been running this program for awhile
and collecting this information.
This program initially started solely focused on the use of computerized
physician order entry in hospitals, with the standard as outlined, suggesting
that 75 percent of licensed care providers were entering medication orders into
a computer system. The other part of the standard was taking a test after
implementation of these systems to understand how effective that implemented
EHR system was. That is the part we will be focusing on.
This particular EHR evaluation tool was initially funded by the California
Health Care Foundation and the Robert Wood Johnson Foundation. Funding has now
been picked up by AHRQ. Many individuals have been involved in developing this
tool over the last seven years, and many health systems and hospital practices,
probably more than 150 of them, have helped develop it along with EHR vendors,
and we have used a number of homegrown sites to evaluate these systems
including the VA, who has been very helpful in the development of this.
The purpose of this was twofold. First, if you are a hospital or a clinic
and you have put in an EHR system, how effectively have you implemented it? How
do you know after implementation that it is actually improving safety? That was
one perspective. The other perspective is from the purchaser perspective, how
does the public know that they are going into a health care facility that has
effectively implemented an EHR.
This test is a self assessment test, web enabled. It is completely free. It
has many aspects that we have been studying EHRs under, both pediatric and
adult, inpatient and ambulatory. It has medication safety, quality and health
maintenance modules. I am only going to talk about the inpatient medication
safety version of it. It basically simulates the process of the effective use
of an EHR with as high fidelity a simulation as we could develop.
We fortunately had access to many research databases on medication safety
from a number of leading organizations around the country such as the Mass Tech
Collaborative on Preventable ADEs. It is one of the databases we were able to
use. We use these databases to make sure as we are evaluating EHRs after
implementation, we are really addressing the things that cause harm. So rather
than addressing any medication error, we were really interested in addressing
medication errors that cause actual harm in studies. We had access to a number
Those databases directed us to the key capabilities that EHR systems need
to have in order to avoid patient harm in the medication arena. Here are these
capabilities outlined. So for instance, can a system pick up a very serious
drug-drug interaction when medications are prescribed to a patient; can a
system pick up overdoses of medications in a newborn; can a system pick up cost
of care issues; can a system pick up corollary orders; when a toxic medication
is ordered, can the system recommend adding appropriate drug monitoring levels
to that particular area.
This web enabled self assessment works as outlined here. I won’t go into it
in much detail, but hospitals and clinics access a website and use it to study
the effective functionality of their system after implementation.
An example of how this might work. We created a simulated test patient.
That would be ordered into a clinic’s electronic health record system, and then
we would ask a physician to enter a fictitious order against that, such as
Coumadin, five milligrams three times a day. We would see if that system picked
up what is clearly an overdose of the medication. This is all after
This test gives immediate feedback to hospitals and clinics, to take it in
all those categories I just outlined. In addition it also gives feedback if you
look at the bottom around nuisance orders where we have over alerted
physicians, which is a major complaint. In addition it has a deception analysis
to pick up gaming of this self assessment test.
Here are the results of the first 74 hospitals in the United States that
took this test. You can see, we have both adult and pediatric organizations
that have been a part of this. These are slides you do not have, about five of
them. They are part of a manuscript currently under review, so I was not able
to distribute them.
As you can see, we have outlined the leading inpatient vendors in this
evaluation. We have broken the test results into a group of categories. Here
are those categories and capabilities. So we are looking at commercial EHRs
after implementation, their ability to pick up critical problems, allergy
problems. We did quite well on that in the hospitals that took it. We didn’t do
so well in drug-drug interactions, and we did less well in cost of care.
More dense capabilities. We did quite well in hospitals with their EHR
systems in terms of picking up the wrong route of administration. We did less
well in corollary orders. So I admit a patient with a heart attack, am I
prescribing aspirin as needed.
Then finally in the more advanced capabilities, the ability to properly
dose a drug for a patient’s age or weight or renal function, we did less well.
But an important observation here is that pediatric hospitals did much better
than adult hospitals, which suggests that meaningful use might vary not only by
site of care, but by type of care.
MR. REYNOLDS: We need you to wrap up if you would, please.
DR. CLASSEN: If we look at one more slide, which is the overall score on
this test based on the leading commercial EHR vendors after implementation, the
first column is for a homegrown well known EMR site. The other columns are for
the vendors, the leading EHR vendors and how they did in this test. You can see
clearly that the homegrown sites still did better than the commercial vendors.
You can also see enormous variability on the part of vendors.
The reason that is important is, many of these vendors were CCHIT
certified. It tells us that meaningful use is probably going to be determined
by implementation rather than the vendor products on the shelf.
What we have done in this test is make sure that we are in complete
agreement that what we test in this test is on the CCHIT road map. But I think
that is not enough. Maybe as we look at meaningful use, evaluating vendor
products on the shelf is important, but also evaluating after implementation
may be as important or more important.
So with that, I will stop.
MR. REYNOLDS: Dr. McGeeney. Thank you. Go ahead, sir.
DR. MC GEENEY: I welcome the opportunity to take a little different run at
some of the EHR functionality issues. I would like to approach it from what I
would call in the weeds or on the ground from a practicing primary care
TransforMED is a subsidiary of the American Academy of Family Physician.
Our mission is the transformation of health care delivery to achieve optimal
patient care, professional satisfaction and success of primary care practices.
The reason that is relevant is that we have a very strong emphasis on HIT, but
as a tool, not as a solution, what is ailing the U.S. health care system today.
I throw up the model that we preach and teach, as it were, and not so much
for the content of the model, other than to show that one of the major
components of patient centered medical home today is health information
technology. The other point I do want to make though is, under the multiple
bullet points under electronic health care information technology, electronic
medical records are only one of the bullet points. I think there are a lot of
ways that we need to be thinking about leveraging and scoring technology
functionality in the context of a primary care practice.
Some absolute requirements under EHR functionality, which was one of the
questions for today. We often think about safety. We often think about patient
centricity. We often think about quality. But I challenge us that we also need
to think about efficiency, that we are not going to be able to address what is
ailing the health care system unless we leverage technology to create
efficiency and measure efficiency in our system.
There are two things that I think are absolutely required. One is a self
populating practice based registry, which could address many of the other
issues that we think about with EHR functionality. The second one that is
absolutely essential is interoperability. Without interoperability we are not
going to be able to coordinate care. We all know that coordination of care is
something that is critical to the efficiency issues within our health care
There are some other meaningful user functionalities that I would like to
touch on. First, e-prescribing, but it is e-prescribing that interfaces with
electronic health record of the practice and automatically populates and
updates medication lists.
Secondly, we know this, but population list management. This subpopulation
from the electronic health record is critical.
Point of care reminders are critical. I will make the distinction here in a
moment between point of care reminders and evidence based decision support.
More interoperability as I mentioned is critical. Self populating
registries are an essential critical function of any electronic health record
technology going forward, as well as electronic order entry and order and test
The starred item at the bottom of this page is important. While evidence
based decision support gets a lot of discussion, I do think it is important
early on to make a distinction between point of care best practice reminders
and evidence based decision support. As we prioritize our needs, I think it is
extremely important that practices have the capability of a registry to search
a database and remind a provider of all of the important wellness preventive
measures and quality measures.
I don’t know that it is so important that if a diagnosis of a sore throat
is made, what the appropriate treatment is. Certainly somewhere down the road
it is, but I think when we are looking at early wins and critical early
implementation, we have found that reminding providers at the time of care of
what should be done to get patients in compliance with quality metrics, and
also to remind providers of what immunizations need updating and such, is a
critical factor in improving the chronic care of the patients.
We have further found that being able to leverage this technology to manage
populations of patients proactively, to be able to identify patients that are
not in compliance with any of our multitude of issues, and be able to reach out
to them, rather than simply wait until they show up in the office, is critical.
It is very, very difficult to do without technology, and technology can be
registered via registry functions to address many of these issues.
So when we talk about population based registries, they certainly could
have the ability to report quality metrics and capabilities. They have the
ability to track outcomes. They have the ability of point of care reminders, as
I mentioned. They certainly have the capability to have a searchable database
at the clinician level and at the time of care to leverage maximal care, and
they do allow that proactive management of populations of patients.
So as we look at critical things going forward as something that addresses
many of the needs of a small practice, man y of the needs of addressing what
ails health care systems today, and doesn’t necessarily have to be a large
Other issues that are important that you heard before. Certainly meaningful
use needs to go fast at a station. There needs to be validation and
verification of use.
It is an important point to keep in mind going forward that we need to keep
in mind current practices and a dissemination strategy that impacts current
users immediately. I think it is one thing to think about what we can do in the
future, 2011 or on, to change how technology products are introduced and their
capabilities, but in the world of family medicine, over 40 percent of family
physicians in this country have an electronic health record today. They may not
use it to its full capacity or appropriately, but they do have one. As we move
forward, we must think about how we introduce new capabilities in those
practices as the health care agenda is moving very quickly.
The point of meaningful use to me should not dictate or mandate necessarily
a comprehensive electronic health record. It may not be reasonable for a small
practice, a two physician or more practice. As we know, 60 to 70 percent of
primary care in this country is practiced in groups of six physicians or less.
It may not be that early on particularly, a comprehensive legacy system is the
answer, but maybe the focus should be on functionalities and capabilities that
are required by that practice to something quality and efficient care.
You could see a scenario where an a la carte solution of a registry,
e-prescribing and a templated ASP model EHR may create an immense functionality
for a practice and be quite affordable and help them to move their agenda along
very quickly. So as you think about issues of meaningful use and the issues
related to that, I would ask that you consider this as an option for
particularly the small group practices that we see today.
There are gaps that we are seeing in the system. Again, self populating
practice based registries are the exception, not the rule, in existing EMR
systems. There are add-on systems that are starting to be developed and show
up. Meaningful interoperability and solutions to actual practices as I
mentioned, and not just new products down the road. We need to very quickly
think about how to create interoperable solutions and not just be spending our
time on creating an agreement around a common language. Health care is being
provided today and patients are being put at risk today by the lack of
coordination of care.
We need to think about coordinating health information nationally, or at
least creating that interface nationally. We all know that. I think we need to
be careful not to have over reliance on legacy systems that are not meeting our
needs today, and thinking of other potential options, particularly with the
We need to continue to think about patient portals to facilitate the
portability and coordination of care. For us to spend a lot of time in this
space as you do, the empowerment of the patient I think is going to be critical
to solving some of our technology needs and leveraging the opportunities going
Then lastly, the focus on total practice based functionality, with the
center on the focus of quality and the patient, efficiency rather than systems,
that maybe we are trying to fit a round peg into a square hole, as it were.
With that I’ll stop. Thank you.
MR. REYNOLDS: Thank you. Dr. Middleton.
DR. MIDDLETON: Good morning. My name is Dr. Blackford Middleton. I am the
Corporate Director for Clinical Informatics Research and Development at
Partners Health Care, and chair of the Center for IT Leadership in Boston. I am
also a member of the National Committee on Vital and Health Statistics, but
today I am speaking in my capacity as someone who has been involved with
electronic medical records use, design, implementation and evaluation in
academic and industrial settings for over 20 years.
It is an honor and a privilege for me to be with you today to discuss this
critically important issue of meaningful use of health care information
technology, or HIT for U.S. Health Care, which I believe is essential as a
prelude to health care reform and transformation of our health care delivery
system into the information age at last.
Through no fault of our own, I think the physician emperor has no clothes.
We physicians have not invested figuratively or financially with HIT. There may
be good reasons for that, given that he who pays for HIT is not he who gains,
but we can overcome this barrier with stimulus and health care reform.
Briefly allow me to describe several issues which motivate and frame my
discussion of the meaningful use of HIT, which then will guide the discussion
of essential functionality and capacities required of HIT which should warrant
The motivation for HIT is clear. In our current fractured and unwired
health care system, evidence suggests that physicians routinely practice in a
state of incomplete information, with an incomplete knowledge base required to
effectively apply best evidence where it exists for clinical decision making.
With the knowledge in medicine continuing to explode, especially with the
revolution in personalized care, it is likely that the information needs
problems for providers and their patients will only escalate.
When considering assessment and measurement for meaningful use, I would
suggest that we consider how HIT itself may play a role in producing reliable
measures. HIT process measures could be a byproduct of the use of HIT itself,
for example, the per provider percentage of prescriptions written with an
Measurement of meaningful use will require a strong signal, as David
Classen has highlighted already, something which can be measured reliably and
one that can discriminate between what it is judged as meaningful use or not,
and does so without bias, and cannot be used fraudulently to misrepresent use
of the technology. Importantly, as Dr. McGeeney has already identified, the
measures should be independent of the particular technology per se, and apply
whether a comprehensive thick client EMR or lightweight web based system is
While we consider our goals for transforming health care with the
application of HIT, I would suggest we must be wary of setting our sights too
low, or settling for only an intermediate goal state in the journey toward
robust HIT adoption.
At the Center for Information Technology Leadership, for example, we find
in our analyses that potentially many billions of dollars may be saved with the
broad application of HIT. Our findings suggest that the two most important
factors associated with the maximum value potential of HIT are seamless
interoperability of health care data and advanced clinical decision support at
the point of care. It is incumbent upon us to keep this long rang goal in mind,
even as we begin the journey with small steps and incremental targets.
Let us now turn to the goals of HIT adoption. Our first goal is improved
care delivery, lower cost, improved quality, but I hope also we design a system
that can enhance individual and community wellness and research on what works
and what does not. It is not about the health care technology itself, of
course, but rather what outcomes and redesign of our delivery system we can
achieve with it.
I suggest that defining clinical goals only is insufficient, and will not
lead to a future state of coordinated and seamless care across different
locations of care health care organizations and different technology solutions.
Such systematic goals require us to think about requirements of the system
itself, to work efficiently, things like the ability to correctly identify the
patient in any technology system, the seamless exchange and portability of data
and information across systems, the availability of the best knowledge at the
point of care, and clinical decision support within any technology, and more.
While these goals may not relate to one or another specific clinical goal, more
importantly they are relevant to helping us redesign our health care delivery
system, as HIT is the enabler and prelude to other dimensions of health care
reform to come. To achieve this vision, we will need a phased approach to
incremental technology adoption that incrementally and inexorably moves us
toward our combined goals.
Let me now turn to the questions at hand regarding EHR functionality.
Several excellent descriptions of EHR capacities and functionalities exist
currently. These include the 2003 Institute of Medicine letter report
describing the EHR, the HL-7 EHR standard functional specification, as well as
the certification criteria for ambulatory medical records 2008, and inpatient
criteria in development from the Certification Commission.
Time will not allow reviewing each of these now, but common to all of these
functional descriptions of EHR are the following core capabilities.
Data access. The technology must aggregate all the relevant patient care
data for review by the provider. This may include organizing the data in a
variety of presentations and numerically and semantically normalizing data from
disparate data sources.
Knowledge access. Provide access to knowledge base tools and services in
the context of the clinical work flow to readily provide an answer to the
clinical question at hand one click away. Work flow and clinical decision
support. Provide support for the clinical work flow, both for the individual
care provider and care teams. In addition, provide clinical decision support in
the context of the clinical work flow, such as alerts, reminders, documentation
assistance tools, care plans and guidelines and diagnostic and therapeutic
decision support as appropriate to both providers and patients.
Health care information exchange. Facilitate information exchange between
disparate sources for the purposes of data review for each and every clinical
encounter. This should include all essential data elements such as current
medications, allergies, problems and past medical history, laboratory and
diagnostic test results, and patient demographics, for example.
Let us now turn to the specific technologies that should be in use by 2011.
The language in high tech within ARRA describes three broad areas. I concur
with these capabilities being areas of focus for 2011. In each case, in the
written testimony I provide pointers to illustrative capabilities as defined in
CCHIT criteria. These lists however are not meant to be comprehensive.
Number one, e-prescribing, defined as electronically prescribing and
managing a medication list with patient specific medication decision support
such as drug draw, drug lab, drug allergy, medication history, allergies,
including inpatient medication reconciliation and medication administration in
the inpatient care setting.
Health care information exchange, defined as the ability to aggregate,
collate and display in appropriate views all relevant patient care data from
ancillary systems, for example, laboratories, radiology reports, patient
demographic, administrative data, et cetera, and in addition provide the
ability to produce a visit summary and exchange select data across and between
health care organizations.
Lastly for 2011, quality data reporting, defined as the ability to submit
standard quality data sets to the appropriate authoritative agencies reflecting
care quality, patient safety and compliance with select standard quality
measures as we have heard about already.
Should a distinction be allowed between provider types? With respect to the
functionality described above, this observer would suggest that distinctions
not be allowed between provider types. The above functionality should be in
place and used by the relevant accountable providers for each and every
clinical encounter or inpatient hospitalization.
As the country moves towards universal adoption of electronic health
records, the depth of functionality and interoperability to fully achieve the
value potential of health care information technology must match the breadth of
By 2015, in addition to the above, the following functionalities should
also be widely adopted to warrant incentive payments. Demonstrate the ability
to widely submit, without tampering or fraud predefined measures of HIT use and
quality data reports. This is critical to allow meaningful use to be assessed
as a byproduct of use. Enhance structured and coded documentation such as
problem list, medication list, advanced directives, blood type, adverse events,
number of events, family history, medical history, surgical and procedural
history, social history, chief complaint, health care maintenance and screening
items, for example. This is to improve the data we collect and input to
clinical decision support and quality data reporting systems, and a progressive
standardization of documentation templates and controlled medical terminology.
Effective work flow and clinical decision support, the team based coordination
of health care maintenance and screening reminders, preventive care services,
chronic care management, acute care management, triage, remote consultation and
Keeping my eye on the clock, I am going to skip some of the other written
testimony and just talk about the highlights.
Knowledge management services. Given the depth and breadth of clinical
content and knowledge artifacts required to support effective work load and
clinical decision support, health care information technology solutions should
have access to a national knowledge repository and knowledge management
services to facilitate keeping all systems up to date with current best
evidence and best practices. Seamless and secure information exchange and
interoperability, and I would note also the need for personal health records as
an adjunct to the EMR strategy.
There are three important issues that I just wanted to touch on briefly in
closing that I think we need to think about as we move toward our national EMR
adoption strategy. First is the national patient identifier. We have had a long
debate in this country about the merits of a national patient identifier. This
observer would suggest that the benefits far outweigh the risks and that we
should move toward developing and implementing this identifier with all
appropriate policies and protections in place to insure its safe and effective
Standardized clinical data model and clinical terminology. While we have
made dramatic progress in recent years, thanks largely to the efforts of
leading standards development organizations such as HL-7, integrating the
health enterprise and the health care information technology standards panel’s
efforts on harmonization, we do not yet have a universally accepted and
complete clinical data model, nor do we have standardize clinical terminology
or even a subset of clinical data elements. Every effort should be made to
advance our progress in these areas, as they undergird the successful design,
implementation and use of all health care information technologies.
Lastly, a standardized user interface. We now recognize that the
human-computer interface of health care information technology itself may
impact clinical work flow and outcomes. There is a growing literature on the
unintended and untoward consequences of HIT use. This observer would suggest
that a critical element for the successful design, implementation and use of
HIT is to standardize the user interface so that providers using any system can
use it appropriately. This will of course also decrease training costs and
assist providers and patients who provide and receive care in more than one
Thank you very much.
MR. REYNOLDS: Dr. Tooker.
DR. TOOKER: Good morning. I am John Tooker. I am the CEO of the American
College of Physicians. I am also on the board of the Certification Commission
for Health Information Technology, and chair the National Health Collaborative,
the AHIC successor. It is my pleasure on behalf of the American College to
provide this testimony to the National Committee on Vital and Health
Health information technology offers many opportunities to improve health
care delivery, but only if coupled with major changes in other areas such as
care delivery, reimbursement and acceptance of new responsibilities by all
stakeholders, including clinicians, payors, employers and patients. The process
to define meaningful use presents a rare opportunity for us to encourage
fundamental changes to our health care system, and focus on what important
changes are attainable through wide implementation of health information
We have a number of specific suggestions, and some concerns regarding EHR
certification and the definition of meaningful use. These points fall into five
broad categories: time limitation, certification requirements, functionality
for practice, a measure of meaningful use and the financing of the health care
system. I will touch briefly on these. There is a large explanation in the
testimony provided to you.
Certification criteria should be based on existing HITSP specifications and
CCHIT requirements. We support the use of a single standard for medical
summaries. All CCHIT certified vendors will support CCD in time to meet the
2010 requirements. The coordination of care document is the ASTM and HL-7
standard for medical summaries that has been recognized by the Secretary of
Health and Human Services.
The certification requirements for 2011 must be sufficiently achievable to
encourage widespread adoption of health information technology that
incorporates sufficient functionality now and lays the groundwork to assure
that more robust levels of meaningful use can be achieved over time. Otherwise,
high tech will fail to achieve its goals. We recognize that EHR certification
standards and requirements for demonstration of meaningful use can and should
increase over time. We recommend that NCVHS consider a multi year pathway to
achieve the functions not ready for adoption in 2011.
Clinical relevance, especially with a focus on patient centered care. I
would urge all of us to think about the development of meaningful use standards
from a patient’s point of view, must be the primary criterion for choosing to
implement and use EHR functions.
Certified HIT must be safe, secure, protective of patient privacy and
supportive of all relevant legal requirements for proper records management. In
addition, all of the HIT needs of a medical practice must be served by
certified technology. Also, an electronic health record system that meets all
the meaningful use requirements but fails to provide the fundamental features
that every practice needs, such as maintaining a patient problem list, or
linked to billing records, will slow adoption by physicians and other
clinicians in practice.
With regard to health information exchange, exchanges must insure that
common security functions are properly implemented. If data exchange is
involved, the exchange partners must be able to participate fully. If proposed
exchange partners are not able to manage their data exchange, practices may be
left without a feasible exchange of a partner.
Accordingly, we recommend starting a consensus building process regarding
the criteria that definition of meaningful use should include. They should be
linguistically clear, concise, evidence based where possible, valid and
reliable over time, least burdensome and disruptive measurement options
available, operationally defined. An operational definition identifies one or
more specific observable conditions or events, and then specifies how to
measure that event. Measurable with currently available measures. Visibly
linked to care quality and efficiency, including safety and having face
validity. Practical for small practices and hospitals, realizing the
extraordinary fragmentation of the American health care system, and protection
of patient privacy.
Quality measures should be used appropriately. We support the move to
electronic health record based reporting, but the data from electronic health
records satisfying the performance measures must be reliable. Reporting
requirements must be varied to fit the necessary differences among practice
types and medical specialties in care settings. Also, there should be multiple
pathways for reporting such as through intermediaries to account for variations
in practice. Keep in mind the differences and additional difficulties faced by
groups such as rural small practice and safety net providers.
We support electronic health activities that enhance patient-physician
coordination, and believe all of these data including test results, not just
orders, should be available to patients.
What are then the capacities and functionalities to enable the safe patient
centric quality health care system that optimizes patient outcomes? We believe
in the context of the issues noted above that key functions should include the
ability to generate condition or need based registries, the ability to
reconcile problem lists and medication lists, the ability to monitor common
nationally recognized quality indicators appropriate to the practice and
populations served with options to view data at both the practice and
individual clinician levels. The ability to generalize customized clinical
reports using simple query methods for specific indication, for example,
medication recall, identifying gaps in care, and electronic prescribing with
useful alerts and clinical decision support to provide guidance tailored to the
needs and preferences of the patient based on nationally accepted guidelines
and protocols, customizable at the clinician and practice level.
We are concerned that the proposed measures of meaningful use not exceed
the capabilities of the EHR product certified in 2009, CCHIT, for reasons
explained above. Therefore, with respect to the critical EHR functionalities
for which providers should be required to demonstrate use in order to earn an
incentive as a meaningful user of certified EHR technology in 2011, we
recommend the consensus building process previously described to define
meaningful use with the goal or initial metrics that leverage information and
data accumulated through the normal work flow and use of an electronic health
record. A list of possible targets should include reconciled problem list,
reconciled medication list, prescriptions, e-prescribed when appropriate and
permitted, allergy information updated at least annually, lab and imaging
results received electronically, and tests ordered online.
These functionalities should be broadly applicable in ambulatory care
across all specialties. Adequate metrics for hospital based uses of electronic
health records will need to include additional or different metrics based on
comparable levels of complexity and importance now.
We recommend that the road map for future years consider areas that begin
to leverage health information exchange such as test and referral tracking as
well as medication reconciliation and transitions of care across the health
care continuum. As noted previously, we strongly recommend that new
functionalities and measures of meaningful use be thoroughly validated before
they are required.
In closing, we support the objectives of high tech and the payment
incentives offered to stimulate adoption of health information technology. We
must take this opportunity to significantly stimulate HIT adoption and
implementation and place the United States on the logical evidence based where
possible pathway towards technology enhanced quality improvement.
