Watson Health 50 Top Cardiovascular Hospitals Study, 2020 - November 18, 2019
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
November 18, 2019 Watson Health 50 Top Cardiovascular Hospitals Study, 2020
IBM Watson Health™ 75 Binney Street Cambridge, MA 02142 800-525-9083 ibm.com/watsonhealth Watson Health 50 Top Cardiovascular Hospitals Study, 2020; 21st edition © 2019 IBM Watson Health. All rights reserved. IBM, the IBM logo, ibm.com, Watson Health, and 100 Top Hospitals are trademarks of International Business Machines Corp., registered in many jurisdictions worldwide. Other product and service names might be trademarks of IBM or other companies. The information contained in this publication is intended to serve as a guide for general comparisons and evaluations, but not as the sole basis upon which any specific conduct is to be recommended or undertaken. The reader bears sole risk and responsibility for any analysis, interpretation, or conclusion based on the information contained in this publication, and IBM shall not be responsible for any errors, misstatements, inaccuracies, or omissions contained herein. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or by any information storage and retrieval system, without permission in writing from IBM Watson Health.
Contents Introduction
Each year, IBM Watson Health™ conducts objective,
03 Introduction quantitative research to shine a light on the nation’s
highest-performing hospitals, health systems, and
09 This year’s winners
cardiovascular service lines, through the Watson
11 Findings Health 100 Top Hospitals® program. The goal of the
23 Methodology program is to deliver unbiased, guiding insights that
can help all healthcare organizations focus their
37 Appendix: Methodology details improvement initiatives and move toward consistent,
54 Footnotes sustainable top performance. Organizations do not
apply to participate in the study, and award winners
do not pay to market their honor.
Now, as we have since 1998, we have analyzed
public data with our proprietary methodologies to
provide the industry with this year’s Watson Health
50 Top Cardiovascular Hospitals study.
Like all 100 Top Hospitals program research, this
study creates a balanced scorecard of metrics to
identify top-performing cardiovascular providers in
the United States.
But the study is far more than a list. Since our
cardiovascular award winners have achieved an
outstanding balance of clinical and operational
excellence in a complex and changing landscape, we
believe their success can help provide a clear and
bright path for others to follow.
The information contained in our 50 Top
Cardiovascular Hospitals study is designed to put
impartial, action-driving, and attainable benchmarks
in the spotlight for hospital and clinical leaders
across the country to leverage as they work to raise
their own organizations’ standards of performance.
3The Watson Health 50 Top Illuminating achievement for a value-based
Cardiovascular Hospitals study focuses world
on one of healthcare’s most important By finding ways to take clinical and operational
service lines that affects hundreds of performance to the next level, the winners of our 50
Top Cardiovascular Hospitals study are identifying
thousands of patients’ lives annually
opportunities to deliver healthcare value to patients,
and adds billions of dollars to our communities, and payers.
nation’s overall healthcare costs.
Repeatedly, we see that these hospitals lead the
cardiovascular healthcare industry, often inspiring
the clinicians and staff within their own walls and
That’s why publishing new and systems, as well as their peers and competitors, to
achievable benchmarks for better understand data and benchmarks, and close
performance gaps.
cardiovascular service line performance
is important and has the potential to It is a kind of leadership that is perhaps becoming
make a large and lasting impact on the even more important as the industry continues to
transition to a value-based payment environment.
quality and cost of care for heart
patients across the US.
Why cardiovascular hospitals?
A 2018 report from the American Heart Association
states that about 92.1 million American adults are
living with some form of cardiovascular disease or
the after-effects of stroke1. Cardiovascular diseases
have a significant impact on mortality and cost,
accounting for about 2,300 Americans dying each
day, an average of 1 death every 38 seconds and
costs the United States about $200 billion each
year2. In addition, prevalence of cardiovascular
disease is expected to increase to a point where
approximately 40 