MR. REYNOLDS: Thank you to all of you. Jeff Blair, you have the first
MR. BLAIR: Thank you, everyone, for very informative presentations. My
question is for Dr. Middleton. You wound up getting right to the heart of the
matter and pointing out with your study that in order to achieve the biggest
improvements, we need to improve interoperability, clinical decision support
and quality measures. Then you went down the next layer. You took us down to
the next layer, which essentially was the need for clinically specific
terminologies to support all of those three things.
With the aggressive time frame that we have, at least we have made a lot of
progress in those areas during these last several years, but during this next
year and a half or two years until 2011, 2012, clearly we have to continue to
strengthen those areas as quickly as we can. What suggestions do you have —
maybe I would expand this, if anyone else on the panel also wishes to answer
this question — what suggestions do you have during this next year and a half
or two years to be able to expand and enhance the clinically specific
terminologies that will support interoperability, clinical decision support and
DR. MIDDLETON: Thank you, Jeff. I’ll take a crack at the question, which is
large and comprehensive, and ask for help from my co-panelists.
I think the first point, just to underscore again, is that our objectives
for 2011 have to be real and doable and practical with existing technologies,
largely in the form they are currently in. So that to my mind implies a robust
definition of electronic prescribing and the feature set that goes with it to
achieve clinical goals, again underscoring this is in pursuit of clinical goals
But around electronic prescribing and information exchange, and soon in the
area of quality data reporting, we will have sufficient specification and
definition of the relevant data elements and the technology requirements to
make all of that first set of technology possible.
I do want to underscore the problem of developing a goal which becomes the
end in itself. I am very fearful that our sights will be set too low and that
we will achieve the 2011 objectives and never get beyond them. Our analyses
that I did allude to would suggest that the full value potential comes with the
more advanced and robust implementation of a comprehensive suite of
technologies and whatnot.
But let me at the same time modify that comment. To get to the full-blown
EMR, I think the strategy that one can elaborate and be successful with is to
consider a lightweight, a medium weight version and then a heavy version if you
will of this set of technologies as incremental steps towards the goal.
Frankly, it may be only by defining these incremental steps that we will ever
even achieve any of the goals, because to do it in all one fell swoop is just
Marc Overhage and I are working on a paper which now the pressure will be
on to actually finish, in which we describe the value that results from
lightweight, intermediate and full featured EMR implementations, based upon our
CITL models and evidence and experience from the Regenstrief Institute.
MR. REYNOLDS: Any other comments from the panel? DR. MC GEENEY: One brief
comment. I think to echo that, that was the reason I was pushing the registry
functions a little bit. They are not part of EMRs today. There is an awful lot
of work being done across the country by multiple entities that are solving
many of these problems that we are talking about. I think the light, medium,
heavyweight add-ons, those kinds of things may allow us a more rapid engagement
of the physicians that are out there seeing patients and doing the work and
having meaningful change in our system.
So I think that transition strategy, working around today’s realities, what
actually is occurring and happening, can be very valuable.
DR. TOOKER: I think it is a very good question. It depends what our goals
are. It sounds to me like we are trying to incentivize widespread, not just
adoption, but meaningful use of a health information technology electronic
health records. I think we have to acknowledge the challenge. Others including
Dr. Blumenthal have demonstrated the relatively low penetration of electronic
health records into the smaller practices in the community.
Just to put some perspective on this now, if you look at the 2006 national
ambulatory care survey, there were in the range of 900 million ambulatory
visits in this country, and the vast majority of those ambulatory visits were
in small practices. Each one of those visits is a unique clinical record.
There is another article that I think David Blumenthal was an author on
published in the Annals of Internal Medicine in 2005 about the costs of a
national health information network, and estimating that there were about
160,000 practices in this country of four or less. We aren’t talking just about
primary care where Dr. McGeeney and Live, but we are talking about across all
So to get meaningful use and accelerate adoption across the scale of all
the disciplines in medicine in these thousands and thousands of small practices
in a relatively short period of time I think is an enormous goal we are
pursuing, but I think we have to estimate what the challenge is before we start
thinking how quickly we can solve that problem.
DR. CARR: Thank you all, great presentations. Again back to Dr. Classen’s
findings from Leapfrog, we have our CCHIT criteria, and then we have
implementation. So my question is, as you looked at the display of results for
the certified EHRs, are there some criteria that do correlate with higher
performance, or is it all implementation?
Again, as we hear about the small, medium, large programs, are we
allocating our focus on implementation sufficiently to achieve safety and
DR. CLASSEN: We are currently studying the linkage between the criteria
that are in CCHIT and the results of the tests, so I don’t have an answer for
you. I have some anecdotal observations. Several of the criteria are on the
road map and are not currently part of the criteria within CCHIT. But we have
not been able to find a clear correlation between any of the criteria and the
performance on the test, which would suggest that implementation is so variable
and so different, that what you see on the shelf and what you get after
implementation are so radically different that there needs to be some other way
to assess these systems after their implementation.
It is not just after the initial implementation. We have been to many
organizations where as a byproduct of the ongoing upgrades of systems, critical
safety checks get turned off, and nobody knows about it until we arrive with
the test and uncover the problem.
So I think we have to look at these systems as very dynamic systems. Even
if you implement them correctly the first time, three weeks later the things
may change, and no one knows.
The answer to your question is, we are looking aggressively at those
criteria. As we collected more and more information and more and more hospitals
and clinics participate in this, we should have a better answer. But I would
say variability of implementation is a huge issue.
DR. WARREN: This is for Dr. McGeeney. You had talked about self populating
registries. I was wondering if you could give two specific examples and then
talk about how this works into the work flow of family practice documents, and
why you see this as an essential part of meaningful use.
DR. MC GEENEY: I see it as an essential part because it is a quick fix and
it addresses many of the issues. When we talk about self populating registries,
we talk about a tool that searches all available databases from the electronic
health record to the lab systems to the hospital lab systems. It populates the
In one example, the system would run a data run of a patient that is coming
into my office. You would take all of the quality metrics that are in
compliance in green, and all the ones that are not in compliance are in red,
including all the wellness protocols. The nursing staff would act on all of
these things before the physician or provider ever walks in the room.
So in theory, the patient is 100 percent compliant in all quality metrics
before the physician or provider ever walks in the room. You have accomplished
a lot towards chronic care, chronic disease management.
The same is true — you hear a lot today about proactive population care,
care managers, the Wagner model of chronic care. Part of that is leveraging all
the resources possible to reach out to patients. So no longer do we just want
to know which diabetics maybe didn’t get their blood test because a payor
informed us of that, but being able to search the database to know which of the
diabetics have an A1C greater than eight and call them up, and reach out to
them proactively. There is a long list of things.
But the point being, being able to have a tool that can pull all this data
together from multiple sources, remind the providers what needs to be done, and
allow you to manage a population of patients proactively. It would move our
implementation system dramatically forward very, very quickly.
That is the reason we need to think broadly. As John said, we need to focus
on the patients, or how we can meet those patients’ needs in health care. So
that is what the registry is all about.
DR. MIDDLETON: Just to follow up, it is a central question actually to
think about how we measure the quality measurement, how we marry the quality
measurement if you will with HIT. I think the notion of a self populating
registry is consistent with a progressive adoption of HIT.
But what we have to do is figure out how to correlate the development of
measures in for quality. Some of the work that is now being done by the
National Quality Forum and HITEP II which Paul is chairing and supported by HRQ
is aimed at this very task; how can HIT produce the quality measures without
fracturing the technology into a distinct registry. Many of the measures come
from work flow or physician based documentation, et cetera. So I would be
concerned about there being a splinter path toward registries which don’t
actually stay on track toward a robust EMR eventually.
DR. MC GEENEY: Just to echo that, it is a transition strategy. We need to
fix problems today and we need to address problems tomorrow. I would hope that
in a very short time new renditions of EMRs will have registries, but there are
One thing I didn’t mention is, it would also create a massive amount of
reporting functionality, various databases for pay for performance programs,
all sorts of things that could be leveraged now as the interim fix for that
DR. FITZMAURICE: Blackford, you mentioned in your testimony the need for
systematically normalizing data from disparate data sources. If we had national
standards for these data and their changes, could clearinghouses and billing
services perform the functions of health information exchange, and data
services for physicians maybe assisting information exchange among registries,
essentially being a part of the Nationwide Health Information Network? Is it
possible to build on what we have, or is the transformation so precise, so
accurate, that it could do the job?
DR. MIDDLETON: Excellent question, Mike, thank you. It is a very big issue.
The clearinghouse question or the information exchange question is not so much
about how broad we do it at first, but where we focus our attention at first.
Certainly we can standardize on a subset of data as opposed to trying to
boil the ocean and standardize the entire clinical information model and all
clinical terminology. We should in a value based way focus on the paying points
for the payers or the national priorities, as the case may be, and insure that
those subsets of data are adequately represented and being exchanged.
The mechanism of exchange I think you alluded to as well, whether or not
HIE’s, clearinghouses, regional hubs of interoperability can meet this need. I
think they absolutely can. It is not a far stretch from where we are today.
MR. REYNOLDS: Jorge, you had a question?
DR. FERRER: To follow on that, Blackford, I had a question regarding your
recommendation regarding the standardization of user interface. How do you
balance the need for standardized user interface with product differentiation
to allow for competition in the marketplace? From the marketplace, that is
where we differentiate our ability to do quicker, better, faster, better
management of information. Do you feel like today the marketplace is mature
enough to move in that direction? Could you expand a bit more on your
recommendation on the standardization of user interface?
DR. MIDDLETON: I’m not sure I heard the entire question, but let me try. I
think there is still the ability to differentiate in the marketplace when
common user interfaces do exist. Our automobiles, many other systems we use of
course have a more or less common layout of essential features that allow us to
operate them in a safe and effective manner.
I am not willing to trade off safety for market differentiation. I think we
have to design the best products that meet the work flow requirements to
optimize care. But it doesn’t mean necessarily impeding on basic dimensions of
the user interface, which can dramatically improve patient safety and quality.
MR. REYNOLDS: Paul Tang, last question. We have got three minutes for the
answer, so make your question quick.
DR. TANG: I think the problem is trying to use a $32 billion incentive
program to affect a two trillion dollar health care system. The tension is
where to set the bar.
The main problem is, we have a migration path that has high switching
costs. So although it is tempting to go incrementally, the switching costs are
such that you can’t just swap out things as you would a car.
So I would like to hear your opinion on where to set the bar, knowing that
we don’t have a seamless or a low cost switching — that we don’t have a
seamless migration path.
DR. MIDDLETON: Gosh, I’ll try, Paul. Big question. I think the trick now is
going to be at some level to determine the appropriate value proposition which
fundamentally makes the 2011 target the right first target, but then with
incremental value being attained within the system and its implementation to
make the sustainability of the model clear, that we move from the two billion
to transforming the entire system.
Two billion may in fact be the tail wagging the dog, for sure, even $34
billion, if we think about the global incentives availability. So I think that
is the analysis that is required, is to figure out what are the incremental
steps that produce incremental value at any step along the way, so it is not a
push phenomenon, it is simply a pull phenomenon, and that adoption proceeds
DR. MC GEENEY: One comment from a personal concern is, if the bar is set
such initially, because there may be dollars associated with that bar, that the
primary care practices or any small practices that John put out across this
country are not precluded from that.
I think it is and can be a pull phenomenon, not a push phenomenon. There is
tremendous motivation and momentum around this country among primary care to
make meaningful change. But I think the incentives need to be aligned in such a
way that it is not exclusive. That is a challenge. I think we need to think
MR. REYNOLDS: Last comment, and then we will break for lunch.
DR. CLASSEN: I would just add to it that the bar may be set differently if
we are creating a road map, Paul, for the inpatient versus the ambulatory
environment, to take into account the realities of what we found and what
actually happens after implementation.
MR. REYNOLDS: Thank you, everyone. Well done, and I appreciate the good
One bit of housekeeping. We had some questions on submitting written
comments. I will state how you do that, but we will put something on the table
over there that reminds people of what that is. You would send those comments
to Marietta.firstname.lastname@example.org. We will note that on the table over there.
We will start promptly at one, promptly being the key word. So whoever is
up at the table will get to talk first. See you at one. Thank you.
(The meeting recessed for lunch at 12:00 p.m., to reconvene at 1:00 p.m.)
A F T E R N
O O N S E S S I
O N (1:00 p.m.)
MR. REYNOLDS: Let’s go ahead and get started. I will make a few preliminary
comments. For those of you on the panel, you might not have been here this
morning. Welcome. You will notice there is a light box in the middle there. For
those of you that weren’t there, we can eject you from your seat when it turns
red. You don’t fly far, I promise you, we just kind of bump you a little bit.
It will turn yellow after eight minutes, so that will give you your two minute
warning. Again, we want to make sure that everybody gets through it, and that
we have ample time to drill down.
So with that, our panel is William Bernstein, Rachel Block, Dr. Marc
Overhage and Gina Perez. Is there any reason that anybody would want to change
the order that I just read? If not, then we will proceed. So, Mr. Bernstein,
MR. BERNSTEIN: Thank you. My name is Bill Bernstein, and I am chair of the
Health Care Division of Manatt Phelps and Phillips. I am delighted to be here
today talking with you.
My work in the health IT arena is quite varied and covers multiple
perspectives. I have been working with the American Health Information and
Management Association on their state level HIE consensus project. I have been
working with a number of individual states, California, New York, Tennessee,
Nebraska, Colorado, and I have been working with a variety of health care
providers and other types of health care organizations.
Before I get into my specific thoughts on the appropriate definition of
meaningful use, I wanted to start with a few overarching points about what I
think HIE and the use of EHRs should be aspiring to achieve.
I guess my first observation is that HIE and EHRs should not be thought of
separately. Our objective should not be simply to migrate clinical work flows
from paper to digital media, but to insure the timely transfer of patient
information in a standardized and actionable format across the entire continuum
of care, for myriad uses including clinical decision support, health
management, quality improvement and population health reporting. I was pleased
that from this morning’s testimony there seemed to be a lot of agreement about
In my view, HIE is the link that makes the aspirations of EHR use real.
This means that the critical functionalities providers should be required to
demonstrate in order to earn an incentive as a meaningful user of certified EHR
technology should rely on Internet based architecture services that in many
respects will redefine what our current view of EHR products is.
Second, given the link between HIE and EHRs, the incentives should be
structured accordingly. I really think this boils down to two points. One,
incentives paid to providers should be for EHRs, HIE and the programs and
services that result from their combined use, not EHR hardware and software
alone. I think it was referred to in some of the comments today, where
particularly given the two billion dollars up front funding as opposed to the
$34 billion in outlays that CBO estimates might come, in order to make this
work we are going to have to be really explicit about how the $34 billion is
used to support our vision.
Second, we are going to need to establish a process by which providers may
obtain a waiver in my view to have their EHR incentive payments start at a
later date, if it is determined that HIE services are not available, and they
demonstrate a viable plan to achieve HIE activity in the near future. Again,
because I see this link between HIE and EHRs being so crucial because I
wouldn’t want people to lose out on incentive payments, and I am worried about
the timing, I think the creation of a waiver process will be very important.
The third observation is, to succeed I think HIE must be implemented on a
statewide or regional basis in accordance with a well defined federal
framework. Letting 100 HIE flowers bloom will perpetuate the siloed system we
are trying to replace.
This means that the HIE infrastructure should be built upon common
nationwide information policy standards and technical approaches as well as
state and regional bottom up implementation approaches and care coordination to
allow local communities and regions to structure their own efforts based on
clinical and patient priorities.
I think the main point I have here is the importance of governance level at
the state or regional level for us to succeed. Of course we need federal policy
and standards, but I think if we don’t have implementing arms, this won’t work.
I think the federal government accordingly must make the hard decisions and
quite rapidly to set specific rules for governance technology and policy
requirements that all HIEs must comply with. I think we have a number of
laboratories which we can draw upon to identify these standards, but I think it
is urgent that they be created up front so people have a road map to follow.
If we fail to make these hard decisions, high tech will result in our
pouring the same wine into new bottles, with the false hope that it will taste
Now I wanted to turn to driving the market through what I see as minimum
requirement. I am suggesting that there be minimum requirements in two
respects. One, minimum health information exchange organizational and
functional requirements and two, minimum HIE use requirements, very much in
some ways the latter discussion paralleling discussions we had this morning.
Like others, I believe that the minimum requirements can be increased over
time to reflect increased expectations for performance, but it is essential
that we not set the bar too low to begin with. I think this is particularly
true with respect to HIE infrastructure, because I think the nature of an
infrastructure investment as opposed to a use investment is that you have got
to get it right up front and have it working, otherwise the whole system will
I guess I would say from my own experience working in different states
around the country, markets will respond to governmental requirements,
especially when they are backed by significant economic incentives and
The HIE market to date has developed slowly over the last five years in
fractured and disparate ways. Participants have used HIE services to very
different end. Most HIEs develop from the ground up, with very little federal
or state guidance for consistency in use case development, technical
architecture and policy. In large measure, the early HIE structures were
influenced by the goals of their sponsors and the reality of funding
imperatives. Consequently, each HIE has its own stories. From Indiana, with its
initial focus on clinical messaging between area hospitals, to Massachusetts in
its wiring of three separate communities, to New York and its focus on a
publicly financed state health information network.
There are many lessons to be learned from each of these initiatives, as
well as from others in other jurisdictions. But the reality is that each is a
product of its own circumstances. High tech represents a new opportunity. Its
level of investment combined with the urgency of reform of our broken health
care system makes it imperative that a national consistent framework be
developed to guide local and regional HIE efforts.
Let me talk a little bit about specifically what I have in mind. One, I
would suggest HHS should define minimum HIE organizational and functional
requirements and require all participants of high tech meaningful use payments
to contract with HIEs that meet these minimum requirements. I think taking such
a structured approach to HIE in itself, will have a big impact on the market.
Two, I think that these requirements need to cover governance requirements
of insuring a multi stakeholder decision making process. They need to cover
technology specifications, and they can’t just be federal technology
specifications, they have to take into account state law variances and privacy.
Three, there have to be very specific both federal and state policies and
procedures relating to things like authentication, authorization, access, audit
My second point is that there needs to be a way to enforce the requirement
either through federal regulation or accreditation. I would prefer
accreditation for various reasons, but enforcement is a key. Three, I think a
significant portion of the meaningful use payments have to go to HIE
My last point with respect to this is, I would respectfully urge that a
significant portion of the two billion dollar initial investment be used for
HIE functional and organizational requirements are not enough. There must
also be minimum HIE use requirements. I think what I have in mind in terms of
the use requirements go very much to creations of medical homes or clinical
decision support and mirror a lot of the conversation that was happening this
morning, not in the context of merely an EHR implementation, but in the context
of a community wide health information exchange implementation. I fundamentally
believe that a lot of the more sophisticated outcomes that we want can happen
through collaborative efforts.
With that, I’ll close.
MS. BLOCK: Thank you very much. My name is Rachel Block. I am the Executive
Director of the New York E-Health Collaborative. We are a statewide multi
stakeholder collaboration that has been working at the intersection of health
information exchange and EHR adoption efforts in New York. I think what you
will hear from me is consistency in some of the themes which Bill Bernstein has
included in his testimony, but a couple of very specific examples of how we
have actually been operationalizing this kind of an approach in New York.
From the earlier discussions, I think I don’t really need to remind this
group as to why we are having this conversation. But I do think it is important
to remember that our efforts here are based on a strong imperative backing
health care reform. The first panel today elaborated on these themes. We have
quite a substantial basis of consensus based reports that have painted a road
map for how we should be thinking about the significance of health IT adoption
But the other thing that those reports have pointed out is that health HIT
alone is necessary but not sufficient. Health information exchange is an
essential component in an overall strategy, both as a way to create further
incentives to support health IT adoption, to track care across the continuum,
to empower consumers and families, create longitudinal information that many in
this room have been looking for, and to strengthen public health surveillance.
Obviously this last point is further amplified by a recent event.
So in terms of some of the key concepts of what needs to be included in
health information exchange, there has correctly been a lot of emphasis on
standardization of technical requirements. We along with my colleagues here on
this panel and others in the room have been participating in the Nationwide
Health Information Network project, and a big focus there has been to take
those technical standards and figure out how they actually can be implemented
in the real world.
But what all of us I think also would tell you is that the technical
standards alone are not sufficient to implement health information exchange.
The broad acceptance of use of other policies and standards is also required. I
will touch very briefly on these, because I think that they should be
integrated with the committee’s view and the Department’s view of how
meaningful use should actually be operationalized. So governance and
accountability policies regarding the protections for access and use, technical
policies and standards, the clinical imperatives and finally, financing
incentives, which of course is a big focus for today’s discussion.
New York State’s involvement in this in addition to the formation of the
E-Health Collaborative, which is a voluntary public-private partnership group,
has also been spurred by the significant investment of public funds that New
York has committed to advance this. Already $150 million invested, another $100
million just recently announced in additional activities.
A big reason that the state of New York and the New York E-Health
Collaborative have the ability to move forward with a very comprehensive
approach is we are providing the funding, or putting our money where our mouths
are, so to speak. I think that this example hopefully will be one which the
federal government can now look to as it moves forward with the further
development of its own policy framework in light of the stimulus funding.
Governance and accountability. I think few people would argue that this is
a necessary component in the overall health information exchange
infrastructure. Public trust is key. There has to be, as I was just mentioning
with respect to the funding, authority, structure and the means to set policy
to monitor use and to ultimately enforce standards.
The federal government to date has established some policies and
mechanisms, but thus far this has not really been a comprehensive approach. As
Bill suggested, we believe that it is critically important to establish
governance and organizational capabilities at the state and regional levels
that will complete the picture.
Speaking of pictures, this is a picture of how we have organized things in
New York. Because it is an HIT slide, there is a cloud in the middle. But what
I would like you to notice here is the top part, which is the statewide
organizations that are involved in the development of policy and standards, but
the bottom part as well, which are the regional health information
organizations and the other groups that are focused on adoption support
services. They represent the bottom up part of our process. All of that feeds
into how the statewide health information network for New York will operate.
A key piece of this is policies governing access to and use of data. This
is not the topic for today, but the committee has done extensive work in this
area, so I don’t have to explain why it is important. Similarly, I think
through the earlier presentations we talked about the importance of technical
One of our points would be that technical standards alone are not enough to
move us away from individual interface specifications to something which is
more broadly based on a service-oriented architecture. That service oriented
architecture itself has to include standards and policies relating to security,
authentication, authorization and audit. These are all key components in terms
of how a health information exchange needs to operate in a way that would
support meaningful EHR use.
We could implement meaningful use of EHRs through this model, which is
basically the confusing spaghetti of how the health care system currently works
and how EHRs and other systems connect to each other.
But what we are advancing and many others around the country are now
working to adopt is to work with more of a service oriented architecture model.
My job here today isn’t so much to go into the technical details of this, but
there is a cool slide there for those of you who are in the propeller head
segment of the audience.
So we have governance, we have technical standards for interoperability.
There are two other key components that I think are really important for health
information exchange to be able to play a role in supporting meaningful use.
One is the clinical imperatives for health information exchange. As Bill
suggested and we are implementing, structured use cases that address the
continuum of care, safety and effectiveness of care, quality improvement and
reporting and population health surveillance and management. These are all
integrated into both our health information exchange process as well as our EHR
I particularly want to draw your attention to the middle part of the slide
which I highlighted in yellow. These are the key steps which we felt were
necessary in order to thoroughly understand not only the definition of the
technical requirements associated with use cases, but also actual clinical
scenarios that would illustrate how these use cases work, work flow
requirements associated with them, the clinical requirements involved in
technical development and implementation of EHRs and HIE, and finally, the
important relationship to payment policy. We see these all as being very
integrally related in terms of how EHRs and HIEs will work together.
One key final point in terms of how we view EHR adoption as being critical
to the success of our health information exchange efforts is through a
comprehensive system of EHR adoption support services, which is represented in
this slide. Basically what we have here is what we refer to as our soup to nuts
approach to support clinicians at the front lines in terms of adoption. But it
is a key element to our health information exchange strategy.
This is one final illustration of how we envision this all fitting together
and the integral relationship between meaningful use, which you could view as
the combination of the top part, those EHRs and PHRs and other tools being used
at the front lines, the clinical informatics services that are necessary to
aggregate and manipulate data in order to make it useful for measurement and
reporting, and finally, all of this supported by the statewide health
information network of New York as the infrastructure.
What we have done is cross walked from our perspective the requirements in
the statute with respect to meaningful use. You can see that there is a direct
alignment from our perspective in terms of how EHR meaningful use could support
this kind of a comprehensive approach. It is critical for financing and
incentives to be aligned to support the development and use of the whole
So we think it is critically important for HHS to make sure that the HIE
infrastructure investment, which is another part of the stimulus legislation,
should go hand in hand with the policy and funding support for EHR use. This
will require HHS to establish specific roles for states to coordinate
implementation, to develop broad standards for HIE governance and functions, to
align the incentives for HIE development, and EHR implementation, and finally
to create a better road map for how we will support future systems enhancements
that will address false certification and payment policies.
MR. REYNOLDS: Dr. Overhage.
DR. OVERHAGE: Good afternoon. Thank you for the opportunity to testify
before the committee on this important topic. While I am a member of the
committee, I am testifying today as the Director of Medical Informatics at the
Regenstrief Institute and the CEO and President of the Indiana Health
Information Exchange. Through those organizations I have over 15 years of
experience creating, managing and studying health information exchange. In that
exchange today, we provide a broad variety of services supporting data flows
from over 80 hospitals as well as independent laboratories, radiology centers,
public and private payers, public health and over 10,600 physicians and over
We deliver about three million clinical results each month, and our
centrally managed federated repository contains over and a half billion
structured results for approximately nine million patients. More importantly,
we have been able to use this infrastructure to demonstrate improved efficiency
and have strong anecdotal evidence for improvements in care process.
There are these little nuggets that keep you going sometimes. Last week we
got one of those, which was an e-mail from a practice that said, thanks,
through the HIE certification services, we detected three patients’ colorectal
cancer that we otherwise would have missed in the last month. It is those
little nuggets that make this all worthwhile, and I think frankly what all of
this work is about.
I think health information exchange involves talking about at least three
distinct roles when you are considering meaningful use. First, electronic
health records cannot fulfill their potential unless they contain the necessary
structured data about a patient. While important clinical data is certainly
generated at the practice, at the point of care, whether those are vital signs
or social histories, a large amount of the data that is important for clinical
quality improvement, for clinical decision support and other sorts of uses come
from outside of the practice, laboratory results, diagnostic studies for
cardiac abnormalities, and medication dispensing events, for example.
Much of this information is already available electronically in a
structured form in systems that are capable of delivering these data through
electronic interfaces. But providers can’t be expected to enter these data or
create the spaghetti, point to point electronic means to capture these data,
and EHRs cannot deliver on their promise without these data.
Second, health information exchanges provide certain community services
that have to be shared and are difficult or impossible to implement within an
EHR silo. This recapitulates Bill Bernstein’s point about EHRs and HIEs
together. Examples of these community services include patient matching,
provider matching, data aggregation, message routing and data normalization.
Just to highlight a few of those a little bit further, secure private data
aggregation is an essential community service that health information exchanges
can provide. While a variety of different technological approaches may be used
to provide this service, any single EHR with the possible exception of a
primary care provider in a medical home setting is unlikely to aggregate and
maintain longitudinally and provide access to all the others who need it the
patient’s information over time.
Message routing is another example of a valuable service that manages the
flow of information between providers and brings significant efficiencies by
centralizing the management of the points of contact for clinical trading
partners, and providing secure transmission.
As an example of that that is timely, the Indiana State Department of
Health was able to deliver today to selected physicians by specialty, by region
of the state that they practice in, into their routine clinical work flow
alerts around the swine flu outbreaks, and appropriate clinical management of
patients, thanks to support from the CDC contract.
Data sources like laboratories can have a very large number of providers
with which they exchange information. Physicians have a significant number of
other providers with which they share patient care, and a physician’s
prescriptions may be processed in a large number of pharmacies, for example.
As a final example of a community service, HIEs are uniquely positioned to
normalize the data across sources. If we want to wait until every vendor system
can deliver structured normalized data in a common format with common codes, we
are going to be waiting a very long time, and I am frankly not willing to wait
that long. HIEs can bring expertise, understanding of core common data sets and
structures and bring visibility across organizations that helps normalize that
data and make it most useful for care.
The third issue that I think you have to consider when you are thinking
about health information exchanges in regard to meaningful use is, an HIE can
provide functions that are more efficient or effective when provided at the
community level. An example of these might include electronic laboratory
reporting for public health, electronic health record light, as Dr. Middleton
described this morning, and quality reporting.
Electronic laboratory reporting for public health for example requires the
development and maintenance of sophisticated rules as well as the ability to
deliver results to appropriate — and that is an important issue — public
health departments, while full function EHRs supported by the HIE are likely to
ultimately provide the most improvements in quality and efficiency.
A less sophisticated tool with fewer features, sometimes referred to as EHR
light, may be able to deliver significant value by organizing and delivering
data to providers along with simple decision support. Health care informatics
researchers demonstrated many years ago that these simple tools can have
significant measurable improvements in quality, safety and efficiency of care,
and they shouldn’t miss that opportunity.
Quality reporting is the final example of a function that may be more
effective at the community level than at the individual EMR level. The effort
required to maintain the logic, the validation processes, data flows and
supporting the required reporting formats may be distributor shared at a
community level rather than every practice having to do that work themselves.
Indiana’s Quality Health First program for example supports quality
measurement and reporting to multiple payors and programs, including CMS’
physician quality reporting initiative for almost 1.7 million patients and over
900 primary care physicians throughout the health information exchange.
Unfortunately not every community has an established infrastructure.
Despite their potential, there is only a modest number, and you can debate a
lot about what that number is, maybe 30, operational health information
exchanges in the country today. Those information exchanges support information
flows for probably less than 30 million people in our country.
There are a variety of reasons that experts have identified for that slow
growth in size and number of health information exchanges, including
misalignment of incentives, a lack of standards at option, legal and privacy
concerns, and lack of leadership. While certainly the Recovery and Reinvestment
Act require that EHRs be connected in order to qualify for incentive payments,
it will undoubtedly stimulate interest and energy around health information
exchange. Given the challenges and time required to advance this work, the
landscape is unlikely to change much before fiscal year 2011, in that short
18-month interval. Proper incentives and approaches may overcome these
challenges, and you have heard some examples already today.
Just for example in the incentives, if the Recovery and Reinvestment Act
requirement for an EHR to be connected was measured based on the availability
of data a provider generates in other provider systems, provides will be
strongly incented to overcome the barriers to health information exchange.
From the provider’s perspective, meaningful use of an HIE services or
functions begins with making structured data from the provider’s EHR available
through the HIE, progress through using the HIE for routing, and incorporating
data delivered by the HIE into their electronic health record.
There is a parallel path of progression for health information exchanges.
They can progress from routing messages, which require functions like provider
matching and normalizing data structures through data aggregation, which
requires patient matching and common data coding, and on to sophisticated
community functions like quality reporting and clinical decision feedback that
requires advanced query capabilities, and logic engines.
In conclusion, I believe that the availability of data that a health
information exchange enables is necessary, though not sufficient to improve the
quality, safety and efficiency of care that our patients receive. In addition,
health information exchange offers value independent of electronic health
records it connects. It also amplifies the value that EHRs can deliver.
Thanks for the opportunity to discuss this topic with you. I look forward
to the questions.
MR. REYNOLDS: Thank you.
MS. PEREZ: Thank you. My name is Gina Perez. I am Executive Director of the
Delaware Health Information Network. There is always a challenge with going
last. I hope that I won’t bore you with repeating, but rather will summarize
what you already have heard today.
DHIN is a statewide health information exchange operational since 2007. It
includes more than 80 percent of hospital data and outpatient lab data in the
state of Delaware, in both ambulatory settings and in hospital settings. We
have almost 50 percent of the physicians using the system in Delaware, and we
share about 40 million transactions per year, with about 600,000 patient
records, and growing.
I would like to start with what we believe in Delaware have been the
benefit of health information exchange from two years of experience in that
data moving around the state. We believe that health information exchange
creates time savings for physician practices and for hospitals and labs and
radiology facilities that are sending data into the system. It is one pipeline
to get information from multiple sources into an electronic medical record
system or other options like a portal, to physicians who need that information
at the time and place of care.
We have one standard format for how those results and reports are
delivered, easing the burden on the practice to manage that information. We
even improve care and patient safety through real time reporting resulting in
faster access to critical information. It facilitates the flow of information
to those who need to know, and improves patient care in emergent situations, as
well as faster response to outbreaks which we are all facing right now with the
In terms of reducing costs, we see this every day in physician practices,
as they have been able to reallocate resources from managing the paper to being
able to provide patient care. They are spending less time moving information
around or chasing down information that they don’t have, and more time caring
for patients, thus better patient outcomes and saved time and resources and
We also believe that the health information exchange enhances privacy.
Through access control and system auditing, we have safer records. We know who
has been in that system and we know what they have been looking at, and the
patient has the opportunity to see that information as well. In Delaware we see
these benefits every day, exemplified by the widespread use of this system and
the growing adoption.
In terms of meaningful use, I would like to talk in terms of three
categories, critical masses of information, functionality in a system and
In terms of critical mass, this is about what is happening in a community.
The community must define what critical mass means to them. To be successful,
health information must include data available when the provider is looking for
it, data that is useful at the time of care, and information that is easily
accessible. That means when a doctor logs into the health information exchange,
you need to find the information that is clinically relevant. They need to have
it at their fingertips, and it needs to be there every time they turn that
system on. There needs to be relevant sources of data, in other words, we need
to have as many opportunities for information to come into the system as
possible. We need to have as many patient records as possible so when that
doctor searches for their patient, that patient is there and they are not
getting a blank screen, and there needs to be enough history that that
information is meaningful. It is up to each individual to understand and to
quantify what enough information is. For some physicians six months of data
might be good enough, for others they might want two or more years of
longitudinal information. The system needs to be flexible enough to work with
the needs of that provider.
In terms of functionality, I believe functionality is local. It depends on
what that community looks like and what is important to that community and what
the business needs are for that community. For those communities at risk for
natural disaster, the opportunity to query a system and retrieve information
might be much more important than a community that has a high rate of
electronic medical record adoption, where getting results into that electronic
medical record system through one source is most important.
But there are some generally accepted roles for health information
exchange. They include results delivery at all levels of health information
exchange adoption, so those paper based practices need to have as much
opportunity to receive and access information as those that are fully
automated. Patient records first is a function whereby a provider can search a
patient’s history in the health information exchange. This is usually driven by
a master patient index and record locator service.
Medication history is something that every provider we talked to says that
is the number one most important thing they have. It helps them manage patient
care, it helps them address costs for both the patient and the health care
system. It helps them identify those that are inappropriately accessing
medications. There are also other opportunities in which we can better interact
with electronic medical record systems. The ability to pull data into the
electronic medical record system on a query. The ability for a provider to be
able to order out of their electronic medical record system and the results
come back and match with that order in their EMR. All of these types of
functionalities make that investment in EHR more valuable for the provider. It
makes the EHR more valuable in terms of the information available and the use
of that information, and it results in good patient care. I think the
alternative to not having that connectivity is a siloed environment where that
electronic medical record system is simply an additional cost to provide care.
Public health reporting is another function that we in Delaware believe is
very important, and I think is going to be a growing need as we look at
In terms of interoperability, there are multiple levels in which we need to
look at interoperability. There is horizontal integration, which is looking at
integration across hospitals and labs and radiology facilities and other types
of data sources.
There is vertical integration, which is the need to get data out of the
hospital’s labs and sources and into the hands of the providers providing
patient care, whether that patient care is in an emergency department or in a
first responder situation or in a physician practice.
There also needs to be EHR integration, again providing additional value
for that EMR and the ability to do better reporting from that EMR, and the
ability to place that result in the patient’s record in the EMR so that the
physician has less work to do, and knowing that that result is there.
My concern though is, what about the rest? What about those providers who
have made an investment in electronic medical record systems that are not CCHIT
certified? What happens to them? Will they be penalized, or will CCHIT
certification be looked at in terms of casting a wider net?
What about for those who have not adopted electronic health records for
reasons of financing or capacity within that practice to do so? We all know
that it is not just a plug and play environment, that there needs to be a lot
of work flow change and support for that EHR to operate correctly in that
Regardless of incentives or disincentives, there may be practices who
cannot implement electronic health records. What about those practices? Will
EHR light be an opportunity to meet the needs of those practices, for instance,
providing reporting analytics and e-prescribing and connectivity to practice
management systems through the health information exchange. I think those need
to be considered as opportunities or options for those providers.
In terms of where HIE might be in 2011, I think we can be making great
progress, or I think we could be regressing. So I think the way that money is
spent is very critical. The stimulus funding can be a great benefit in growing
successful models for health information exchange across the nation, and moving
us toward the NHIN goal of having a network of networks. Alternatively, it can
result in poorly spent funding that can be devastating for the health care
system, further fragmenting information and patient care.
So we have the opportunity to insure that adequate planning and
collaboration occurs in building a strong foundation for health information
exchange, going back to those themes that you heard about, governance and
sustainability, needing to build upon best practices. Let’s learn from those of
us sitting here and those others across the country who have made great strides
in health information exchange, but let’s learn from those who didn’t quite
make it because there wasn’t a strong foundation, because not all the
stakeholders were sitting around the table, because there wasn’t good planning
involved, and understanding what the needs and the business challenges were in
Let’s look at engaging all stakeholders, including consumers, providers,
payers and government, and implementing standards that foster meaningful use
and connectivity across networks. We need to make sure this opportunity is not
wasted and this is not money poorly spent.
I believe that the benefits to having meaningful use through health
information exchange are again back to that one interface with multiple
sources. We cannot underestimate the benefit to a hospital, to a lab who can
start managing their information through one pipeline versus multiple
pipelines. That is very costly and difficult to maintain investment.
One standard format for results in reports. Let’s streamline the system so
that we make it easier for those who have to use the system, those who have to
provide patient care. Meeting the needs of the providers along the adoption
curve and helping them transition from paper to paperless, and providing
information at the time and place of care, providing critical alerts and out of
range results so that the providers can take immediate action on those things
that are most important to their patients.
MR. REYNOLDS: If you could wrap this up quickly, please.
MS. PEREZ: But let’s also recognize that there are significant challenges.
Government can help foster efficient and effective health information exchange
for those who benefit from the system, and those who benefit must also be
sharing in the costs as well as the savings. HIE is about building consensus
and focusing on what is best for the patient. In order for health information
exchange to be successful, it requires that participants must put aside IT as a
competitive advantage and focus on technology as a facilitator to improve
patient care and clinical outcomes.
Stimulus funding is to stimulate HIT, not to sustain it. So therefore we
must look at ways to evaluate and replicate success and establish a baseline
for long term sustainability.
Thank you for the opportunity.
MR. REYNOLDS: Jeff Blair. Since we don’t have much time, if everybody would
precisely get to your question, please. No stories.
MR. BLAIR: It is nice to see the successful HIE networks. You have pointed
out that there is a lot who have been struggling, whether it is financial self
sufficiency or privacy issues, getting data or the fragmentation among HIEs.
HIEs mean e-prescribing to some folks and clinical data to others, and HIPAA
transactions to others and personal health records to others.
The nation is expecting that meaningful use —
MR. REYNOLDS: Cut to your question, please.
MR. BLAIR: Yes, is going to be there in a couple of years so we can have
data coordination. What is the one or two things that you feel the nation needs
to do to make sure that HIE networks are part of this infrastructure two to
three years from now?
DR. OVERHAGE: I’ll take a first crack at that. I think part of the answer
is, we have to simplify and focus on a fairly narrow range of opportunities.
If you look at the data that we need for clinical care, for quality
improvement. It is a broad spectrum, but there are certain high value areas to
focus, laboratory results, medications, radiology results, for example, are
enough to take us a long way down the road. Rather than trying to do everything
and deal with everything, focusing on those few narrow things, and doing it in
a way that can be implemented within our current infrastructure, recognizing
the constraints of where the technology that is in place today is.
If the laboratory systems today can generate data and deliver it, they are
not going to send it out in a CDA next week. I think we have to meet the market
where it is and evolve from there if we are going to be successful at providing
the infrastructure that we need to help EHRs reach their potential.
MR. BERNSTEIN: There is a lot I love about the stimulus bill, but I think
there is not enough money up front to build the required infrastructure that we
have all heard about in this panel. So if I had to say, there is a lot to be
done with governance and technology, but the single most important thing that
is going to move the market, because I have faith in the market out there to
respond to incentives in some ways, it is to look at the meaningful use
payments and to get very specific about what the expectation is about how they
are going to be spent and the relationship between EHR and HIE, so that some of
the benefits that are substantial and that this panel has demonstrated can be
realized, and to do that in a way which — I agree with Gina — empowers,
whether it is state or regional, it is going to be different in different
The nice part about this is that all across the country right now, there
are really motivated people who want to get into this and do it and are
organizing themselves to do it, so take advantage of that energy in terms of
how you fund.
MR. HOUSTON: As part of the stimulus package, meaningful use doesn’t
include information exchange as one of the key components. As I listened here,
I get this very strong sense as others, there is a big difference as to how
active HIEs are. One simply facilitates passive information, others almost seem
as though they are the repository for a community of information that
physicians will draw upon in small practices and otherwise.
So the question I have is, I guess I have two questions. One is —
MR. REYNOLDS: One.
MR. HOUSTON: Is there a possibility, and what are the standards that we
need to apply here? I hear from this panel there is such a variance in opinion
or application as to what is part of an HIE. Is it the lowest common
denominator? I guess I am looking for some type of guidance as to how far
active these HIEs are. I get the sense that they are all over the map, at least
from what you are saying. DR. OVERHAGE: I think Gina said it. At their core,
HIEs are about freeing the data, making the data look good, whatever phrase you
want to use. Then you build on top of that. A stream of normalized data is a
powerful tool. It can be used by an EHR, it can be used by an application at
the community level. You can choose, but at its core it is about freeing the
data and making it available in a private secure way that empowers innovation
at the edges. That is the thing that is at the heart of all of them.
MS. BLOCK: Just to add to that, I think what I noticed in listening to the
panel is, we have only been talking about this in an active way for about two
years as a group. There is actually remarkable consensus, I thought, across
this group in terms of what the necessary attributes are to put a system
So I think that while Marc has expressed the goal and Gina did very well in
their testimony as to the ultimate outcome that we are seeking, the building
blocks for how to get there, I believe, have been identified. HHS through its
guidance for the various funding streams that will go out for stimulus has the
opportunity to reinforce those building blocks in a consistent way across the
DR. FRANCIS: This is a question for both William and Rachel. Each of you
emphasized the importance of state and regional variations in governance. I
take each of you to be saying that an aspect of this was state variations in
privacy laws, with presumably a federal floor.
I wonder if each of you could comment on what you see as some of the
issues, particularly if we move to a RIO of RIOs or networks communicating with
each other and whether you see a role for something more like a stronger
federal preemption with respect to privacy in this area.
MR. BERNSTEIN: I have done a lot of work both in New York and California,
which have privacy laws that go much further than HIPAA. I take as a starting
point for this discussion that there is not going to be federal preemption of
state law. So I think the reality is, in states like that you need to design
around current state law.
Then the second thing, and somebody said it in this morning’s panel, health
information exchange is fundamentally different than point to point. It is the
capacity to get information that is so much more powerful. I think you have to
take that into account in how you divide your privacy policies.
In New York it took two and a half years to get there, but the RIOs were
all adopting a standardized consent form which applies to all data, sensitive
health information as well as non-sensitive health information. I think as a
community we are going to be much better off for having standardized the
process and done it in a way where we didn’t exclude filtered data from the
exchange because it can lead to much more powerful results in terms of clinical
MS. PEREZ: I would just like to add that I think it is important that the
privacy aspect be handled locally, because it is that community that has to be
comfortable with how the health information exchange is operating. So a
preemption of that community’s requirements would be a challenge potentially.
A concern that we have also is that there are requirements that put
constrains on what data can be available. A provider may actually make
detrimental decisions by not having a full complement of the information. So in
other words, if they are assuming that by clearing the system they have all the
information, that that patient has opted some information out, that decision
making could be more challenging because not all the information is there.
So I think we need to be very careful about how privacy is looked at and
what kind of options are put on the table.
MR. REYNOLDS: Walter, you have the last question. DR. SUAREZ: My question
is very quick. You all have said during the testimony in different ways the
need to have some standardization of the concepts around functionality and
characteristics of HIE. My question is, given the fact that as a fundamental
building block of EHRs and NHIN and HIEs, do you believe that there is a need
to formally establish a certification process for HIEs, just like we have a
certification process for EHRs?
You all have said concepts around the idea of having some sort of an
oversight or a body that can look at whether the functions and the
characteristics of HIEs are being met. But I wanted to push it to that level.
Do you believe that there is a need to have formal certification of HIEs in
MS. PEREZ: DHIT does have a network certification work group that is
looking at what those requirements might be. There is also a belief that it is
not the HIE that should be certified, but rather the technology that supports
the health information exchange, just as it is not the physician practice that
gets certified, it is the technology that provides electronic health record
So I think there are two fields of belief in terms of that certification
process, and I think they both have merit.
MS. BLOCK: Just to add, some of the work that we have done was looking
beyond just the certification bodies and standards, but looking to
accreditation models and similar kinds of public-private regulatory frameworks
that might be applicable to and cover the broader governance policy components
which would not be addressed through the certification process alone.
MS. TRUDEL: I’m not directing this to anyone in particular. What is your
anticipation in terms of 2011 would be the penetration of HIEs across the
population, and in areas where those HIEs are not present, is there any other
alternative for information exchanges?
MS. PEREZ: I think I spoke to the idea of an EHR light, which I think is an
opportunity for an interim step for those that might not be ready or able to
adopt an electronic medical record system.
In terms of adoption, I don’t know. I think each community is going to be a
little different. I know in Delaware we have a pretty significant adoption of
electronic medical record systems today, but I think each community is going to
be different. How the funding is distributed is going to make a big difference.
MS. TRUDEL: No, my question was about HIEs, not EHRs. In communities where
there is no HIE, what is the alternative for any health information exchange at
MS. PEREZ: I know there are some communities that are looking at networks
of HIEs, so that is certainly an opportunity to exchange information across
providers. Hospitals are also doing a pretty good job in communities where
there is not HIEs to get information out to providers in various ways through
portals and direct interfaces with EMRs. It is not the most efficient, but
there are opportunities.
MR. BERNSTEIN: I think one answer to that question is if federal policy is
strong related to HIE development, you will get one result. If it is weak you
will get another result.
What my hope is, none of us really know how this works, but as you get
critical mass there will be clustering. What I mean by clustering is that there
will be many governance entities, because I think we will organize around care
delivery systems. But from a technology standpoint, a lot of the same
technologies will be used throughout many, many different governance networks
and we will get quicker at it, as a result of mass. That would be the hope. But
I think it is a hard question.
MR. REYNOLDS: I’d like to thank the panel. We really appreciate it. If the
next panel will please move forward we will begin immediately. Thanks very
The next panel is on meaningful use capacity as it relates to functionality
in quality reporting. I would like to go in the order that you are listed on
the agenda, unless that is an issue for anyone. If not, then the first
presenter is Kristine Martin Anderson
MS. ANDERSON: Thank you very much. I am Kristine Martin Anderson, and I am
with Booz Allen Hamilton, where I work in the health information exchange
practice, consulting primarily to the federal government, but also to nonprofit
organizations and states.
I am speaking today mostly about quality. That is an area that we have done
quite a bit of work.
At Booz Allen we have done some recent work for clients to review the
landscape of quality measures that are created and regularly reported by
hospitals and physicians. Compared to some work done in 2006 we do see a
growing tendency toward reporting measures that are in public programs, but we
are not necessarily seeing a decrease in measures that are being reported that
are not necessarily in those programs.
I think there are two reasons for that. One is that there is still huge
gaps in the measure set. The National Priorities Partnership is working on
trying to fill some of those gaps. Another reason is that some of the existing
measures today aren’t easily implemented with our current health information
technology systems. We are increasingly seeing provider organizations making
tradeoffs between the burden and the value of precision, and beginning to take
what are normally started as standard measure sets and adapting them to those
in a way that allows them to report them with less burden. Therefore there are
versions of measures that actually vary according to the implementer. Very much
like the discussion earlier around health information technology or EMR
So even though in the beginning it looks like they are all implementing the
same measures, if you look post implementation there is actually quite a bit of
variability in what is actually being reported.
I start with a discussion of quality measures, because when the AHIC
Quality Work Group first started, Carolyn Clancy talked about a vision of
pressing the F7 key and automatically generating quality measures. Our first
reality check is that the measure set is still evolving to fill those critical
gaps. If you had an F7 key today that was enabled, it would have to be
customized programmed, just like most other things that we have today that is
already implemented in health care.
Now let’s take a bit of a closer look at the use of health It to measure
quality. We are in a challenging position as it relates to using health
information technology to improve care and enable quality measurement.
It is imperative that we do all we can to speed adoption of health
information technology, because it is simply not possible to get where we want
to go using paper. However, we are also aware that today’s health information
technology is under evolved as it relates to quality, due in part to the
just-mentioned under evolved quality measures. We have made progress, but not
yet hit sufficiency in many areas, including harmonization and adoption of
standards described treatment, which I’m sure will also be described by the
next speaker, development and adoption of effective clinical decision support,
harmonization of the quality measures themselves, and most importantly in my
mind, delineating a clear role for health IT including electronic health
records in quality measurement and reporting.
Since we simply can’t get comfortable with trying to change the tires on
the car as we are driving down the highway at full speed, I think we really
need to get to the point where we have good pit stops to find, where we can
define what is the road map, what is the management plan for how we are going
to get from where we are today to where we are going, how many laps can we take
with each set of tires before we have to stop and change them. We have to be in
this position where we are running down, we are going as fast as we can, and we
So working with the AHIC Quality Work Group, I became acutely aware of the
second law of thermodynamic supply. So perpetual motion machines are
impossible, and if we don’t apply direct energy to holding quality and health
IT together, over time it will average out to some uniform dead nothingness, in
other words, chaos prevails over order without intentional energy to prevent
So I think that is where we are with quality and health IT. There have been
a lot of attempts to hold the two together. They do tend to move apart every
couple of months, when you look back and say where did that effort come from,
and here we go again.
The next question becomes, how can we be intentional about keeping quality
and health IT aligned? How do we actually do this? There are many stakeholders.
Just naming a few, and there are lots of versions of the kind of slide that you
are seeing now, there are those that create the evidence, those that provide
the care, those who evaluate the care, those who pay for the care, those for
whom the care is given. Those stakeholders are not necessarily easily aligned.
AHRQ has been exploring how to migrate those who develop guidelines to
communicate in a way that can ultimately be understood by those who provide
care, and — and you don’t want to lose that part — to those who evaluate care
or the vendors that are trying to create some of the measures in an automated
way. We have got vendors, informatics professionals and measure developers who
have been focused on reducing the burden of measurement and trying to correctly
— as evidence is correctly applied, help prompt at the point of care those
providers to give care that matches the evidence.
Payers have been trying to move towards more universal measurement, trying
to amplify the signal to providers regarding what they expect the outcomes of
the treatments to be. So the good news is that these organizations generally
know about one another now. That is progress, just in the last couple of years.
So now that we have done a drive-by of the complex system of measures,
health IT and EHR development in the quality enterprise, I think we also need
to step back and think about the long term. The truth is, we just don’t want to
optimize the system around today’s measures. I think there is universal
agreement that that is not where we want to be. In fact, there is general
agreement that our current system of measuring quality, one doctor, hospital,
nursing home, at a time is temporary. It is all we have got for now, and we had
to start somewhere. We have already seen what happens when we create incentives
for each care provider to optimize their own situation.
We envision a future where care is coordinated and quality measures are
truly patient centered, where we can truly measure whether or not the system
works for any one patient. To do this, we need interoperability both at the
point of care and retrospectively for care measurement.
I want to just reflect back on Dr. Classen’s statement earlier about after
the implementation, when you do the comparison. When we think about quality
measures, we want to be looking at what happens for the patient, so that can be
the ultimate test. The question is, how do we get the inputs right so that in
the end we can get the outputs that will be uniform.
What should EHRs do in the short term? I tried to think about this quite
explicitly. I think there is a very clear role for health IT for supporting
quality at the point of care. I show it in this slide in the red part as an EHR
because that is the main focus, but it applies generally to other technology.
It is about the what, and not about the product in the end.
We need standardized data input. By that, I don’t just mean the work that
is being done on standards harmonization or the work in NQF or the HITEP, but
also about data that comes in from outside sources, how do you bring lab
results in, how do you bring medication history in, which probably shouldn’t be
done by EHR, but maybe more of a shared service, and how do we bring in some of
the effectiveness data that we are planning on having.
But we also need more clinical decision support and the ability to identify
panel certification where eligible for interventions, and whether that is done
through self populating registries or whether it is done by panels that are
done by others, critical functionality, and then critically, the data export,
how do you move data with the interoperability, what needs to be moved and
What is not so clear to me is that some functions like data aggregation or
measure calculation are inherently part of an EHR. I lived in a world where I
reprogrammed those measures, and it is a really tough thing to get it right. We
don’t need hundreds of entities trying to do it.
There is also an important role for EHRS in audit and audit and validation
and in feedback. How do you make sure that the clinician at the point of care
knows how they did in the past with similar patients?
The last point I want to make is around where I think we are. I think it is
possible to specify EHR requirements around quality. In New York they have
already done some of this work, where you can be very explicit. You can know
what EHR is capable of today, what the commercial EHRs are able to do. You can
know where your adoption risk is, so you actually know where you want to apply
your resources around technical assistance to help those who are adopting the
electronic health records to provide high quality care and have an ability to
I think we are at the point where we need that road map, and we just need
to decide in the big picture of all development of scope, resources and time.
We have begun to already look at the what and then the when and then the how
and who and the how much, and get to the point where we have a management plan
around how we are going to get where we need to be with regard to automated
measurement and high quality at the point of care.
MR. REYNOLDS: Thank you. Dr. Theresa Cullen.
DR. CULLEN: I am Terry. I am with the Public Health Service. I am the CIO
at Indian Health Service. I am going to share with you very briefly a large
view, and then I am going to answer the questions that we were given very
explicitly. I think this may be a unique opportunity to give some specific
items based on a long term personal and organizational history.
The Indian Health Service has had an electronic health record for 25 years.
Starting in 2000 because of performance based budgeting, the Government
Performance and Results Act required us to start reporting on our performance.
At that point, the Indian Health Service as a health care agency and a public
health agency elected to start reporting on clinical performance.
As you can imagine, our field pushed back. They are overworked, they are
tired. The last thing they wanted to do was collect data that they in fact
thought was already in the electronic health record system that they were
So starting in 2000, we developed an application, and you have a small
handout on that, the clinical reporting system. At the current time, CRS
reports on 300 quality measures. It reports quarterly on approximately 100
measures that are reported from 400 disparate sites in 36 states through an
export file. However, at the same time, the site is enabled to run on the fly
at any time they want those clinical quality measures.
They are used for accountability at a national level. They are used for
improvement at a local level. I think that when we discuss the value of
clinical quality, we need to recall that it has multiple uses, and that the use
that I may see as a CIO may be different than the provider at the local level.
I want to give you one other example. In the midst of the swine flu
epidemic last Friday, we have a reservation that straddles the United States
and Mexico, the size of Connecticut. Many tribal members are in fact Mexican
citizens. We were able through our health care system to put in early
surveillance based on the ICD codes that were given to us from the CDC to guide
The reason why I think that is important is that it once again points out
the ability for us to do that with dependence on a standard code set that had
already been imbued and embodied in the health care IT system and that were
being used regularly.
So your first question. I also have to tell you that I am a federal
employee for a very long time, so I recognize federal time. So even though
there is a sense of 2011, there is a much more impending deadline, which is the
beginning of 2010.
So what I have chosen to do is give you some guidance on what I think needs
to happen in what our agency would call a very rapid time frame, in six months.
You can see this here. We need to look at the data currently available in
I recognize there is an ongoing discussion about what a certified EHR is
going to be. I think there is some agreement however, and it was already
discussed. There are things like vitals, measurements, pharmaceutical data, lab
data that we can probably get consensus fairly rapidly on.
I would recall for you that the vast majority of labs that are produced in
this country do not have LOINC codes with them, so even though it seems like
there is a standard, right now at least for our reference size, only 25 percent
will report using LOINC unless we pay them a premium price.
Identification of a small set of performance measures. I think this is
critical, and I think these measures need to be decided upon very rapidly,
because without that, there is going to be pushback from not only the public
sector, but from the private sector, and this inability to understand where we
are going. That does not preclude that there should not be an attempt to do
rapid enhancement as we move along, but I think that this first step is a small
step that needs to understand where we are.
Development of performance measure definitions, and that was already
addressed, identification of the source of the data. Clinical decision support
remains large. It is in the ARRA bill. I think that to develop performance
measures without doing concurrent attention to clinical decision support will
insure that we do not move quality. They have to accompany each other.
The beginning of a structure for us would be called a meta data dictionary,
so that we really have a sense where these data fields are living. Once again,
because there is an assumption implicit in the bill that there will be
reporting, there is development of at least some open source guidelines for
what that reporting software would look at.
Now, I was very presumptuous. I took the opportunity to even give you some
measures. Obviously these are based on my personal experience, and you can see,
they are very low-hanging fruit. That was intentional.
The ten years that we have developed clinical quality, I was the front
woman for four of those years. I have the bruises to prove to you what 400
different sites did to me. In fact, I have been known to cry publicly because
of being tortured by our providers.
But what I looked at was the capture, retrieval and reporting. I elected to
not give you ideas about retrieval. I think retrieval is vendor dependent and
system dependent. The one thing I would tell you is that it is overcomable. It
is not a difficult thing to do. Once again, I did take the presumptuous liberty
of giving you some inpatient measures.
The real reason why I did that is because we have been subject to a
majority of measures. We are a federal agency; we seem to get named in every
order that can come out, so we have struggle with making measures consistent.
We have shared measures with the VA and DoD, and as you can imagine, because we
are all at different places in development, we all have different insights into
what those measures look like.
So detailed specification of a few measures. By the start of FY ’10, as
David Hunt reminded me earlier, really FY 10 and 11 start May 1, so it starts
this week. It is a little worrisome. Then current measures that are not defined
I proposed this previously. I think there needs to be a sandbox, a quality
software sandbox. A logical place may be at CMS where the public-private
centers come together. They work on these things, and they make sure they are
doable prior to throwing them over the transom. We have been accused in the
past of throwing software over the transom, and that can clearly cause
Quality data set. I think the most important thing here for me and
obviously I am a public health officer, is to improve quality of care and
achieve health equity. We need to imbue every baby step we do in this with that
belief, that we are trying to achieve that in the country. There are clearly
accountability measures. They may not match. They are both important, but I
think we cannot lose sight of that target.
The QDF needs to be developed for 2011 based on the information we have,
not on the what may be. Then we can rapidly move forward from that. In
addition, I have posited for a very long time that long term quality data sets
are important to perceive and review, not only from the patient perspective but
the population perspective. I believe that data should be available at the
point of care, the providers caring for a patient within a family and within a
population, and needs to have all that information.
Interim milestones, you can look at these. Once again you see a focus or at
least an attention to population health functionality within the architectural
design of the EHR. If we don’t get this right soon, we are never going to get
it, because everybody is going to have to go back and redesign.
Barriers are many. I think they are all overcomable. I would urge you that
based on ten years of experience doing clinical quality measures that a
two-year lead time is probably a minimal requirement to promulgate, develop and
produce new measures.
I want to remind people however that providers are committed to quality,
and that patients are committed to activation. This is a unique opportunity for
us to meet the needs of both the patient and the provider.
Infrastructure or policy requirements. These have already been addressed by
some of the previous speakers. Standard fire structure, standard secure
transmission methodology, a file acceptance criteria; if these files are coming
to CMS, what is going to cause them to reject, what is going to cause them to
include, and what is going to give the providers the incentive, which are huge.
The proposal is for some kind of data aggregation. The issue is, is that a
data warehouse, is it a QIO. I don’t know, I just put those things here.
Determination of which definition will be used. This was already talked
about. How do we know which performance measure is the performance measure to
be used for the control of hemoglobin A1C in diabetics? As you know, that is
all over the place.
Clear goals of quality reporting. Why are we doing this? It is not just to
get the incentives for the providers; it is to go back and improve the care and
the health status of the patients and the population.
Determination of the inclusion of administrative versus clinical data sets.
As you know, they can give you vastly different conclusions, and they are both
Question four was a difficult question. I elected to not look at privacy. I
could bring my security officer and my privacy people, but that is an ongoing
debate for us regularly. However, there are practical measures to help insure
accuracy and validity, a standard core measure set. I put in federal adoption
here; sometimes I see the world from the federal perspective. There doesn’t
need to be federal, but there does need to be agreement on what is the standard
for a measured data set and how it affects other people.
Precision of the definition of the components of the measure. As you may be
aware, we do a lot of work on comprehensive measures. Some of them require
searching in 20 different data fields. That is not the place to start. We need
to start small. The ability of EHRS to record comparable data that can be
identified, extracted and aggregated.
For us, our docs wanted apples to apples. They didn’t want that clinic down
the street to change the denominator. They had better data. As you know,
physicians are competitive. Even if CMS isn’t rewarding for results, the
physicians are tracking those results, so it is really important to insure that
there is comparability.
Finally, there are policy issues that need to be defined and addressed. The
technical issues continue to exist, but they are surmountable. We are a good
example is the little engine that could. We had a lot of stress on us to do
that, and we were capable of doing that with a fairly limited physical
And implementations should move slowly. I don’t mean implementation of
EHRs, I mean implementation of quality reporting. You have to give everybody a
first step and then let there be breathing room.
MR. REYNOLDS: We love presenters who answer our questions. Thank you. We
get excited about this. Dr. Halamka.
DR. HALAMKA: Thanks very much for the opportunity to speak to you about
quality reporting. I will be represented the Health Information Technology
Standards Panel, but also make a few comments in my role as the CEO of an
integrated health care delivery network and a chair of a health information
I think today so far, I have been listening to all the testimony, and there
have been some great common themes. The electronic health record and some
baseline functionality rolled out very broadly, health information exchanges
that have local variation because they are more about data use agreements than
they are about the technology; create databases where it is appropriate for
such things as population health and quality reporting, and make them as big
enough as is politically feasible. In my case, I think it is probably a city or
at most a clinical service area or region; and protect privacy.
So if that is the goal, and we have heard these things over and over, what
are some of the things that we need from a standards perspective to get there?
First I want to start with this statement, because I know you guys are
going to ask hard questions like, how would I define meaningful use in one
sentence or less. My sentence is, the processes and work flow to facilitate and
improve quality and increase efficiency. If we are implementing technology and
it is meaningfully used, we hope the quality and efficiency increase.
From a Health Information Technology Standards Panel perspective, this is
the road map that we have worked on for the last four years. If we are going to
achieve enhanced quality reporting, we know we are going to need some basic
data elements. All the panelists today have been very practical. They have said
it is absolutely great that in the future we would have common data models and
wonderful if we would have clinical document architecture everywhere. But alas,
we are dealing with some data quality issues that aren’t so great today. We
have a lot of anti-administrative data. We have a lot of old standards and
legacy systems. So we have to deal with that.
So what HITSP has tried to do is both be practical and forward looking. In
this road map you will try to see that we tried to cover the lab, for example.
Everyone has talked about the importance of the lab. Lab does include that
LOINC code, and a unit of quality and a reason for ordering, some things that
we think are going to be pretty important for measuring quality. But we know
that not every lab in the country is yet reporting LOINC codes, so we will get
there as a journey.
We also specifically focused on quality. As you will see in our 2007 work,
we worked with the National Quality Forum’s health information technology
expert panel which Paul Tang chaired and now continues on as the HITEC II; we
have a meeting next week to continue that work.
That was some really great work. How is quality defined in this country
today? Well, quality metrics organizations look at the best evidence and create
quality metrics that are academically sound. The only challenge is, it is not
necessarily the data we actually gather in EHRs, and much of the data we gather
is free text. So what you have to do is paper based chart abstraction that gets
you a quality measure six months after the care is delivered wrong. That is
probably not the way we want to go.
We probably want to ask the question, which is what Paul and the team did,
if you were to define quality metrics in light of the data we gather today, and
you did a Prader diagram and said what are the data elements that are going to
be most useful and then say okay, problems, meds, allergies, some labs; if we
do that we could get quite a lot of these quality measures done.
So specifically what HITEP did was to take these 13 disease states and
processes of care. These are the big ones for American health care, so you have
got asthma and cancer and diabetes and ischemic heart disease, mental health.
If you look at the AHA and HQA measures, you find 84 different metrics, but you
can break those down into 34 different data types. That is to say, if based on
this work we did in 2007-2008, you had 35 data types reported using a lot of
the existing standards we use today, you could hit a lot of these quality
We are going to continue to refine this. Over the next week we will be
getting I hope to an even tighter minimum quality data set.
What are some of the standards you might see? These are ones that you guys
have all been implementing and the HIE folks have been talking about. HL-7 2.X
for things like messaging of demographics and problems. NCPDP scripts, the XML
versions. There are some service oriented architectures here for the aspects of
medications, the CCD or continuum of care document architecture for summaries.
Also vocabularies, SNOMED-CT, Rx Norm, structured sig is still in pilot and
For allergies too. You want to look at fluids and substances, medications,
medication classes and there are standards for that.
There are reasons to gather text data. I know that text data isn’t perfect,
but it is in fact what we have. And there are departmental reports that are
also useful, and many of those are encoded in SNOMED-CT. Laboratory results as
Marc highlighted are very important, and HL-7 2.X with LOINC and some SNOMED,
You will notice in this list of data elements, I didn’t highlight a lot of
administrative data. Administrative data, it is true, what we have today, but
one caveat for you. You read the Boston Globe, and you will see that I have had
some fun over the last two weeks investigating the nature of administrative
In the interest of transparency, I share data with patients, both clinical
data and administrative data. And patients said, I see that ICD-9 isn’t
granular, is it? When I see murders, war, ICD-9 not granular enough, you know
we are in a new world. What the patients discovered was, you can’t really
measure their clinical experiences very well on administrative data sets.
So I think the answer is we use what we have got. If you have a choice, use
clinical data. Use the kinds of standards that I described, the labs, the
medications, the problems encoded in SNOMED-CT. Hopefully with our $19 billion
spent wisely you will get EHRs rolled out and more and more clinical data used
for quality measurement.
Just a quick comment on what HITSP is doing to try to help this process.
The use case process was a good construct to get us started as a country. But
now as we are looking at the stimulus bill, we recognize there is a new game.
It is not about a use case specifically, it is about meaningful use. And
meaningful use is all about data elements, certain aspects of functionality. So
HITSP is taking all its work that has been recognized and approved over the
course of the last four years and reformatting it so that it will be much
easier to access, much more streamlined, simpler, easy to maintain. So when you
say I want to gather problem list information, here is the one way to do it. I
want to gather labs and medications; here is the one way to do it, as opposed
to going through multiple use cases to pull out that information. We will have
that done by July 15. We have focused our work over the course of a 90-day
period to get that done.
A quick comment on regional health information exchange and quality. We
have recognized it is quite important to aggregate data in a community to
measure quality and get feedback to physicians. We do in our health information
exchange care about individual patients, so we are sending clinical summaries,
we are doing e-prescribing, but we are also using the data that is coming out
of our hospitals and clinics as a feeder for a regional quality warehouse.
Mickey Trepathi will be here tomorrow and will describe that in some
detail. What he has been able to do is create a set of data elements — and we
worked with New York on this as well — to get some of those high priority
process and outcome measures, use the HIE, use our EHRs, and feed a
de-identified population-based quality warehouse that we use for benchmarking.
That kind of thing has worked pretty well.
Again, I would just try to be realistic. I don’t think that we can create a
giant database in the basement of the White House that is used to measure
quality for everyone. There has got to be some federated nature to this. It may
very well be some federated nature to this, and it may very well be regional or
it could be domain based, Indian Health Service, VA. It could be hospitals that
are cancer care, pediatric hospitals. But there are going to be natural
organizing principles for the aggregation of quality data for reporting in a
way that is technically and politically feasible.
The last thing to close on is, I think that this journey is something we
can begin today. There are standards that are good enough. Yes, we have work to
do on the policies, but those are evolving. As you heard from some of the other
speakers, there are HIE constructs that are working today. There are examples
of quality warehousing that has been done in Massachusetts, New York, Indiana
that shows that all this is doable.
HITEP and the work that has been done over the last year and the work that
will be completed in the last couple of months will give you a quality data set
that is going to be sufficient for this purpose. The HITSP standards that we
first worked on over the last three years and are now reformatting specifically
for the American Recovery and Reinvestment Act will give you the standards you
need to achieve and transport an aggregation of that minimum quality data set.
MR. REYNOLDS: Thank you. Dr. Kmetik.
DR. KMETIK: Thank you, and thank you for your time. I know I am between you
and the break.
I am Karen Kmetik. I am Director of Clinical Performance Evaluation at the
American Medical Association. I am speaking to you this afternoon from three
perspectives, one, as a measure developer. The AMA convenes the Physician
Consortium for Performance Improvement or PCPI. We are the body that develops
the measures. We maintain those measures for the current evidence base. We
specify them using code sets available, et cetera. We are now specifying all of
those measures for integration into electronic health record systems.
The other perspective that I wear is, for two years now we have been coming
together, measure developers with EHR vendors and physician users, to begin to
say what are the issues still and how can we together as three communities work
to solve those.
The third perspective I have, which is the most relevant for today, is the
opportunity to participate in some real world incubator groups, I call them,
which right now have electronic health record systems and are trying to query
and report out on national performance measures, which I think is where we want
to be able to go.
A tremendous opportunity. We have all said that today. I think it is just
Others have talked about the fact that we are moving beyond claims based
measures. I think it is important that we don’t forget that. We don’t want to
designate measures for EHRS that are replicating what we were forced to do in a
claims based environment. So things that we can now do are not just things that
are under use, but I call it the if not-why not. If you did not prescribe the
ACE, why not? That is the information that is going to help us to improve care.
Others have talked about putting the measurement data quickly into the
hands of the physician. That is absolutely critical, and we have this
opportunity in our hands. Also being able to produce the measurement data for
100 percent of patients, not just a sample.
Obviously what we are talking about here on this panel is the quality
performance measures. The trajectory for this work in electronic health record
systems has to be in parallel with the trajectory for e-prescribing and the
trajectory for exchange of information. Obviously we need that information for
But my main point today is, today is the time to lay out the game plan.
Today is the day to lay out the game plan for the quality measures that we want
to have integrated into EHR, because as Theresa and others have said, it is
going to take time. We have got to let everybody know the game plan.
Just to bring it down to a real world example, if this is helpful in your
discussion. This project we call Cardio-HIT. We pulled together six practice
sites. It ranges from Dr. Houston’s site in Pittsburgh, which has hundreds of
providers, to a site in the suburbs of Illinois that has three physicians, a
three physician group. They all have electronic health record systems. They are
different specialties. They all came together to say, what would it take for us
to do this today.
So we laid out the game plan. That was very important, so that everybody
knows, the providers, the EHR vendors, where we want to get. Where we want to
get to is what you see in step four, using this data to close gaps in care,
increase efficiencies and enhance patient specific care.
So we had that as our sites, and we chose coronary artery disease and heart
failure. We go back up to the top and we say, we have got to start with a set
of national performance measures. They exist today for these conditions. We
then needed to identify what are all the data elements that we need to collect
the data for those measures. We can do that today.
Then we said, our next step is, as Theresa mentioned it, we want the
information on measurement to be available at the point of care, and we want
the practices to be able to report out in their practice how they are doing for
their whole population.
Third, we want to be able to export the data, and we also want to think
ahead in terms of, could this data eventually, if all practices in the U.S.
have EHRs and are exporting data on quality measures, that data should populate
I would suggest the AHRQ national reports.
So what do we want to see down the line in those reports? Sometimes I think
in discussions I am a part of, we skip over number two. We say, we want the
measures, we want the data elements, we want to export it. We want it first to
use in practice where it can have the greatest and the quickest impact.
So if I go through these rather quickly, first starting with a set of
measures. I suggest that we can do it today, granted, not for every specialty,
but for an awful lot of them, particularly the chronic care conditions. When we
went to select a set of measures, we looked at coronary disease and heart
failure. There are measures that we developed with the specialty societies.
They are NQF endorsed. We have test results available. They are in four CMS
programs, and we have a way to transmit the specifications to EHR vendors. What
we would like to continue to track now is, are these the types of measures that
could make their way into an AHRQ report? Are these the types of measures also
used by HRSA and other agencies? Now I will add the Indian Health Service to
this list, thank you. Then, are these the data elements that are being embedded
into CCHIT? So we started there.
Second, we needed all the data elements for those measures. We can identify
these. They are part of the measure specifications. We know we need ejection
fractions, we know we need lipid levels, we know we need allergies. For these
six practice sites who have electronic health record systems, they have been
using the systems for four years, each of the sites had to themselves, create a
discrete field and manually enter the ejection fraction. Fine, we know that. I
was thrilled to see on the CCHIT website there is a cardiovascular optional
module that has a discrete field for ejection fraction.
I say this with great passion, because we can do this. It starts with the
measures. Everybody accepts the measures. They will evolve, but let’s start.
What are the data elements, get them into CCHIT, then into the products, et
Moving along. The second piece, having those measures available at the
point of care, the measure results. In some ways it acts a little bit like
clinical decision support and having practice wide reports. The only sites that
we work with that can do this were those that had an IT expert on staff to help
them figure out how to produce the report. So I would suggest that this is an
area we want to focus on, we want to prioritize.
A third area about exporting the data. It is no surprise, none of those
sites, even though they have been using EHRs for at least four years, are using
a common data export file. We accept that. We enable them to send us the data
in a very sample way, but others are working on this, and I believe this will
We found an interesting note. Even if they were using codes in their
systems such as NDCs, there is a difference between the system using the coding
and then being able to easily export that coding. This is another thing for us
to think about when we get to this stage three.
But all the sites now are sending data to a warehouse on these national
measures. They are getting feedback reports at a website. One can begin if you
are a health services researcher like me, get a little bit excited bout how
this data could then find its way into national reports.
Just by way of a simple example, there has been some articles written about
exception reporting in measures. I call this the if not-why not. We want to
know this. We don’t want to hide this data. We want to know this.
We see published literature from the U.K. on exception reporting for
measures. We can look at exception reporting from the 2007 PQRI data. In our
project now with EHR data sent to a warehouse, we can reveal the exception
rates in this. So you can begin to imagine, if every physician practice is
reporting out an EHR, these data could populate the AHRQ reports.
Lastly, of course this fourth step which is most important, is being able
to use the information to improve care. This could happen once we have that
functionality and the data for the practices.
Let me just close by saying, I encourage us to lay out a game plan. We can
do it starting today. There are measures available. We know what the
specifications are for those measures in an EHR. We need to prioritize the
functionality to have the measures available at the point of care, have the
reports there before we start shipping it anywhere, I would suggest.
Lastly, I will just put in a brief plug for the fact that this is going to
take practice. We are not going to get it right even if we have everything
coded, everything in a discrete field. Questions come up; which blood pressure
did you mean, I’ve got three blood pressures recordings in there. We can solve
that. We need to continue these forums where we bring together the physician
users, the hospital users, the vendors and the measure developers. But as you
can tell from my passion, I think we can do it.
MR. REYNOLDS: Thank you. The panel has passion, standards and a road map.
What more could we ask for in mid-afternoon? With that, I’d like to open it for
DR. BLUMENTHAL: Thank you very much for that very stimulating conversation.
I would be interested in Dr. Cullen’s response to, you were much less
ambitious, it seems to me, in what you thought was possible than the people on
either side of you. I wondered if there is more agreement than I perceived.
DR. CULLEN: What I think is possible is looking at an 18-month trajectory
to have something in place by 2011 that will encourage implementation and use
of electronic health records and not freeze people out and discourage them.
My recommendation would be that you start slowly with things that are not
dependent upon other people giving you information at their point of care.
Vital signs are pretty easy to get. If you are not getting them, there is
probably something wrong. Weight, pretty easy to get.
So I think that there are some metrics. Even though we know for instance
there is lab data, the issue about LOINC transmission or LOINC aggregation of
that data makes it very difficult in our site.
So my response to you is based on ten years of having done it slowly, and
now we are on a rapid logarithmic increase, and we can add 20 measures a year
that people may or may not use. But I think the overarching goal is clearly to
get — the first step is to get the electronic health record used in a way that
people will not feel are intimidating to them, and/or they worry about the one
thing that was in this about data quality, which I think is a huge issue.
I will put a plug in. Regional training centers may encourage rapid
escalation of that if they can get in place very soon.
DR. HALAMKA: And just to comment, of course there will be gaps. Every data
element has a cost, the cost to make sure it gets entered, the cost of
reviewing the quality of that data element.
I know that as we roll out these measures there will be a lot of scrutiny,
and that the physician will say, did you really risk adjust that, is that data
accurate. Wait a minute, all my patients are sicker, that can’t possibly be
true. So we will be forced as you said to re-look at everything we are
submitting, make sure the data quality is as good as it can be.
We will have this problem that the physicians have 12 minutes to see the
patient and get the data in and get it accurate. So how do we balance the needs
of the physician, making them more efficient and their work flow with the need
to capture good data in a structured form. So there will be a journey.
MS. ANDERSON: I have a little bit of a different view, in that I don’t
think that the measure itself is the most important first question. I think it
is the role of the electronic health record, and you can layer in —
measurement has to occur, and you can layer in pieces for each measure that
comes on board.
To be more explicit about that, dealing with the issue of how and when do
you capture discrete data, what data needs to be captured in a discrete way in
classes of data, not necessarily always down to the element level.
We also have a fundamental problem of not getting data out of electronic
health records. Setting as a requirement that the data must come in, it must be
able to go out, so we can enable health information exchange.
The fact that we need to be able to create panels of patients. For anyone
to manage a group of patients, how can you create that, and an electronic
health record should support that function, even if we layer in over time
different types of panels that are related to different denominator statements
that go to measures, or the ability to do clinical decision support. It may not
be every practice does the same kind of clinical decision support, but there is
a piece of that that is relevant.
So this is the role we see for the electronic health record. This is the
role we see for health information exchange. This is the role we see for web
services. This is the role we see for public reporting. Then you layer in the
measures as maturity comes. Some folks might still have to have an electronic
health record and do chart review. Sorry, but that might be a reality for a
little while until it all syncs up.
DR. KMETIK: I would just echo and say, it is about no surprises, about
saying to everyone now, this is what we are looking forward to, and this is how
quality measures can help you in your practice. You would have this data
available to you
I guess my main message is, just lay out that plan, because if we wait,
then it might not be implemented.
DR. OVERHAGE: Thank you very much for covering the waterfront so well with
your presentations. Dr. Cullen started from an ER perspective, and then Dr.
Kmetik commented on moving beyond claims.
As my state Medicaid director told me last week, certainly Medicaid, their
managed care organizations are based on HEDIS measures because that is what CMS
measures them on, and therefore that is all they are interested in.
I would be interested in the group’s perspectives on how do we — there is
one set of things in the standards and the data, and a parallel set of
challenges is, how do we move this critical mass that Gina Perez talked about
before for quality measures, from a focus on claims to a focus on clinical?
DR. KMETIK: I’ll start. I probably sound like a broken record, but it comes
back to laying out the staging of it. I will just use the CMS PQRI program as
an example. The program started as a claims based program, and then there was
added the option of reporting via registry, and an option under study is
reporting via an EHR.
I thought that was a good example of saying we would like to get out there.
We know we are here, but everybody knows you can do it with that system if you
have it in the future.
DR. CULLEN: I would agree. I think having a stated trajectory that has no
surprises is critical. When discharge for CHS came out with ejection fractions,
we didn’t have ejection fraction in our system. So we got a field in right
away. We control our system, we can do it. It took two years to get the field
populated enough that we felt the data was okay, because somebody had to come
into the hospital, he had to put the ejection fraction in. Now it is a standard
part of our set.
The one other thing that I think is really critical, and we didn’t talk
about this, is, we are talking about quality from the provider perspective. In
many places, care is a group practice or a group. It is important to recognize
that when you move from claims data, you want to recognize that fact too, that
there is quality for a group practice and how critical that is. I think that is
dependent on clinical data.
I would say starting in 2011 we say the data is clinical.
DR. HALAMKA: Changing behavior depends upon aligning incentives. You have
to figure the incentives of each stakeholder. So in the case of clinicians, we
have said first, condition of practice. You must have a problem list, a
medication list, a reconciliation and e-prescribing, and you must clinically
chart, and that is a condition of practice in our bylaws. And by the way, all
your pay for performance measures are going to be based on this data, so you
will get paid whether you fill in the data accurately or not, and other members
of the ecosystem.
If there was regulation that defined meaningful use as a certain kind of
quality measure, then you would expect that CMS would change. So carrots and
sticks and aligning incentives.
DR. CARR: Thanks for great, great stimulating and inspiring presentations.
We heard earlier today that one of the requisites of anything funded should be
that the patient access — if we move to our patient centric and move away from
our clinician centric model, where does EHR or patient functional status fit
into all of this?
MS. ANDERSON: That is part of why aggregation and quality measurement
aren’t necessarily EHR centric functions, because of this issue that a practice
based EHR is likely to be one ultimate data point around quality. That personal
health record information both should be an input that could go into an EHR or
an input that could go into a quality metric, or both ideally. We need to plan
for that to be happening over time, but without expecting that it is
necessarily going to happen right out of the gate, relative to how folks have
the ability to get that information in place.
The same with any type of functional status or patient satisfaction type
data, or patient exposure type data. It is imagining that the longitudinal
record of the patient is not necessarily something that lives within anybody’s
DR. HALAMKA: I think of PHRs as having two important roles of quality. The
first is, the patient can be the health information exchange. We all agree that
there are issues of state laws preempting HIPAA, and everyone has talked about
today the issues of policy and privacy. If the patient is the agent, the
patient is the hub and all of us give data to the patient, if the patient
elects to share that data as appropriate for clinical care clinical research,
et cetera, that solves a problem for us; care coordination and quality improve.
The other aspect is measurable quality based on data provided by the
patient. Many of the new PHRs are allowing home glucometers and blood pressure
cuffs and maybe subjective migraine scores, other things, to be reported via
Now, the only challenge is, I don’t think from my institution’s perspective
we are yet mining such data, because we are concerned about its quality and its
sourcing. But I think in the future, a PHR may be a very novel and important
source of quality data.
DR. CULLEN: I would like to remind people that there is still a digital
divide in this country, and hopefully the ARRA funds for broadband will help
The one thing that we have done is look at what is appropriate technology.
We give patients a patient wellness summary. It is paper; we are moving to an
electronic one, but we have done that for years.
But I think this whole concept of patient experience of care as a quality
measure is critical. We are just now testing that in a fairly agile software
development. The problem is, where do you stash that data and then what do you
compare it to. I think that is a whole field that we need to go into.
In fact, incentive dollars are really helpful to have people measure
patient experience of care. The issue is, because it is so inconsistently
measured right now and it is not tracked longitudinally other than survey data,
it is a difficult quality measure to get at. But I think we should not forget
that some measures need to be from the patient’s perspective.
MR. REYNOLDS: I’ll ask the last question. As we go forth and you all
recommended some exciting opportunities, we talked about tools earlier. The EHR
is a tool, the HIE is a tool, quality appears to be little more than a tool. It
appears to be possibly even a destination or something.
So if it is probably the quickest thing that a small doc and others would
understand, how do you package that quickly? How do you put these things
together that we have seen in all the slides? So as you first go out and start
talking about bringing the ones that don’t have right now along, how do we use
that in a most appropriate way, and how does that happen?
DR. HALAMKA: Regional health care IT extension centers, we hope, will
implement EHRs in a consistent fashion that makes the process and work flow
both quality and efficiency oriented. One would hope that as these EHRs get
rolled out by these SWAT teams of local experts, that it will be a requirement
for that practice to submit quality data as a byproduct of their use of the
EHR. So intelligent rollout plus requirements to gather the data.
One of the things that you said was quite important, is feedback to the
doctor, that quality data, pretty rapidly. We have found as long as you work
with the clinician, you give them that feedback, you give them benchmarks, they
find it pretty useful. Now they understand why they are gathering it and the
benefit to gathering it. So with good rollouts, SWAT teams and feedback from
the doctors rapidly, we should get it done.
DR. KMETIK: I would just add, in putting it down as to step by step, what
will be the benefit then. The benefit is, you will know of all your patients
with heart failure who is not on ACE R and why. You will know when you are
seeing that patient with coronary disease, the reminder to check, did they have
a prior MI and should they be considered a beta blocker.
That is what you will have available. To get there is going to take time.
The first step, the problem list. Second step, everything in discrete fields
and coded. But to say at the beginning, it is not just about having the result,
it is about having that information so you can care for patients in this way.
MR. REYNOLDS: With that, we will thank the panel. We are five minutes
early, so you have until 3:15. Then we will be back together and proceed
through our last two panels. Thank you very much to all of you. We appreciate
MR. REYNOLDS: I’d like to welcome everyone back. We are in Panel 5, which
is the use for 2011 and beyond for vendors. We have got Justin Barnes, Glen
Tullman, Dr. Mark Segal and Fred Trotter. So we will go in that order unless
anybody has any concerns about that. Justin, if you would please proceed.
MR. BARNES: Thank you very much, Chairman Reynolds, Dr. Blumenthal, and
distinguished members of the Executive Subcommittee and staff. I am going to be
giving a formal testimony today, so no slides, but I think you will enjoy it.
My name is Justin Barnes. I am a vice president with Greenway Medical
Technologies, a leading provider of integrated electronic health record
practice management interoperability solutions.
In addition to my work at Greenway, I am also Chairman of the HINS
Electronic Record Association, the EHR associations comprised of the nation’s
leading 40 EHR companies, currently representing over 90 percent of all EHRs
implemented in America today.
It is always a great honor and pleasure to work with members of the
Administration and its advisors, as I believe you all have a common goal to
shape the new face of the health care industry by utilizing the vast
contributions that information technology offers providers, patients, payers in
achieving goals of reduced medical errors, lower costs, better quality and
improved efficiency within the nation’s health care system.
My perspective today will be a balance between expertise and experience
with a broad range of software companies that are in the ambulatory and
hospital market. Greenway as well as the EHR Association has been focused on
advancing standards based interoperability, standard based quality measurement,
clinical decision support, computerized physician order entry and encouraging
effective custom implementation of comprehensive EHRs.
Based on this expertise and commitment, my colleagues and I stand ready to
work with HHS, NCVHS and others to further develop and refine the definitions
and implementation of meaningful use. In implementing the meaningful use
provision, we suggest the following goals: Improve the health and health care
of Americans, improve the health care delivery system, support health care
reform, improve quality and patient safety while reducing costs, achieve a
critical mass of physicians and hospitals using interoperable comprehensive
EHRs, target high costs, high morbidity of chronic disease, counter identified
barriers to adoption, and support collection of data that support
biosurveillance and public health.
Beyond these broad goals, we recommend the Executive Subcommittee operate
from a few guiding principles: Following closely the structure and priorities
for the meaningful user established in the American Recovery and Reinvestment
Act, ARRA. A key objective of this provision should be to insure that care
providers and hospitals are using the EHR, especially for the priority function
specified by ARRA. To maximize adoption and insure administrative simplicity,
the number of criteria should be small in number while consistent with the
statute and overall goals. The criteria should be simple, and ease of reporting
should be primary. In the hospital’s case, it should reflect the differences
between hospital departments.
The primary initial goal should be to incentivize as much point of care
adoption and use as possible of comprehensive, interoperable and certified
EHRs. It is particularly important to understand hospitals have relatively long
implementation and update cycles, and in addition face substantial departmental
complexity. Therefore it is essential to balance the need for accelerated
adoption of EHRs with the need for care providers and hospitals to implement
these in a careful non-disruptive manner.
To this end, it is recommended that the Secretary set meaningful user
criteria for the initial years, especially 2011 and 2012, at achievable levels
with a road map for steady uplift over time in expected breadth and depth of
use. Such uplift should be on a less than annual cycle, perhaps 24 months, to
allow for predictability and effective provider adoption. Certification
criteria and product functionality should however anticipate and support
projected increased levels of meaningful use. Thus, while we support an
initially relatively simple approach to meaningful use from the beginning,
providers should be encouraged to adopt EHRs that contain high levels of
While initial meaningful use needs to allow for the currency of EHR
deployment and implementation, we emphasize that to insure national goals be
met, the features and functionality of certified EHRs must be comprehensive and
highly functional from the beginning so as to support the rapid widespread
transition of physicians’ practices and hospitals to high levels of meaningful
Current CCHIT certification provides the fundamental features and
functionality for comprehensive EHRs that are available in the market today.
Future certification can and will evolve over time, and should be based on a
24-month cycle to insure widespread product availability. In implementing
quality reporting provisions, we urge HHS to build on applicable current
reporting programs and developing criteria that can be created well before the
end of 2009, and that can be adopted by providers using certified EHRs.
From the eligible professional and ambulatory environment we must learn
from the successes and failures of previous programs and seek non-intrusive and
low cost reporting options such as using Medicare Part D data to track
e-prescribing, using EHR data as a source for quality reporting, using surveys
and/or attestation in lieu of claims based reporting for the dimensions of
meaningful use. For the hospital environment, we must also learn from the
successes and failures of these programs and seek low cost and non-intrusive
reporting options such as reporting measures that are a byproduct of meaningful
use of the EHR and/or using surveys and/or attestation in lieu of claims based
reporting for the dimensions of meaningful use.
Meaningful use criteria should support the inventory standards based
interoperability, and interoperability from the beginning should only be
measured using HITSP harmonized standards. Specific eligible professional
proposals in the area of meaningful use, proposed criteria using certified,
qualified EHR for encounter documentation in a way that supports quality
reporting through the creation of structured data sufficient for such
reporting. The use of electronic prescribing, the use of clinical decision
support, sufficient discrete data capture to support interoperability.
In the area of information exchange, the overall requirement would be for
connection for exchange of clinical summary data using HITSP harmonized
standards with other clinicians, hospitals, patients in other health care
settings. We believe that eligible professionals should be required to satisfy
an initial level of standards based data exchange to meet this provision in
2011, with increased requirements in out years. I have included suggested
criteria examples in my written submission.
In the area of reporting measures, consistent with ARRA, the focus of this
criterion should be primarily on a subset of existing National Quality
Foundation endorsed measures that align with national quality and performance
goals. There should be a priority for measures that support chronic disease
management such as diabetes, congestive heart failure, hyperlipidemia and
hypertension, as well as measures applicable to the broad range of specialties
and provisions who will be reporting.
For transport of quality measure information, we favor the use of HITSP
standards as available. Specific hospital proposals in the area of meaningful
use, clinician view access to hospital clinical data, use of CPOE at a
reasonable percentage for medications, with broader use of CPOE in out years,
use of electronic medication administration at a reasonable percentage with
barcoding, and more in the out years, use of clinical decision support to
improve medication safety, sufficient discrete data capture to support
interoperability and quality reporting.
In the area of information exchange, the overall requirement would be for
connection for exchange of clinical summary data using HITSP harmonized
standards with other clinicians, hospitals, patients or other health care
settings. We believe that hospitals should be required to satisfy a limited
criterion to meet this provision in 2011, increased requirements in out years.
I have included suggested criterion examples with my written submission.
In the area of reporting of measures, consistent with ARRA, the focus of
this criterion should be primarily on a subset of existing NQF endorsed
measures that align with national quality and performance goals. Transport of
quality information, we favor the use of HITSP standards as available.
To meet the anticipated increased rate of certified EHR adoption starting
in 2010, vendors recognize the need to substantially increase their
implementation capacity. To that end, we need extensive education of health
care providers and clarity in regulations and interpretations to avoid market
confusion and to support the vendor investment and quality training in hiring
In 2016 we expect ARRA to result in widespread adoption and use of
innovative and comprehensive EHRs that reflect substantial advancements in the
areas of interoperability, decision support, quality reporting and clinical
research. These advances in functionality and usability drawn on the extensive
expertise of our industry and customer driven product development will lead to
improved health care quality and outcomes and hopefully, significantly reduce
growth in our nation’s health care costs.
We will continue to work together in public and private collaboration,
build on present successes with CCHIT, HITSP and NQF, take the prudent and
fiscally responsible steps necessary. We will achieve our health care
transformation goals and recognize and estimated $100 to $200 billion of annual
savings that will come from fully integrating the interoperable health care
We are ready as an industry to continue our progress and partnership with
Congress and the Administration in making these goals a reality.
Chairman Reynolds, Dr. Blumenthal and distinguished members of the
Executive Subcommittee and staff, I want to thank you for this opportunity and
your dedication to meaningful EHR use, interoperability and the evolution of
the health care information enterprise. I hope that comments will help steer
ideas and thoughts that can be transmitted into innovative policy shift in the
future health care for this country.
Thank you very much.
MR. REYNOLDS: Thank you. Mr. Tullman.
MR. TULLMAN: Chairman Reynolds, Dr. Blumenthal and members of the Executive
Subcommittee, I want to begin by thanking all of you for the opportunity to
speak today as a part of the hearing on meaningful use. My name is Glen
Tullman, and I serve as Chief Executive Officer of Allscripts, which is the
largest publicly traded electronic health record company in the United States.
I am also pleased to serve on the board of trustees of CCHIT and as co-chair of
the National Electronic Safety Initiative, which is a program in cooperation
with Dell and Microsoft where we give free electronic prescribing to physicians
across the United States. I am pleased to say that that program launched some
two years ago now has some five million transactions a year running through
But today I am also representing some portion of the 160,000 physicians who
are our customers, who use evaluation records, electronic prescribing, practice
management systems and our connectivity solutions. To these clients who span
the entire breadth of the health care equation, they range from some of the
largest academic medical centers in integrated delivery networks in the
country, folks like Columbia University, Thomas Jefferson, Iowa Health System,
an advocate, and George Washington Medical Faculty Associates, only a few
blocks from here, to midsized multi specialty practices, Leahy Clinic, Health
Care Partners, Mt. Cisco Community Care, to very small groups, Stanford
Pediatrics, Wilmington Health Associates and Oliver Family Medicine.
I reference that in part because of two reasons. One, we are very focused
on insuring that all of the meaningful use requirements are focused across
America, from the largest to the smallest. I also mention that because in that
smallest category, out of our 160,000 customers, 90,000 of them are in the
category of less than ten physicians, and those 90,000 don’t have an electronic
health record. So it is very critical to us that they go through.
If you look at one of the challenges that we have as a group, what you will
find is — and I had a wonderful chart that showed that in the 513,000
physicians that are measured by outside research organizations, about 300,000
of those range in the less than ten physician category. Our challenge is to, as
Dr. Middleton said this morning, be practical in terms of what the meaningful
use requirements are, but at the same time to make sure that we don’t set our
sights or our goals too low.
As we look at that, we look at the goals of ARRA, of the American Recovery
and Reinvestment Act, and what we see is that they fall into three categories.
Category one is to focus on two things as we see it, that is, safety and
quality. The second piece of that is to eliminate waste and fraud and abuse and
add efficiency. Last but not least is to insure that as we move to electronic
health records that it ultimately allows us to use those not as an end but a
means to getting those physicians better information.
So the question is, how do we do that, how do we get that leadership and
those standards. If there is one thought that I would like to have all of you
agree with today, it is this quote: The future is already here, it is just not
evenly distributed. I think if you look across our base, what you would see is
that health care organizations of all sizes today are already using electronic
health records and e-prescribing to provide safety, more cost effective care.
This is doable now. It is here today. Our view as an organization is to start
with what we know works, bring on acceptable standards, continue to build on
those with an aggressive road map, but as one of the earlier speakers said,
make certain that we telegraph everyone where we are going, what direction we
are headed in.
So if we take an example of what is working today, electronic prescribing,
this year I am very proud to say that physicians using Allscripts software will
write about 80 million electronic prescriptions. One of the things that we have
seen is, since the introduction of the CMS electronic prescribing stimulus, the
roughly $5,000 that a physician can earn, since the introduction of that, we
have seen a 30 percent per month growth in our prescribing base. We are up
about 300 percent from last year. So again, what you see is that even a small
amount of stimulus money can have a very major impact. We see that as a
harbinger of what can come with the stimulus package that we have here.
We also look at electronic prescribing and we see that as an on ramp to the
full electronic health record. As we talk about meaningful use standards, one
of the things we say very clearly is that we believe electronic prescribing
should not only be a part of electronic health records, but in fact it should
be a requirement to make sure that we move forward aggressively on that.
I wanted to address a few questions that the committee asked us to face.
One was the speed of standards integration. That is to say, to the extent we
come up with new meaningful use standards, how quickly can vendors like some of
my counterparts up here and Allscripts address those standards.
The answer is that we have been addressing those standards through CCHIT,
which we are very supportive of. We would like to see us build on the CCHIT
standards, and we believe that those standards should focus on what works. In
other words, tell us what you want the physician to be able to do, not how to
make it happen. That is very important, because we want to make sure that all
of the vendors are free to innovate in ways that we see fit.
Similarly again, we think that the current standards are a solid
foundation. We don’t think that they go far enough, but I can tell you, as a
trustee of CCHIT that there is a road map in place that substantially raises
the bar. All the vendors are ready and willing to go forward along those lines.
Similarly there was a question on bandwidth sufficiency. This was the
question of, do the vendors have enough bandwidth to cope with the demand that
we see coming, and are they able to adjust to the coming standards.
What I would tell you is that at Allscripts we are dramatically increasing
our R&D investments. We have introduced software as a service product,
because a lot of people have said we want easier to install products that are
out there, easier to learn, easier to install. Those have already come to
market with us as they have with other vendors. We created a distribution
network to allow for more rapid distribution. Last but not least, we have had
some creative approaches like paper-free Tampa, where in association with the
community we are hiring over 100 what we call information ambassadors to visit
over 8,000 offices to talk about not only electronic prescribing but a full
electronic health record.
Last but not least, industry consolidation was a question we were asked
about. Allscripts has spent about two billion dollars in the past three years
in consolidating the industry and helping, most recently six months ago with
our merger with MISIS Health Care. We expect that is going to continue to
happen. Why? Simple reason; this is going to cost more money to do what all of
you are asking industry to do. So we need a lot of base to distribute those
R&D costs across.
Let me conclude with some recommendations. We believe that all of the
systems should offer the ability to easily import and export patient data in a
discrete fashion. We are strong supporters of registries. We believe that we
should work through organizations like SureScript to make electronic
prescribings routable, transferable.
Last but not least, we think it is important to insure that we don’t leave
the patient out of this new health care system we are building. So we would
require all EHR vendors to have connectivity with the personal health record.
Finally, I would just conclude by saying we have the leadership, the
software, the technology. We have best demonstrated practices, and with your
help, we believe that while the standards must evolve, we think we have the
ability to allow us to get to a world class health care system and allow all of
our citizens to connect to health.
Thank you very much.
MR. REYNOLDS: Thank you. Usually the vendors are pretty resourceful. If you
notice, they have built a water fort around the indicator up there as to how
they are doing.
MR. TULLMAN: I was close.
MR. REYNOLDS: So in the spirit of this hearing, we are instituting
standards. The meaningful use is that we would like to see that. However, we
also have another copy as a backup, so you can’t hurt us.
With that, I’d like to call Dr. Segal, please.
DR. SEGAL: Thank you very much. Good afternoon, Mr. Reynolds, Dr.
Blumenthal, members of the Executive Subcommittee, staff, and folks in
attendance today. I am Mark Segal, Director of Government and Industry Affairs
at GE Health Care IT. On behalf of GE, I am pleased to have this opportunity to
discuss meaningful use of health information technology and particularly
electronic health records.
GE Health Care is a leading vendor of health care IT, including electronic
health records for a range of medical practices of all sizes, for hospitals and
for integrated delivery networks. We have been a strong support of the HITEC
Act, and this comprehensive focus on infrastructure, on incentives and on
standards will accelerate adoption and meaningful use of interoperatable EHRs,
laying the foundation for health care reform. Along with leading policy
experts, with Congress and with the Administration, we believe that the outcome
will be a great leap forward in health care quality and efficiency.
I am especially pleased to be testifying today with such distinguished
colleagues, and notably with the Chair of the EHR Association, of which we are
a founding member, and whose positions on meaningful use we are very
As you know, the requirement that providers be meaningful users of
certified qualified EHRs has three main components. First, a general category
of demonstration of meaningful use. Second, standards based data exchange, and
third, quality and performance reporting. Together these will help our nation
see a real return from the investment in health care IT.
We suggest that the Department consider several points, reflecting our
perspective as an experienced supplier of EHRs for a wide range of provider
organizations. First, policies should be carefully tied to the actual
legislative language for meaningful use, and mindful of the distinction between
functionality, driven in part by certification, and meaningful use, what the
provider does with that functionality.
Second, meaningful use and related policies should encourage adoption of
comprehensive EHRs such as those that would meet current CCHIT requirements,
meaning widespread adoption first and foremost in order to get to meaningful
use and its benefits. Policies must be simple, issued with considerable advance
notice, not impose excessive costs on providers, and create a predictable path,
something we have heard a lot about today, for providers and for vendors. We
should avoid the temptation to create too many requirements and stay focused on
the primary goal, to assure that providers have and are using interoperable,
Finally, we need to insure that clinicians and hospitals can implement EHRs
in a deliberate and non-disruptive way.
These principles in consideration of your questions lead to several policy
recommendations. First, there must be a careful transition from standards to
certification criteria to products, and then into initial and ultimately
meaningful use. Not every element of every standard meets a current market need
for vendor or provider capability. Thus, CCHIT has considered market readiness
in its decisions on where to place a standard on its road map. This focus on
market readiness should be continued.
To speed time to market, we urge as much reliance as possible on the mature
standards already harmonized by HITSP and the certification criteria that have
been through the CCHIT requirements development process. We also must bear in
mind that implementation and upgrade times are generally much longer for
hospitals and for medical groups, often up to 18 months or more across our
industry. The timing for medical groups is often in the three to 12 month
For 2011 and 2012, we urge that meaningful use start at readily achievable
levels but with robust early requirements for HIE and quality reporting to
produce the data that are needed for informed progress in meaningful use levels
over time. Consistent with HITEC, we envision a steady growth over time in
meaningful use requirements. Increases should stair step upward with each level
and use for two years. Because providers need time and careful change
management to integrate this functionality into their work flow, they should
via certification be encourage from the start to adopt EHRs with a high degree
of functionality and a high degree of interoperability that anticipates the
future growth in meaningful use requirements.
In contrast to the likely pace of change in meaningful use requirements, we
envision a relatively small growth rate, a shallow slope, for EHR certification
requirements relative to an initial robust HITEC certification baseline.
The committee asked us how vendors such as GE will adopt our product
development and upgrade cycles to synchronize with increasing requirements for
meaningful use, information exchange and quality reporting. As a business we
are laser focused on helping our customers qualify as meaningful users, and
they are making sure we stay focused on that.
Even before the stimulus, we have oriented ourselves around standards road
maps and the CCHIT certification cycle in road maps. This effort has been aided
by our active involvement in HIT standards organizations, HITSP, CCHIT and the
This focus can however divert resources from other customer sought
enhancements. Hence, we support a two year period between certification
requirement uptakes. We believe that there should be at least six months
advance notice following an open development process before the certification
process begins for a new set of requirements. Finally, we believe that
certification should remain valid for at least two years.
The other panels have addressed the specific elements of the definitions of
meaningful use, but I would like to provide a few summary recommendations.
First, for the general meaningful use category, whose components will surely
vary between hospitals and health care professionals, potential criteria
include use of the EHR for documentation at a level that can support quality
reporting and interoperability, advanced e-prescribing for health care
professionals, and use of CPOE in hospitals, and point of care work flow
integrated clinical decision support. Again, this general criterion should be
less stringent initially and more stringent in later years.
We favor an emphasis on non-intrusive reporting, appropriate to the type of
meaningful use in question, including surveys and at a station where
appropriate, and support for registry based reporting such as being used
successfully for PQRI.
Second, for information exchange, GE is a very strong proponent of
standards based interoperability, and has been an active participant in the
applicable standards organizations. As part of the path to expanded standards
based interoperability, we believe that HITSP harmonized standards should come
2011 be the basis for the requirement of standards based electronic exchange of
Also, the initial focus should be on information exchange that can improve
quality and not on the specific parties to which a provider is connected.
Finally, 2011 requirements should be substantial enough to start useful data
moving from its current silos via standards based HIE.
Third, on quality and performance reporting, where critical standards are
now coming together, we urge use of a subset of the NQF endorsed measures being
used in Medicare quality reporting, focusing on those that can be derived from
data elements within EHRs, and that align with our national priorities.
HITSP harmonized standards as available should be used to define how the
quality information will be submitted. For maximum flexibility, submissions
should allow either patient level data or population level computed measures.
Similarly, EHR derived quality data should be able to be sent either directly
from the EHR or via a registry or other intermediary or reporting application,
and consistent with the national objectives for HITEC, quality reporting should
focus on all patients, not just those for Medicare or Medicaid.
We believe that quality reporting should begin with a fairly deep level of
required usage. In particular, care delivery organizations should do more than
just report externally on quality data, however valuable and challenge that may
be. Meaningful use in this area should include provider access to and active
use of evidence base quality data and measures to track and improve in priority
clinical areas. We need this closed feedback loop.
So in conclusion, HITEC will transform our industry and the health care
system. We expect a substantial increase in hospital and professional adoption
of comprehensive EHRs, with most of the impact occurring over the next five
years. We also expect changes in the structure of the industry and how our
products are developed, priced and deployed.
GE is reviewing key aspects of our business to meet this increased demand,
including how we develop and deploy software and how we support implementations
and upgrades. We are delighted with the new education and training funding that
was in HITEC, as well as the work that is going to be able to be done by the
HIT extension program, to assist providers in acquiring and implementing EHRs
in a cost effective manner.
So in conclusion, we look forward to working closely and cooperatively with
the Department, with this committee, with Dr. Blumenthal and his colleagues,
with CMS and with the many other agencies that will contribute to the
development and implementation of policy for HITEC.
Thank you very much for your time this afternoon.
MR. REYNOLDS: Thank you. Mr. Trotter.
MR. TROTTER: Hello, good afternoon. My name is Fred Trotter, which brings
us to the end of my prepared remarks. In all honesty, I am going to be changing
what I had about to say substantially, because when you represent a community,
which is what I am trying to do, sometimes they threaten to fire you if you
don’t substantially change what you are going to say, and then you have to do
it. Otherwise you lose that privilege.
First, I would like to define what community I am representing. I represent
free and open source software vendors and community at the same time. That is
very important, because a point that I would like to make is that for free and
open source software you have to think of the community and the vendor. You
can’t think of just one. A lot of the stuff that is provided by a proprietary
vendor is actually provided by a community and not a particular open source
vendor at all. So what I am going to be talking about is going to be somewhat
First of all, I don’t have a PowerPoint presentation, of course. That is
because it is a Microsoft product, and how would that look, really? I’m not
going to use Open Office. Instead I am going to fall back to an open source
standard. We are going to use mental PowerPoint.
This is going to be pretty quick because I have only got one slide to show
you. I want you to imagine in your mental PowerPoint a slide that says, open
source works, proprietary doesn’t. That is my point, that is my slide. So
everything I am going to say falls from that.
Now, to back that up, I am going to ask one single question: Why are we
here? Why is it necessary to fund an industry that is a high tech, high science
industry to computerize? There was a gentleman in here who was testifying
earlier who said he was doing meaningful use in 1972. If the auto industry was
doing meaningful use in 1972, what are they doing now? They are about to send
the Ford Mustang into space or something.
So why is this industry here? We are here because we have had a massive,
catastrophic market failure. Some people have said in this testimony that
doctors’ buying is a part of that, and that is probably true. That is 50
percent of the market, the buyers. The sellers have also failed. The sellers
have failed to provide a compelling option that the doctors can buy.
1972 we had meaningful use. What is the one substantial exception to
meaningful use today in the world of EHRs? It is VA Vista. I understand that a
very qualified researcher recently raised evidence that about 50 percent of the
very complicated EHR systems that are in deployment are Vista systems. You can
take it on good authority that that is the case. So only Vista has been widely
adopted. Only Vista has been doing all of the quality stuff that we want to
talk about. All the problems that we are talking about, quality, getting
movement of data, all the VA hospitals all talk to each other, and they have
been doing it for a long time. So if you factor in the size and the time, Vista
— and what I hope to show, the Vista Spirit, which is the open source Spirit
— is the Microsoft of this industry. We are the ones who have been
substantially solving the problems in this industry.
So I hope I have done enough to catch your attention here. Now I am going
to move on.
A lot of the complexity that you are facing right now is these delicate
balances. Do we fund the HIE or do we fund the EHR? Do we set the bar high? do
we set the bar low? I want to point out that many of these problems are based
in a mindset that is a black box mindset.
Let me give you an example. Let’s imagine that our goal today was to get
doctors into automobiles. The proprietary strategy is basically the same as
saying we should buy doctors a limo with a driver, and the driver has a
biometric thumb based thing so only he can start and drive the car. The doctor
can’t drive the car, only the vendor can control what the car does.
So now we are asking all these questions, what do we do if it goes left
instead of right, what do we do if it goes too fast? The answer to many of
these questions can be, don’t put the vendor in the driver’s seat.
The whole point of open source software is that the customers are equal
owners with the vendors and the rest of the community. So this notion of
control and who has it actually creates a lot of the problems that you are
trying to address. I would like you to keep in mind that some of these things
just go away with open source.
To continue, one of the examples that I have been asked to give is the
example of alchemy. Alchemy is where you have this culture of people working
for these kings, and each alchemist said to the king, I am providing you with
serious value. I am your guy, I am your man.
Now, every alchemist had to figure out on his own that mercury was
poisonous, which kind of sucked for him. Then we made a shift to chemistry,
which is essentially openness. It is the same information, but we are now going
to open up. We are going to have shared data. This is exactly the same decision
we are making in terms of electronic health records now. Open, closed, trapped,
free. It is the same basic conception. That wasn’t my idea, that was Dr.
The other summary from Dr. Valdez that I love to quote is that doctors have
no idea what they want, and programmers have been giving it to them for years.
Now I want to take a second and I want to zoom out a little bit. You have
asked me to say 2011 and beyond. I would like to take you way beyond. I’d like
you to zoom out.
My grandmother took a drug while she was pregnant with my mother, and that
drug predisposed my mother to ovarian cancer. My mother had ovarian caner and
died from it. I carry my mother’s genes. I will pass those genes on to my
daughters and my granddaughters.
Now, did my mother get ovarian cancer because of her genetic predisposition
to it, which my granddaughters will share? Or was it something that happened in
nature? Was it this drug? If you do the math on this, that means the contents
of my grandmother’s record are relevant for about 150 years. That means that if
we say a proprietary vendor should be responsible for maintaining and holding
these electronic health care records, that is about the same thing that saying
that the East India Trading Company should be responsible for the contract to
the Apollo space missions.
When you think about it that way, it is ridiculous. No offense, guys. They
are not going to be around in 150 years. It just isn’t how it works.
I would like to zoom out in two directions. One, this has been working for
an awfully long time. Two, we need the stuff we are doing now to last an
awfully long time. I know you have some serious deadlines. I have heard 2010
and 2011 over and over again. I don’t mean to pretend that those aren’t serious
things. But I would also like you to keep in mind that we need this stuff to
work for centuries, for centuries. So I would love it if you would consider
funding only open source. I would love it if you would consider mandating that
all these guys go open source. I think it is unlikely that you are going to do
that. So I am going to provide an alternative suggestion.
Because of what I am saying, what I would like to show you is that we are
doing something. The open source Vista community — also, I want you to
understand that Vista is a way to do health records in the open source. You
pair a programmer with a clinician and you have them work together to develop
code which you would collaborate on. That model, which is essentially an open
source model on comparison to how they develop the LOINC kernel, the Apache
model. Its own development model works. So it is not just Vista, it is the
Vista ideals. It is the Vista Spirit.
So what I am going to ask you to do is not to hold us back. Please don’t
hold us back. What holds us back? If you go into these requirements assuming
that you are dealing with a proprietary vendor, that you are funding a
proprietary vendor, you create a lot of problems for us. For instance, CPT
codes. Some of these proprietary ontologies.
I want to paraphrase something from David Olman, the CEO of Clear Health.
If meaningful use ends up requiring American Medical Association’s current
procedure terminology, proprietary editions of ICD-9 and ICD-10 codes, or
direct electronic transmittal of prescriptions through SureScript, then they
are precluding a completely open source solution out of hand. Why? Because open
source requires that we give away copies of the software, and that the people
who we give them to have the right to give it to someone else.
That is retransmission. It is part of the open source credo. We can’t
change these rules, it is in the license. We can’t pass on the right to use
CPT. So it seems trivial. This is tax that everybody else is paying. For us, it
is a deal killer. It is a serious impediment.
SNOMED. You guys had something to do with that. Awesome work. Use that.
The other thing is that we need to have a certification model that is not a
feature bucket certification model. Laundry list stuff does not work for us. We
are having to dumb down our Vista installations so that they will meet lower
standards in CCHIT.
You have got to remember, Vista is awesome. It does everything, and it
MR. REYNOLDS: You’ve got one minute, sir.
MR. TROTTER: It does all this kind of stuff. So please don’t give us this
certification. All us to create our own certification boards, not because CCHIT
is not working with us. They are, they have been showing us some love. Mark,
thanks for that. But we need to be able to be able to have another
certification body in 15 years. Again, we are trying to solve this problem in
the very long term, so please help us do that.
MR. REYNOLDS: First question, John Houston.
MR. HOUSTON: Thanks. I have a question for Justin Barnes. I was interested
in your thoughts on the HINS analytic seven stage model for EHR adoption and
whether it might be a good model to show progress on the adoption of hospital
EHRs over time. Additionally, whether a similar model could be adopted for
positioning EHR adoption.
MR. BARNES: I’m sorry, I didn’t hear the entire question.
MR. HOUSTON: I am interested in your thoughts on the HINS analytical seven
stage model for EHR adoption, whether it would be a good model to show progress
in the adoption of hospital EHRs over time, and if there is an analogous model
that maybe could be developed on the ambulatory or physician side as well.
MR. BARNES: Did you say that I mentioned that, or do you want me to make a
MR. HOUSTON: No. I noticed you were related to HINS, and I know that HINs
analytics though is somewhat different an organization. I thought you had some
experience with it.
MR. BARNES: Sure, I do have a lot more experience on the physician practice
side, but I certainly think that we probably could create a stage such as that
on the physician practice side. But obviously the HINS analytic version is
completely focused on the inpatient side. But I certainly would see it as
helpful in the physician marketplace, and to get into a stage four, five and
six or something like that would show a clear benefit like we are starting to
see now in the inpatient space with those hospitals. I think one just hit stage
DR. SUAREZ: One of the challenges with certification and the certification
process that I have seen is the balance that the certification criteria itself
has to have with respect to capabilities versus requirements.
In other words, many of the criteria that are used to certify products
focusing on, the product is capable of doing this or it is capable of doing
that. Whether the product’s installation actually turns on or keeps off that
type of capability is a decision of the customer.
What has been found in many respects is that that turn on-turn off
capability has made it more difficult for interoperability to actually happen,
because at the end the change doesn’t happen someone doesn’t turn on or off
MR. REYNOLDS: The question?
DR. SUAREZ: The question is then, how do we move from having capability
criteria to having requirements in terms of interoperability in the products
MR. BARNES: I’ll take the first stab at that, and then pass it down,
because I’m sure we have a lot of feelings on that.
I think there is a healthy balance there. There is a balance between
providing a solution in your practice that your customers can use and want to
use effectively. I think that is one of the reasons we are here today, to
figure out that meaningful use criteria, those areas that they can’t turn off
necessarily. Obviously many of us are in support of very similar items around
HITSP and interoperability exchanges. We want that information exchange. I
think there has got to be a healthy balance of what — and I think we have all
mentioned that in our statements, that we find that balance between what
systems do and what providers can do at the point of care with their patients.
I know that is the healthy balance, and hopefully we are here to
constructively discuss that and find some solutions.
DR. SEGAL: Just to make a quick point, one of the things I think we are all
going to be dealing with is, the role of certification is really shifting.
Heretofore, certification has been to help purchasers make decisions. If they
know what is in it and they choose to turn something off, then that meets their
I think one way to handle this is meaningful use. Again, there are specific
things that have to be done with it, in terms of interoperability. I think you
need that balance of reflecting the fact that there is a range of user needs
But I think we need to reflect, and I think we will have some conversation
tomorrow, on just the changed nature of certification, and the fact that now
providers will in effect will need to continually keep a certified product to
be eligible for bonuses and to avoid penalties. That leads to a variety of
questions that I think need to be identified and answered about this
fundamental nature in the nature of certification.
MR. TULLMAN: Just to make one comment, as software developers we can make
the software whatever we are asked to make it do. So to the extent that the
meaningful use requirements say that a switch ought to be set a certain way, I
guarantee you that everybody sitting up there can build it that way.
The question is will it get used? We have been in a situation where we have
had to try to convince physicians, who all wanted to do it different ways, to
use the software. So that led to software that was very broad and allowed you
to do many different things. It also led to lengthier implementation cycles and
So to the extent we firm up those requirements, I think what you will see
is that that will happen very quickly. Now, because we have the incentives,
physicians have a reason to change their behavior. So I think it is a unique
opportunity that we have.
MR. TROTTER: This is actually one of those areas where the transparency of
open source really changes the game substantially. One of the things that you
can’t do very easily with the proprietary model is reach deeply into the
software to see what it is doing. But you can do that kind of thing with open
I have already heard it here. This is the Leapfrog model, where you
basically have this testing requirements and the script that you are going to
go through. We send you a patient. You do these certain things to the patient,
did you catch this, did you catch that, what happened here, what happened here.
There are three areas that are basically the edges of a certification. They
would be security, interoperability and quality. There are good standards
available already that we have already heard about today to measure those
things as they come out of the system.
So using the Leapfrog model, if you throw into the system, including doing
so by injecting the interoperability standards, then you run a bunch of tests
that are specifically geared — the EHR is capable of testing that it is doing
what it says it is doing, then you can certify continuously at the provider.
So that is a totally different model that is possible with us. This is
essentially what we are going to be proposing to CCHIT. They have seemed that
they are friendly to it. So that is a different way of thinking about it.
DR. MIDDLETON: Thank you, gentlemen, for a very stimulating and
entertaining panel as well. I have two fundamental questions, and maybe
everyone can get a crack at them.
One is, Fred, you highlighted the issue of how the knowledge base that we
use in our EHRs, and the various knowledge when they are at cost for any
developer, whether open source or the vendor, they may present a significant
barrier to both use and open source, and perhaps accelerate the national
adoption of HIT.
So the question might be, which knowledge bases would you suggest should be
in the public domain and freely available for all HIT vendors to use, private,
public or open source?
Hold on to that thought. That is question number one. The second is, if
there is a set of commodity technologies which do not differentiate between
vendors, should they similarly be in the public domain at some level, and
should ONC or the relevant national agency be in the business of creating the
commodity technology such as record resource locators or other identify
technologies and stuff that we all have to use, whether open source, private,
public, et cetera?
I agree with you, Fred, that there is a whole set of current open source
technologies, widgets and various forms of things that we are all using
already. So the question is what is that common set of stuff which you wouldn’t
care about being in the public domain and is not a market differentiator?
MR. TROTTER: First of all, I would like to point out that if you are going
to do a really big push on open source that way, what happens is that you have
competition, but it is in a different place. It is in the service and support,
which is exactly where you want it.
So I think on that angle, then the question becomes, are you interested in
being in the role of an open source project manager. The answer to that
question is no, but you do want to be the project starter. On this, you guys
are a little bit pregnant. You guys are doing all kinds of really cool open
source stuff right now. You are already in this game.
This is a hard game. The government starts a project and then hands it over
to the community and says, run with this, guys. That is something that has not
been done a lot, but it is possible.
That is what I think you should be looking at. I don’t think you should be
trying to be a project manager, because being a project manager in open source
means you have to be a real bully sometimes. It is a meritocracy. You have to
tell an entire group of people, you are wrong and you are not going to get your
way, and you have to do it on a regular basis. That is not exactly the role I
see for government.
DR. MIDDLETON: On the question, then. If you focus on the question, which
of the knowledge bases would you suggest need to be in the —
MR. TROTTER: Going back to the knowledge, you don’t have to fight for
SNOMED, so that is cool. Besides that, surprise us. I think the SureScript
problem is a pretty significant problem. It doesn’t seem like it, but it is a
proprietary interface that they don’t share. I have gone after CCHIT,
SureScript is next, to see if these guys are willing to open up. But those are
the kinds of things.
The other thing is, SureScript is a good example of something that doesn’t
seem like it is proprietary, but still is.
DR. SEGAL: Two difficult questions. I can’t comment on the full range of
what knowledge bases should be in the public domain. I think probably
critically, and we have heard something about it today, if you think about
evidence based medicine and the need for example to take performance measures,
some performance measures are in the public domain and some not.
But as much as possible having the material that comes out of evidence
based medicine that can be used for clinical decision support, performance
measures, practice guidelines, and having that be in a machine computable
format in the public domain where that is possible, I think you get a lot of
benefit there. You may have innovation, you may have proprietary nature around,
particular ways that is used and transported, but the fundamental information
can move quickly from the public domain standpoint.
Then from the standpoint of — I think you called them commodity functions,
I think there is already a lot of that out there that are utilities that are
open source. I am certain we are using some of those in our products.
I think Fred raises an important point though about the government role
there. Our perspective is to separate consideration of open source from issues
of government developed EHRs. Again, I think there is certainly a role for
tools, such as what is happening with the NHIN.
I couldn’t specify particular commodity functions. I think there is a role
there, but I think it is also a slippery slope. I don’t know if that is the
best use of government time and energy to get into that software development
MR. TULLMAN: Let me comment first on the knowledge base. SNOMED is fine.
Frankly, from our perspective any of the databases are either a cost for us,
which we pass through to the ultimate customer, or are free for us. So to the
extent that our customer wants to use SNOMED, we are happy to use it to the
extent that the government or anyone else wants to make these free. We are
happy to use them and incorporate them into our products. Again, the idea is
The second thing I would say is in regard to the comment on SureScript. I
would tell you that I have no association with them, other than we send them
millions of prescriptions. But a few years back we had an industry that
couldn’t connect. In the span of two years or three years, we have an industry
that has made dramatic progress of connecting over 50,000 pharmacies.
If we want to fix health care, going back and revisiting things that are
actually working very well today I don’t think makes a lot of sense. I think
that is a big success story. I don’t have any connection, other than us as a
vendor going to ten different shops and trying to recreate it each time, to
send all of our transactions to one place and know they are going to be routed
effectively. Now through the merger, SureScript and RxHub, which I thought was
great, now to get back eligibility information, cost information, I think that
is a success story. I think most of the vendors in this industry would tell you
that. It worked, we did something and patients are better off for it.
Finally, what I would tell you is that we don’t have relative to open
source or any of these other technologies. What we are there to do is serve the
customer, whoever is paying. So consequently, some of our products were client
server, and then the customer started saying, we want software as a service. So
guess what? We built it, we delivered it to market, it is available now. They
have more choice than they probably ever had before.
I think what you are seeing is, because of the stimulus, you are going to
see more and more innovation, from not only providing electronic health
records, because one way or another they will get over a period of time
commoditized. They will also look more and more the same because we will figure
out what works best. But then the next stage is where we want to get it. That
is all of the information services.
So if I can give you just one analogy, we all bought computers. We were
told computers were going to eliminate paper. Then the next ten years we had
more paper than we ever had in our lives. The paper industry had a heyday. Then
all of a sudden we connected those computers together, and what happened? We
started seeing enormous innovation in terms of information services.
As all of us start to connect together, and hopefully the new meaningful
use will require interoperability, as we do that you will see an explosion of
new companies, existing companies, with services that really provide better
So that would be my answer.
MR. BARNES: I can answer it really quickly. I agree with SNOMED. I think
that is certainly a knowledge base that we potentially could start with.
I think on the open source area, first of all, probably none of us have
issues with open source. We haw invited them into the association. We would
love to work with them, interoperate with them. I think on the information
exchange area is probably where you are going to get a lot of freeware in a way
for our customers. We are going to create a medical record, that medical record
is going to be shared between practices.
Today open source is available free of cost to a lot of practices in
America. I think what a lot of customers are asking for, every day it seems to
us, is more integrated solutions, not piecemeal solutions. They want high
secure, high integrated solutions, practice management, electronic health
record interoperability. They want it in one single product and flow seamlessly
to point of care with split second response time.
So whoever provides that, I think that is exactly where the market is going
to go. The market dictates what we create, where we innovate, and obviously the
federal government will as well with their incentives. But I think that
ultimately our customers —
MR. REYNOLDS: Judy Warren, last question.
DR. WARREN: Earlier today we heard that certification of EHRs or software
itself has some impact. But there is variable implementation of those EHRs.
You had mentioned in there that what vendors needed to do was to increase
their implementation capacity and then to train and hire implementation
personnel. Could you define what you mean by implementation capacity and
implementation personnel? Do those people need to be clinicians?
MR. BARNES: I will certainly answer that. First of all, I think all of our
companies are ramping up, and a lot of the implementation personnel are
clinicians. When you are down in the trenches in the practice, the peer to peer
helps with the implementation. The explanations around the clinical components
are very critical.
We have physicians, we have nurses, RN, they are a whole gamut helping, as
well as just technical resources that are very important. In this labor pool
right now there are still a lot of resources still out there. We are in a job
market decline in many segments of America, if not all over the country. So we
have very large pools to pull from. I think that we are going to continue to
see training in these areas very extensively. Hopefully those resources can be
So that is what we are doing. I can say for my company, we are enlarging,
and I’m sure GE as well in those areas. So we are certainly ready, and
obviously regional centers will help as well. But I certainly see clinicians as
a vital part of that.
MR. REYNOLDS: Last comment.
MR. TULLMAN: I would just add that again, you go back to the question. We
will have plenty of people to do the training. I don’t think that is going to
be the issue. The question is what you are training them to do and what kind of
standards we are going to have.
So the more flexible all of our software is, in some respects the bigger
problem it is, because people can do things different ways. So what we hope to
see is a very high level plan that says every piece of software in this space
is going to connect to a registry. Every piece of software is going to be able
to share data. Every piece of software is going to have vibrant electronic
health records. The only way physicians and other caregivers are going to get
the dollars is to do what is specified, to use it in that fashion.
So I think we can implement it any way we are asked to or we are told to.
But I hope that what we see come out of here is some very strong guidance,
especially relative to the next stage of where we ought to go. That is, not
only using this to make us more efficient, but using this to provide better
care to patients.
I’ll just put in one last plug. I think it is critically important that we
make certain that all the software that we produce is open to patients as well.
What we don’t want to do is two things, one, take a paper based siloed system
which is what we have, and make electronic silos. We see that today. We have
offices using a full electronic health record, except they can’t talk to
anybody. So then they try to print it out if you want that information.
The second thing that we have to do, we have to use that information that
we are gathering from the electronic health record to provide better care.
MR. REYNOLDS: I know there were other questions on the panel, but we have
come to the end of our time. Always when the vendors are up there, there are
plenty of other questions that could have been asked. So we thank you very
much, and if the next panel could please come forward.
We are starting our next panel. The next panel is the use for 2011 and
beyond for providers. We will go through it in the order, but before we start I
would like to recognize one of our panelists, the Honorable Nancy Johnson,
former member of Congress from Connecticut. As a committee we have followed
your work extensively when you were in Congress, and have deliberated a number
of your positions, and we appreciate all that effort that you did, and we are
anxious to see you back here in your new role. So with that, please begin.
MS. JOHNSON: Thank you very much. It is really a pleasure and an honor to
be with you today. I consider the work you are about to be crucial work, the
difference between whether the vision the President has and many of us have of
transformational health reform succeeds or fails. So I appreciate very much
being here, and it is a pleasure to see Dr. Blumenthal with us to join us in
this important effort, and all of you sitting here today.
I just want to remind you as we start, in January the National Research
Council — I did work hard on health IT legislation while I was in the
Congress, and I share your vision of the importance of the electronic health
record and the role in changing the structure of health care delivery. But I
also helped to start a RIO in my home state, thinking that would help us
connect all of these things together. So I want to start by saying that I am
very impressed by the report issued by the National Research Council on January
9 of 2009. I think it is important for us to remember that they said that what
we have been doing will not be sufficient to achieve the vision of the 21st
century health care, and maybe even set back the cause if these efforts
continue wholly without changing their present course. Specifically, success in
this regard will require greater emphasis on providing cognitive support for
health care providers and for patients and family caregivers.
I have come to feel that if we don’t have a record that patients can
access, that their caregivers can access, our hope that we will be able through
chronic disease management, through prevention, through those kinds of
approaches, to long term reform the American health care system so that it not
only addresses disease, but addresses prevention, wellness and care management.
I was very pleased that the statute passed by the Congress is very focused
on an individual record. I think an individual record with the cognitive
support is the difference between success and failure in the future.
This is just to remind us that in the past, technology development has been
focused on the professional care sector. Some systems have had tremendous
success intra system. But we don’t really have any examples of a similar degree
of success inter system. So our success on the inter system issue rests on our
ability to over a person’s lifetime, through the course of their health needs,
their wellness, their illness treatment, their aging needs, to have the support
of the health system at their fingertips. So this is just to remind us of where
we are starting, and how we have narrowed the focus in past years.
This system has very little to do, the systems we have developed, with the
patient’s own experience. The patient goes from doctor to doctor, particularly
if he has a complex disease. We have not changed the system to address
So the goal of this legislation as far as I’m concerned is to remind us
that we have to stop trying to fix the old structures, the big structures, the
institutional systems. It is curious, isn’t it, that we have great systems for
payers, we have some great institution systems for institutions, we have no
system at all for patients.
That is what this bill is about. It is about the individual patient having
an electronic health record that is real, that is up to date, and that all of
his or her providers can access.
The institutional records are important. They are beginning to develop, the
institutional electronic systems. They have a ways to go. It varies a lot from
institution to institution. But they will be driven forward by quality
standards, by DRG development, by the need for efficiency in institutions. But
we must guide them forward in terms of their use by the patient, their
communication with the patient, with keeping the patient and the patient’s
doctor in their own health loop.
So I don’t worry about those systems developing. I worry very much about
whether or not we will fix the process problem for patients. The easiest way to
help the institutions is to fix the process problem, the record problem for the
If we do that, then we will have — if the patient goes to the doctor in
the top line, that doctor has a consistent record until the patient comes
again, and has a consistent record until the patient comes again. But if he
translates that information to the individual patient’s record, that record has
that information, and the next doctor has that information, and the next doctor
has that information. The next time that patient comes to see the doctor in the
top line, he also then has the record of all the other doctors have done.
So that individual persistent patient record, look how simple it is. It
doesn’t require the first doctor in the top line to then send the information
to every other doctor. It isn’t that complicated. It doesn’t open that
information to a level of possibility of error or exposure.
So this concept of an individual patient record underlying everything that
the patient has access to, there is some experience now that shows that if
patients have access to their own record, particularly those with multiple
chronic illnesses, they use it, and they use it to be sure that they are doing
state of the art care, that they are keeping up on what the doctor told them to
do. Their caregivers are using it to be sure that they follow the protocol the
As we move into the private sector redesign and private sector insurance so
that people will be rewarded for good care, they will want to know, did I get
my preventive care, because my premium will go down if I do. They will want to
know, am I up to date on my chronic disease management responsibilities,
because if they are, their premium will recognize that.
If you don’t incentivize the patient to be involved in their own care, to
participate in their own care, and to take some responsibility in their own
care, we can change the rest of the structure as much as we want, but we will
not transform health care. Transforming health care is cultural for patients as
well as providers. It is one thing to try to change providers, but if you don’t
put the patient on the physician’s team, the physician is going to have a very
hard time doing what his team indicates he should be doing.
In my testimony — there are some copies back there — there are some
examples of how you can use an individual health record. The patient gets a
record, the doctor gets the kind of record he needs. You get the interchange of
information. You get the platform for rules driven, evidence based medicine
rules and benefit optimization rules and pay for performance compliance and all
of those things. So the platform can be there, and it is current technology of
things that people are working with now.
So what has Congress provided? It is important to go back to the test of
this legislation. It does talk about patient records. It doesn’t talk about
doctor’s records or hospital records. It doesn’t talk about use by doctors or
hospitals; it talks about use for health care. It is use oriented. It is not
going to pay you for having this stuff, it is only going to pay you for using
this stuff in behalf of the patient. That is a huge difference from other
technology legislation that has been proposed, and certainly nothing has ever
Furthermore, it is focused on accountability. It is a huge investment for
the federal government. I have never seen anything like it. I do think it is
exciting. I do think it is important. I do think it is worthy, because if we
don’t make this move, if we don’t get where we want to go, we will have no
chance at all of doing the one thing that America could do that no other nation
has done, that is, control costs by improving the quality of health care,
improving prevention, improving early identification, improving early
management, early intervention, chronic disease case and all those things. If
we don’t do it that way, our only alternative is just top down, cut the costs,
cut the reimbursements, deny access.
So this is an experiment that I think every person in the world has an
interest in. I know it will take our consultation with ourselves and others as
well to achieve it. The only way we can get to where we want to go is through
smart records, that is, systems that support major reforms, comprehensive, real
time, relevant, persistent records. It has to be the same record all the time.
It won’t do to accumulate every time it needs to be used, because every time
you will have a different record. You have to have a persistent record for a
lot of reasons, including malpractice.
It has to be an understandable record. You can’t just dump a whole doctor’s
record into the record. Somebody has to sort it out. That is the ontology. It
has to support both individual care and system change.
The challenge you face of detailing what the Congress meant by use and by
quality technology is at the center, at the heart, of whether health care
reform succeeds or fails.
MR. REYNOLDS: You’re down to the last minute, ma’am, please.
MS. JOHNSON: So our recommendations are very specific. You have been
hearing them here today, but I think they have to be detailed out. We have to
insist on integration of patient information, of information from providers and
individuals and formal caregivers, access for them across the spectrum, and so
on and so forth, privacy and confidentiality.
I’ll just get to the last one, which is when you look at meaningful use, we
have to look at whether or not these things are used, whether the patient
center record is then used, whether it is used to communicate with all the
patient’s caregivers on a regular basis by contributing receiving data from
qualified electronic records, and that they have developed access to the
So if it is Internet based, the access is there, and so the money should
flow. We make a terrible mistake if we only let the money flow to those who buy
an electronic health record in their office with all the things that that has
in the past implied.
So I appreciate your attention. It is a pleasure to be here. I look forward
to the discussion. Thank you very much.
MR. REYNOLDS: Thank you very much. Dr. Perlin.
DR. PERLIN: Chairman Reynolds, Dr. Blumenthal, distinguished committee
members, ONC staff and colleagues. Good afternoon. Thank you for the privilege
of being here.
I am Dr. John Perlin, President of Clinical Services and Chief Medical
Officer for a national Tennessee based HCA. I bring the perspective of a very
large health system, operating more than 160 hospitals, 200 ambulatory centers,
380 physician practices, providing about five percent of the major hospital
services around the country, with about 185,000 employees, including 200,000
nurses and about 1500 employed physicians and 80,000 physicians making up the
voluntary staff, talking about small and large practices.
In the early stages of moving from a basic electronic health record born as
a hospital information system, to providing advanced clinical functions
including computerized provider order entry and interactive decision support, I
want to thank you for the support for the HITEC Act of the economic stimulus
legislation, which has unequivocally accelerated our migration. I would like to
share our thinking on the concept of meaningful use.
I would also note that my perspective is informed by the privilege of
having spent nearly a decade in the Veterans Health Administration, working
with colleagues in a number of roles, first as quality officer, then as acting
chief of research, as Deputy Under Secretary and then ultimately as Under
Secretary for Health, successfully implementing electronic health records
across that system. I appreciated the comments made in the last panel.
Results as validated in the academic literature and the lay press were
unmet in terms of advances in safety, quality and efficiency. In short, the
electronic health records improve the value defined as the relationship of
quality outputs to resource inputs to health care, the ultimate objective of
the HITEC Act.
I want to address your questions. I’m sure you as well have received input
from colleagues at the American Hospital association Federation, academia and
In terms of the EHR functionalities necessary to provide safe patient
centered and effective care, the EHR functions that are critical for supporting
these attributes are enunciated with remarkable consistency among researchers
and practitioners, supporting the primary act in medicine, first do no harm,
and secondarily informing care that might be provided. Thus, the list starts
with allergies and medications, and the next level is the functionalities, the
problem lists, past diagnoses, laboratory data and then more detailed treatment
Let me address the questions that you asked about critical success factors
for robust participation by professionals and hospitals, and continuing that
participation in later years as well as the barriers.
The number of hospitals eligible to participate in the incentive framework
in 2011 will first depend on how high the bar is set and defining the
functionalities required for meaningful use and the level of connectivity
required to demonstrate health information exchange.
I would like to comment on three critical factors for attracting robust
participation in the incentive program. First, it is imperative that we provide
clear guidance on the desired outcomes at the end of the defined period. I
emphatically agree with Dr. Blumenthal’s New England Journal editorial, in
which he argued that the point of health IT was improving health care and
value. Thus, the desired outcomes of meaningful use must transcend technical
specifications and include guidance addressing improvements in safety and
quality that are intended.
We agree again with Dr. Blumenthal that health IT is a means, not an end,
and will be inadequate alone to transform our health system into one that
provides the right care for the right patients in the right place at the right
time. It really can be a tool for optimizing care processes that have been
designed to be safe and patient centered and promote high quality care.
Second, it is rational that we build on what currently exists. We should
build on both existing quality programs like the Hospital Quality Alliance, the
Physician Quality Reporting Initiative, as well as from the existing data
standards, framing the outcomes in terms of both technical specifications of
use and measures of health care improvement.
Third, we recommend progressively escalating incentive requirements over
time. As an opinion article by John noted, only one and a half percent of
hospitals excluding the Veterans Health Administration, currently meet the
definition of even modestly using electronic health records with advanced
clinical features such as interactive decision support.
A number of formidable issues underlie this anemic utilization rate. To
date, system implementation, even open source system implementation, is
expensive in terms of capital acquisition and the concern of potential losses
from disruption of hospital and physician practices. We are encouraged that the
incentive program will foster innovation in terms of both cost and usability,
but in the interim the requirements to increase utilization by physicians in
particular require, one, up front, even speculative investment to avoid the
2015 penalties and two, frankly changing an ingrained culture that is paper
Even with full commitment, implementation in complex hospital and system
environments is a multi year endeavor. While we strongly support and see a line
in the sand, we are encouraged that the incentive program will foster
innovation of both cost and usability and support the required culture change.
In terms of technical specifications, a staged approach to defining
meaningful use should be utilized to allow as well for industry innovation over
time. Well developed, well deployed technology exists today to view and
exchange labs, imaging and pharmacy data, making these three functionalities
the logical starting point for meaningful use.
Beyond the hardware and software challenges, we have a warm ware or a human
challenge. Clinicians have not embraced the information revolution with the
fervor of other industries. While this derives in part from unfriendly software
with impenetrable user interfaces, and while we appreciate the motivation that
the positive and negative incentives provide, as it is sometimes said, even
nine women cannot make one baby in a month. We have to achieve culture change,
and that is of course the end, not a starting point. We are also anxious to see
the full field of play, and look forward to learning over the immediate near
There are important factors that will be relevant to continued
participation in the later years of the incentive program and beyond. It is
imperative that incentives that are enunciated at the outset not be diminished
during the duration of the program. Not only would this undermine the intended
activity, but it would render any future incentive programs without
credibility. Additionally, the penalty phase beginning in 2015 will provide
incentives for not only continued participation by earlier adopters, but can
serve as the tipping point for late adopters becoming meaningful users.
I want to address the comment on the anticipated impacts on special
categories of providers. We feel strongly that the incentive program must raise
all ships. Exclusions of classes of providers will magnify disadvantages of the
populations they serve. Functionality without information exchange will not
transform the system. The greatest value added will be in the ability to make
the data flow by creating an interoperable health IT architecture. Outside of
the nascent and limited network provided through the initial NHIN trials and
RIOs, we are still struggling to make health information exchange operational
on a national scale. In short, as we build the equivalent of the personal
computer revolution for hospitals, doctors’ offices and patient homes, the
government needs to assist to foster building the web that makes the Internet
With respect to the administration of the incentives, it will be rational
that early adopters not be penalized for early investment or fear that largesse
disproportionately accrues to late free riders.
In closing, it seems advisable that eligibility for incentive payment
requires that health IT systems are not only installed by providers, but they
are really used by clinicians. Full implementation of health IT will require
critical changes in culture and work flow to insure that clinicians are using
the technology in a meaningful way. This will be a collective learning
experience and will benefit from accruing best practices and lessons learned
successfully around the country.
As well, we hope that there may be technology innovations like scalable
application service provider or software as a service model that can support
physician practices, even hospitals.
It will be beneficial to have clear articulation of five year goals,
redefining both technical use requirements and population and health outcomes.
Meaningful use would involve practitioners increasingly utilizing interoperable
health information technology to one, generate clinical action and two, to
inform those whose patient general histories would suggest progressive
requirements for computerized physician or other provider order entry and
connection to a qualified health information exchange as two central tests.
We strongly support a definition of meaningful use that assures meaningful
change. Given the height of the hurdle financially, functionally and
culturally, breaking the goals into achievable steps that recognize walking as
a prerequisite to running would prove helpful.
We thank you for the opportunity to provide input and we look forward to
working with you to achieve the intended benefit of HITEC all Americans. Thank
you, Mr. Chairman.
MR. REYNOLDS: Thank you. Dr. Mostashari.
DR. MOSTASHARI: Good afternoon. Thank you for the opportunity to testify. I
am Dr. Farzad Mostashari. I am Assistant Commissioner with the New York City
Department of Health.
Congress was wise to offer funding in the HITEC Act for the meaningful use
of electronic health records. By doing so, Congress created the mechanism to
safeguard their intent that this unprecedented public investment not treat
technology as an end to itself, but rather results in concrete progress towards
policy objectives, meaningful improvement in health, and reduction in costs.
To fulfill this intent, we must work backwards, from the desired outcomes
in order to determine meaningful use metrics and the scope of work for regional
extension centers. Implementation of the HITEC provisions can then proceed
along a consistent pathway from product development to EHR implementation to
practice redesign, quality monitoring, and ultimately reimbursement reform that
will both benefit from and sustain these activities.
This is the glide path. Virtually every health reform strategy under
serious discussion includes encouraging more organized systems of care and
increasing accountability for health care quality and cost outcomes. The
transformative potential of health IT is to provide the information necessary
for organized care delivery and drive transparency in health care outcomes.
This is the end point for meaningful use, incentive payments tied to
measures that matter, that will sustain improved clinical outcomes like blood
pressure control, evidence based care like smoking cessation therapy and
aspirin use, patient safety potentially around adverse drug events, continuity
of care, patient satisfaction and compliance with public health reporting.
Meeting these end points may require practices to engage in new processes
like electronic prescribing or patient outreach to embrace health information
exchange around medication histories, clinical summaries, notifiable disease
reports and quality reporting.
The midpoint criteria for meaningful use must be strong enough to spur
these real changes in care processes, but focused enough to be feasible for
motivated practices in all settings to achieve.
What fuels these processes is information, electronic data that can be
queried and collated. This is the beginning of meaningful use, collecting and
documenting the critical data elements needed for organized care processes,
continuity of care and quality reporting.
Smoking status must be asked. Blood pressure must be measured, electronic
problem lists maintained, allergies and medications documented and laboratory
interfaces established. There is reasonable concern that even such a stepwise
escalation of meaningful use criteria could overwhelm many small private
practices that lack ready access to IT expertise.
What I can add to this discussion is our experiences at the New York City
Primary Care Information Project. With spending authority from public funding,
we established a master contract with a commercial EHR vendor, E-Clinical
Works, a little more than two years ago. As we detailed in a recent Health
Affairs publication, we worked closely with the vendor to develop new
functionality that supports prevention and population management, including
integrated registry functions, point of care decision support for providers,
and automated quality measurement.
We have reached out to primary care practices that serve Medicaid patients
and have signed agreements with over 1700 providers in small practices,
community health centers and hospital outpatient departments.
In the city’s three most underserved communities, over 50 percent of small
practices are in our program. In a little more than one year, over 1,000
providers in over 150 of these independent medical practices were brought live
with a 99 percent implementation success rate. We provide boots on the ground
project management support for implementation and post implementation technical
assistance around effective use of the EHR to protect patient privacy, to
improve practice efficiency, safety and quality of care. We use our scale and
technical expertise to convince commercial laboratories to grant electronic
interfaces to even small practices.
We work with New York State Medicaid to provide 90-day medication fill
histories back to providers. We have created a clinical hub that could link our
practices to each other, to local RIOs or personally controlled health
repositories when they are truly operational.
Without creating a massive centralized data warehouse, we are receiving
monthly utilization reports that enable us to identify and help practices use
the record more effectively. For example, one practice we noticed was
documenting allergies almost all in an unstructured manner. A little education
and now they are doing it the right way.
We developed through funding from AHRQ standardized quality reports on a
small set of key clinical outcomes like blood pressure control among patients
with diabetes. They are visible to the providers themselves and reported to the
department. Quality data from these practices as well as New York City
practices using two other popular commercial EHR products are already being
used in an innovative and robust trial of pay for quality, and will be shared
back with practices for benchmarking purposes.
Developed from funding from our CDC center of excellence in public health
informatics, the health department is also receiving electronic reports of
immunizations and daily reports on trends in respiratory, febrile and
gastrointestinal disease that are helping us monitor the swine flu outbreaks in
New York City today.
I know there is much work to be done before we reach our goal of improved
community health. But at a cost to government of about $20,000 per provider,
our regional extension center has insured that our practices, including the
solo practitioner in a storefront in Harlem, will qualify for the physician
payments in 2011 and forward by any likely metric of meaningful use.
Some will say that the difficulty of changing physician practice means that
we should set low goals for meaningful use, perhaps simply adoption of
certified records in the early years. But providers don’t want to implement
technology. They want to improve care for their patients. Any effort however
well intentioned to help them by reducing meaningful use to simply adopting
technology will not only fail to improve care or costs, but is also likely to
fail at engaging most small practices in EHR adoption itself.
Setting a standard for meaningful use that is consonant with the mission of
health care providers and supporting them through robust and empowered regional
extension centers was our best bet.
MR. REYNOLDS: Excellent. Ms. Murphy, it is always nice to have a nurse end
the day, because Judy Warren is much easier to live with as we go through the
evening. So thank goodness. This would have been an ugly dinner tonight if you
hadn’t showed up.
MS. MURPHY: I appreciate that support, thank you.
MR. REYNOLDS: So thank you so much.
MS. MURPHY: Thank you for the opportunity to testify. My name is Judy
Murphy. I am a nurse. I also sit on the American Medical Informatics
Association board of directors. My day job is as Vice President of Applications
at Aurora Health Care, which is an integrated delivery network in Eastern
Wisconsin with 13 hospitals, 120 clinics and 28,000 employees. Our providers do
include a thousand employed physicians, about 1800 affiliated physicians, and
over 6,000 nurses.
I have personally been involved in implementing health information
technology for 24 years, with both custom program development and purchased
vendor products, and in all aspects of the system life cycle. I have led
implementations in both the acute and ambulatory care venues, including
e-prescribing, computerized physician order entry, bar-coded medication
administration, positive patient ID, clinical documentation, embedding evidence
based practice for nursing and for physicians. So I live every day in the
trenches of HIT implementation, its governance, structures, its caregiver
readiness and adoption issues, technology challenges and budget constraints.
I am really thrilled to be able to address you today on something near and
dear to my heart, that is, how to implement and use health information
technology to improve the quality and safety of care. I of course doubt there
is no one in this room who would disagree that electronic health records is a
linchpin for transforming health care delivery and a key enabler of health care
reform. The biggest question is, what path do we follow and how do we know when
we got where we need to go. Hence, the need to define meaningful use and plan
the road map to achieve it.
I must add that I am not sure that glide path is the appropriate analogy
here. Of course, in health care we often use aviation as the role model for our
safety record as well as execution. In fact, this morning Carolyn Clancy asked
if we all had the right stuff. But I don’t think we can necessarily draw on
aviation when we talk about reaching this goal, as unfortunately our path I
don’t think is as clearly guided by a radio beam, and a lot more of this feels
like an exploratory expedition than really an airplane descent.
I am going to frame my comments today around six topics or themes, and look
at the definition of meaningful use and a road map around these topics. Each
theme does incorporate one of my biases, which is a very strong patient point
of view. Of course, several other testimonies today, in fact many of them, have
referenced that as well, that patient centricity.
Nurses do have this partiality overall. Just to remind everybody, there are
2.9 million practicing registered nurses, and we do comprise 55 percent of the
U.S. health care workforce. So as the providers who spend the most time with
patients, particularly in hospitals where we are the caregivers and the patient
advocate 24/7, we have always had that solid focus on seeing things from the
patient’s point of view.
So let me talk first about interoperability. Others today and tomorrow will
testify on the importance of interoperability and the electronic exchange of
patient standardized data between clinical and administrative stakeholders.
This is a given criteria from my point of view and for many others. It is
extremely important for the safety and quality reasons that have been
I support the need for standards harmonization between HITSP and IHE as
outlined in the HIMS written statement submitted to you. I also support their
recommendations to have standard implementation guides published by these
groups by the year 2012, similar to what was done in the past with HL-7 and
DICOMM implementation guides.
I would like to focus my personal comments on a more narrow aspect of
interoperability to emphasize one more time why this is so important to our
patients. Just as Aristotle at one time saw the earth as the center of the
universe, so we in health care often have a clinician or hospital centric point
of view and an illness based model. Subsequent to Galileo and Copernicus’
studies, we now know that the sun is the center of the universe, and similarly
our health care models have migrated to a continuum of care view or a wellness
or health maintenance model with increased emphasis on ambulatory and home
care. Now we see that the patient is really the center of the universe.
Yet our patient records have not kept up with this thought. They are not
patient centric. They do not span the continuum of care, and they are not
transportable between care providers. So this is one of the most important
things we really do need to keep in mind.
In terms of ubiquitous accessibility, health care professionals are
knowledge workers and must deal with huge amounts of health care data as well
as ever changing knowledge related to best practice. Two simple yet very key
points here. The electronic health record needs to be able to assist the
provider in distilling the barrage of information into information, deciding
what is significant about that information, and then making good patient care
Secondly, access to that information needs to be anytime, anywhere, and not
limited to any type of proprietary hardware, but ubiquitous, like access to the
The data distillation involves intelligently filtered patient information
and decision support in all of its forms, including data displace and flow
sheets and graphs well as care management alerts and reminders. It includes the
embedding of evidence based practices for nurses and physicians in the forms of
care plans and order sets, and hard wiring the evidence based care steps,
making it easy for our practitioners in the heat of the moment to do the right
thing. This speaks to the ability to get at patient data and make decisions
about patient care in a way that was never seen in a paper chart.
From an implementation standpoint, this is often the carrot that can draw
our providers to value the electronic health record. It includes features and
functions that overall are generally easy to achieve adoption with, and should
be one of the first milestones we consider rewarding. Third, tethered personal
health records. The HIT industry has just begun to tap into the potential of
e-health and the value of patients participating in their own care, using
web-based tools. Often this is seen today as a self service option and possibly
as a productivity enhancer for the health care organization. But when we open
up our registration systems and scheduling books so patients can arrange
appointments when it is most convenient for them, this actually serves the more
important purpose of demonstrating that they are a partner in their care.
Furthermore, to have patients update their demographics, insurance,
allergies and medication lists now puts the accountability for the accuracy of
this information not only where it belongs, but where the source of truth
There is no other part of the human experience where such a passive role is
played as we see with patients in managing their own health care. So this needs
to be turned around if we are able to get to the next level of care quality. It
is the patient who is the constant across the care continuum, and it is by
focusing on the patient and not the care venue that we can create the seamless
integration needed in order to have the information to provide best care.
I believe this area has the largest potential for impact on the quality of
health care. Online storage of things like immunization records, advanced
directives, medication lists, patient histories in a personal health record
that is tethered to an electronic health record will allow the patient and the
practitioner to co-manage the patient’s care. Our challenge is clear here,
though. Not only do we need to give patients the opportunities to participate;
we may need to convince them of the reasons why that also makes sense.
Technology adoption. We are in the business of health care, and successful
IT projects are not about the implementation of technology, but are about the
clinical changes and patient impact that is enabled, supported and facilitated
by the technology.
Many of us who work in health information technology have long heard the
mantra that technology adoption is ten percent about the solution and 90
percent about social cultural issues such as change management, leadership,
risk tolerance, incentives and so on.
So in our implementation of an information system for nurses and
physicians, the importance of planning and executing the project as a practice
change that is being facilitated by technology cannot be overemphasized. The
HIT needs to take a supportive role to the people, process and practice change
being enabled by the technology. As has already been mentioned several times by
other speakers, we need to insure that the HIT implementation is seen as the
means to an end and not an end unto itself.
In terms of measuring success, considerations for incremental maturation of
EHR functionality and corresponding reporting measures must be incorporated
into the definition of meaningful use. I support the maturation model as
proposed by HIMS in their written statement outlining three phases of not less
than two years, each with identified functionality milestones and outcome
reporting measures. But although it will be important to incorporate process
measures with implementation metrics, it will be even more important to be
focusing on process measures, so the true measurement of success then will be
to focus on clinical and business outcome measures such as the NQF endorsed
quality measures that will align with national quality and performance goals.
Other speakers have mentioned this as well.
Lastly, I would be remiss from an implementation standpoint if I did not
mention the need for an informatics trained workforce capable of executing the
daunting task we know is ahead of us. We need nurses and physicians who can do
this right. In 2005 AMIA created a ten by ten program with the goal of training
10,000 health care professionals in informatics by the year 2010. We need more
programs like this, as well as other educational options in order to insure
that the EHR implementation and use we achieve is meaningful.
Thank you for this opportunity.
MR. REYNOLDS: Thank you. Leslie.
DR. FRANCIS: First of all, thank you. I want to ask you to reflect on some
of the themes in a couple of the earlier panels. We have heard data that there
is enormous variation in the error identification rate from electronic records
systems. The last panel seemed largely satisfied with the CCHIT certification
process, but recognizes there are going to be a lot of difficulties with
Two of you have suggested that what we need to think about is patients and
the role that patients can play, another with support for individual providers.
I want to ask whether as the providers involved in implementation, you think
that there is a role at the standard setting level with respect to the products
that you get, that could be helpful in getting products that are easier to
implement, or helpful with the implementation process.
DR. PERLIN: I’ll bite. Good question. Let me take the last part first, do
standards make a difference in acquisition of products. These are typically
expensive propositions, and one wants to be sure that the investment is not
obsolete. So I think something that has some sort of implication of forward
compatibility, including the comportment to enunciated standards, is very
Having been a former commissioner on CCHIT, I think CCHIT offers a great
service in that regard. The HITEC legislation as I understand it also includes
a component for interoperability. As we contemplate, yes, we will be looking at
both standards comportment and the ability to plug into a network of systems
that allows the exchange of information, because we believe that exchange has
to go not only for the formal provider community, but ultimately has to empower
patients to plug in their information.
MS. JOHNSON: I would also mention that innovation is really just taking off
now in this area. So it is very important to require use, to define what use
has to accomplish, but to be careful not to prescribe the development of tools.
I think the Internet is going to be extremely important. I think between
Internet access, which is very cheap — there are some out there that are
$4,000 a year; I worry about our putting much money into some of these more
expensive ones. So our job is to do the functionality and let competition
innovate around both quality and cost. I think with the Internet and
telemedicine, we have the opportunity to basically revise quality health care
across all America. But we really are going to see some innovations we almost
can’t imagine now.
The other thing I would say is that meaningful use is going to evolve. It
is like continuous improvement. We have to have a way for the Administration to
specifically define uses as we go along in that, and try to encompass all the
uses that might be possible, but give core examples.
DR. MOSTASHARI: In my experience, just because a vendor can under a
controlled setting with a known test script can bring a version of a software
and demonstrate something to the satisfaction of the panel doesn’t mean that it
will be used in practice. So it is good, fine, to have functions, but what
really matters is meaningful use of the product.
In terms of standards, probably the most important thing the federal
government could do around standards, it is not standards or certification for
products or HIEs or so forth, it is the code sets. Just make it clear, we are
going to be using these code sets and make them freely available. I think that
would be the most important thing you could do on standards.
DR. CARR: Thanks to all of the panelists. This is really a wonderful
synthesis, very inspiring.
I have a question for Dr. Mostashari. Thank you also for your great
presentation. With your E-Clinical Works, do you have a patient module? Also,
are there specific barriers from the underserved communities that you have
related to PHR?
DR. MOSTASHARI: There is a patient portal with the products. It allows
patients to do tasks, things, jobs they want to do. That is the strength of
having a personal health record that is tethered or linked, as the gentleman to
your left has pointed out. It actually lets them do something. I want to get a
prescription, I want a refill, I want to ask a question, I want to make an
appointment. So that is good.
We have found that it is sometimes hard to get these rolled out, get them
rolled out for the practices to do the work they have to do to enroll patients,
and to get patients to buy into it. But we are starting to put more effort into
getting the value of them appreciated and greasing the skids.
DR. MIDDLETON: Thanks again also from me for a great panel. My question is
for three out of four of you. Nancy, I apologize, this really isn’t directed at
you. For the provider organizations, roll the tape back to when you were
thinking about starting your implementations. At that point in time, what were
the value metrics you used to determine what was target A, what was target B,
what was target C for the technology implementations?
I guess what I am looking for is whether or not there are tools we can use
from your experience that might help define these intermediate implementation
points, and are there associated ways to define the relevant codes,
terminologies, future functions, et cetera, so we have a way to get at the
DR. PERLIN: I have the interesting perspective of being at this rodeo three
times, in a large health system where we are beginning again to move toward a
large electronic health record. I have been through a massive health systems
implementation in the VA, and an academic medical center, the Virginia
Commonwealth University, Medical College of Virginia, which has had provider
order entries since the ’70s.
I think they have been different situations, but ultimately a theme which
has resonated amongst my fellow panelists and in your comments is that it has
been a value to better care. So I think it is very difficult to motivate
people, but I think we do have a challenge we have to recognize in our current
opportunity. If the better care were a sufficient incentive in and of itself,
we would be there, and we are not.
So I think we need to think about our business cases to different
constituencies. To constituencies of clinicians and patients, I think we have
an opportunity to communicate to patients the social value of improvement. To
the comment that was just asked, the VA launched its patient portal, that might
help with that without any marketing, despite the fact that the demographic
among veterans who are older, sicker and poorer, the first week 200,000
veterans refilled prescriptions electronically. So I think we underestimate the
uptake that is possible. Dr. Tang’s work has also proven how quickly that is.
To the clinicians, I think you need to be credible, that it improves
outcomes. I think the thread here that suggested that this is a means to an
end, not an end in itself, is something that is attractive and increasingly
demonstrated as measurable. So drawing from the academic setting, the VA and my
current environment, working toward better performance on measures, frankly
creating both the carrot and stick, a carrot to attract to those, the
requirement that certain things either happen or don’t happen by date certain
is very compelling. I think we see nationally improvements in core measure
performance, and I think there is an analogy here for adoption of health
records as well.
So in summary, I think that there is social marketing to patients, to make
them — the work of the Merkle Foundation — understand the desire, that same
medicine is practiced here. For clinicians, that the end game is better care,
but to the business community, the thread and common glow is the combination of
understanding or connection to the business case and the delivery of care is
I think we have seen some examples of that with e-prescribing, and I think
that the intent here of recognizing some of the economic burden. Even with
quote free systems, there is still hardware and change to implementation,
recognizing that that is a hurdle, but then setting by date certain a line in
the sand that says this is a requirement. That would be my framing.
MS. MURPHY: My framing would be three: Put it in, get them to use it, and
then measure the benefits. So the getting it in, to me that is whether the
system is certified, what its features and functions are, that is what is
important there. You get it in, you get people used to understanding how it
Then the usage part is, I am using it pretty consistently when I see
patients. I am using it in the actual interaction with the patient. Then the
third is getting to the point where I can’t imagine practicing without it, and
I have adapted or changed my work flow around the ability to use that system.
DR. MOSTASHARI: As I understand your question, Blackford, you are saying
where are specific metrics that we have used in our implementation process. We
had an eye towards where we want to go, particularly around cardiovascular
prevention in the community, so the three measures there carry that through.
We use sales force for our implementation milestones, where all the
practices from first contact to signing them up, which often is 23 contacts in
between. First contract to where they are in the program to intermediate
milestones to goal. We track those and we get reports generated in terms of how
we are doing in terms of our slippage, and where the practices are in the
implementation time frame.
Once practices go live, we have a series of utilization reports that are
baked into the product, and we would be happy to share the logic behind those.
I’m sure they could be improved a lot. We look at things like, of all unique
patients seen, how many had unstructured allergies, how many had structured
allergies, how many had blood pressure recorded. We are at 98 percent now, of
our practices are recording blood pressure in the right way, in a structured
way that is machine interpretable in the record. Our problem is being used in a
Then we are monitoring their actual use of e-prescribing, the use of the
system around specific issues that we think are going to be important for the
next element, which is the quality reporting. We are now starting to get to
quality reports that are enabled by the raw data created coming in. When there
is an aberration, then we know that someone is not using the system
MS. JOHNSON: I don’t disagree with anything that the other speakers have
said, but I think there are two things that you need to be conscious of. Some
of the lack of use of the path has been because some of the technology in the
past has been useless. For an individual physician in this office, if the
electronic record looks nice on his screen but only tells him what he already
knows about his patient, this is useless.
I saw a demonstration recently that presented itself as advanced. Every
doctor had a picture of the patient’s record, and they all showed that patients
hadn’t had a mammogram. So I asked the question, what if Doctor A says to the
patient, you don’t have a mammogram, the patient goes and gets the mammogram
and Dr. A checks that off, what happens? Nothing. The other doctors still have
it on their record. So it is an incentive for all eight doctors.
This issue of interconnectedness, it would be too easy to focus our efforts
on making you accountable for the record. I don’t want to simplify this so
much. Everything you say about holding providers accountable for use I
appreciate. But we have to also hold patients accountable, and we have to hold
physicians accountable for the conversation that is supposed to take place
between them. Sometimes that conversation will not result in action.
So it isn’t as if we know exactly how do to this. That is important to
recognize from the beginning. But uselessness, we need to find a way to punish.
It may look very good, but if it doesn’t function well, somehow we have to
figure that out.
I think in all fairness to physicians, and also to avoid penalizing a
physician for taking a very complicated patient who frankly is unable to
comply, we need to do some of the kinds of things you are beginning to see
happening in home monitoring. They have to think newly about how you are going
to talk to these patients, how you are going to educate them, how you are going
to bring their family, how are you going to help the daughter see.
It is amazing. We have seen this years ago with retarded kids, we are
seeing this with homeless people. It is a different way of teaching, and then
you can change behavior. One of the more bizarre things in federal law is that
after six months, the home health agency — remember, we are into episodes of
care; in an episode of care, if you focus on changing the behavior of that
family group, the caregiver and the patient, and you can only follow them for
six months, in six months you do not tie in behavior change. But the law makes
it illegal for you to keep checking, even though the cost of persisting in that
relationship is practically zero.
So I think we need to think about how do we maintain relationships and how
do we change behaviors, as well as how do we make sure that you comport the
concrete entities that we have traditionally relied on to judge value.
MR. REYNOLDS: On that note, that ends our panel. I know there were
questions, but to maintain relationships we have open mike next. We have some
people listed to talk. So thank you very much. We really appreciate it.
Would Jonathan Hauten, Tom Leary, Andrea Pennington please begin lining up
at the mikes. We would like your comments written if you don’t get to speak,
but we will go ahead and begin with John Hauten.
MR. HAUGTON: Hi. I am John Haugton. I trained at work as an engineer and a
physician. I am a bone cancer survivor, was treated at the NIH as a teenager,
and currently sit on the Clinical Advisory Council. As well, I have been a
caregiver for aging parents.
At this point, my personal health record, it was important from 30 years,
in eight inches of manila folders is an Xray, an aureomycin dose and an
ejection fraction. Currently I run a company called Dockside that for a number
of years has offered Internet data, interoperable registries and clinic care
decision support systems that work in paper and electronic offices and
demonstrably have improved care.
In 2008 we became a clinical data submitter for the CMS PQRI program, and
successfully put 2300 physicians through the program. We found PQRI offered
physicians a simple low risk way to start engaging with meaningful use of IT
while getting paid for it.
Last year Farzad testified in front of Congress that CCHIT certified
systems generally lacked four criteria that improve care. Did not reliably
collect structured data, did not have a registry function, did not create
population performance metrics, and lacked decision support alerts for best
practices. I think you have heard some emphasis of that. You can get through
ARRA has the right language and the right elements. Kawamoto and Lovac
showed that decision support within work flow is what ends up improving care.
If you guys can emphasize the what of improved care around the meaningful use,
it will happen. You can get people to do it quickly, you can make these changes
quickly. PQRI from the time the law was signed to the time that the data
collection period was closed was 12 months.
MR. REYNOLDS: Thank you very much. You had two minutes plus two pages of
written if you want to submit it. I want to make sure everybody continually
understands that. Tom Leary.
MR. LEARY: Chairman Reynolds and members of the NCVHS Executive
Subcommittee, I am Tom Leary. I am the Senior Director of Federal Affairs for
HIMS, the Health Care Information and Management Systems Society. I want to
take one moment to welcome Dr. Blumenthal. HIMS looks forward to working with
As you know, we have 20,000 individual members, over 350 corporate members,
and 47 chapters in North America. So we would like to leverage our subject
matter expertise and resources through this discussion.
To that end, HIMS over the last four weeks has convened several different
meetings, and has used our website to gather information around meaningful use
and meaningful users, two definitions. We submitted those via letter and have
it posted on our website.
Four overarching recommendations I just wanted to highlight here, and then
we will submit in the two-page testimony. One, adopting the Certification
Commission for Health IT as the certifying body. Two, achieving incremental
maturation in a three-phase process that is outlined in our white papers.
Three, utilizing HITSP in integrating the health care enterprise to bridge the
interoperability gaps. Four, reconciling certified EHRs with best of breed in
open source technologies.
We look forward to the continued discussion. Thank you, sir.
MR. REYNOLDS: Thank you. While Dr. Pennington is getting ready, if Timothy
Stettheimer would also step to the mike.
DR. PENNINGTON: Hello. I am Dr. Andrea Pennington, a physician here in
Maryland and the D.C. area in private practice, Chief Medical Officer for
Logical Images, a provider of diagnostic decision support software. Thank you
for the opportunity to speak, and welcome, Dr. Blumenthal.
When defining meaningful use, I think you must ask meaningful to whom. A
fundamental question I think remains to be answered here today, that is, how
will HIT affect the digital divide, what does meaningful use mean for the
chronically underserved patient populations such as the racial and ethnic
minorities and families with a household income below the poverty threshold,
and the practitioners like myself that serve them both in urban and rural
settings. Dr. Cullen and others touched on this today. I respectfully suggest
that additional focus be directed here.
As physicians, we are treating a more ethnically diverse population than
ever before. Yet this diversity is often forgotten by our health care system.
To that point raised by Gina Perez earlier, what concerns me most is the
widespread consensus that minorities as well as other underserved populations
and their caregivers have largely been absent from policy and decision making
platforms. So as a starting point, any definition of meaningful use should
include direct input and involvement by the very group who will be affected by
it most and need it most. Such comprehensive input must include equal focus on
empowering minority and other underserved populations.
Finally, I have just three key areas for your consideration in addition to
MR. REYNOLDS: We are down to less than a minute.
DR. PENNINGTON: Okay. Health informatics experts. One of the under reported
problems in our health care system is that of diagnostic error. While
medication error and surgical errors get a great deal of attention in the
media, diagnostic error is an order of magnitude greater. So as we are looking
at EHRs and other health IT resources, I would ask that we focus on collecting
that data in a closed feedback loop.
MR. REYNOLDS: Thank you. Then if Gary Christoph would also come to the
mike, please. Dr. Stettheimer, please.
DR. STETTHEIMER: Thank you. Mr. Chairman, Dr. Blumenthal, committee, I
appreciate being able to be here and just hear all the comments that we have
been making today.
This is a very serious issue obviously that we are struggling with. I am
the Chief Information Officer for St. Vincent’s Health System in Birmingham,
Alabama. It is part of Ascension Health, which is the largest not-for-profit
health system in this country. I am also a board of trustees member for CHIME,
the College of Health Care Information Management Executives, which represents
more than 1300 CIOs in this country.
I really only have one question and one suggestion that I want to make. I
will endorse the statement for the record that has been made for CHIME, that
has been submitted. However, the question I want to ask is, who will we leave
behind. I think we have hinted at this throughout the day, but I think we have
to ask throughout these deliberations which patients, which physicians, which
hospitals will we leave behind, as we set bars and create hurdles and make
definitions. I believe if we leave any behind, then we are not doing what we
I would suggest that as we look at meaningful use, the thing to consider is
not just technologies, because that is not what this is about. It is about
information. It is also about the evidence based practices we use. We should
reward movement along this path, regardless of how it is made, and not lock
into particular technologies, but instead be looking for progress in health
outcomes for this century. So again, my last question is, who will we leave
MR. REYNOLDS: Thank you very much. Gary Christoph, please.
DR. CHRISTOPH: Gary Christoph, Dr. Christoph, Northrup Grumman. Welcome,
Dr. Blumenthal, and thank you for the opportunity to speak.
In building a definition of meaningful use, I want to say that it is
critically important to insure enforcement. Financial incentives alone are
insufficient to drive adoption and use of interoperable EHRs. Only with good
enforcement will trust by the public and providers come.
I say this as the first CIO of CMS and as the father of the HIPAA security
reg. I also served for the last four years on the HISPC, the technical advisory
panel, a collaboration of states trying for years to get to interstate health
information exchange. Interoperability, actionable exchange of data, will not
happen by standards alone. It is very hard work, and it means people work.
For example, hospitals using the same certified EHR in the same town cannot
today exchange health information. Oregon Health State University and the
Kaiser Hospital down the street cannot exchange the data, despite both being on
the same certified EHR.
Congresswoman Johnson said, fix the process. This is very apt, because it
is implementation of systems, provides and configurations that cause these
certified products to not interoperate.
MR. REYNOLDS: Thirty second, sir, please.
DR. STETTHEIMER: As Rachel Block stated, we need meaningful use not just
certified products, but certified processes and certified processes means that
people must be devoted to enforcement and audit of these processes.
I urge the National Committee on Vital and Health Statistics to insure that
a meaningful use in the definition, that it also include definitions of
MR. REYNOLDS: Kristen Richardson, you will close off the hearing for the
MS. RICHARDSON: Good afternoon. My name is Kristin Richardson with Rudolph
Severr. I am speaking today on behalf of Elsavir, a world leading publisher of
scientific, technical and medical information products and services. I commend
the committee for holding a forum on the important topic of meaningful use for
health IT adoption.
We know that in order for health IT to be meaningful, it must be more than
simply automation. It must integrate evidence based content and clinical
support tools that will help enhance the quality of care of patients across the
country. Health information technology that is meaningful will require a design
based on a systems thinking framework and firmly grounded in a professional
provide model that provides standardization of practice and guidance based on
the latest evidence, and that supports professional processes of diagnosis,
assessment, planning, treatment, interventions, evaluation and education.
We must also take an interdisciplinary approach in designing health
information technology where all clinicians integrate their practice and
documentation so that we stop wasteful duplication and fragmentation. This is
no longer just a theoretical concept. It is now a best practice available
today, and patient outcomes are most impressive.
On behalf of Elsevier, I appreciate the opportunity to share our views on
this important topic. We look forward to working with the members of this
committee and other stakeholders in developing solutions that will help improve
the quality of health care in this country.
MR. REYNOLDS: Thank you. At the end of a long day, I will remind everybody
that tomorrow we are in the Virginia Suite, which is across the hall. Please
take your personal belongings. If you leave them here, they are at your own
The other thing I would like to say is, I would like to thank everyone,
remind everybody about written comments, please submit them. Dr. Blumenthal,
thank you for being with us the whole day.
The last thing is I would like the committee members to at least touch base
at some point between now and tomorrow with Margret. We are working through
some themes. Tomorrow afternoon I am going to ask you to participate in one of
those themes, because we have got a lot of written testimony and we are going
to have to pull out those themes and other things out of them. So I need you to
do that so that tomorrow you are ready to help us make sure we make this
Thank you very much to all, and the hearing is adjourned.
(Whereupon, the meeting was recessed at 5:30 p.m., to reconvene Wednesday,
April 29, 2009 at 9:00 a.m.)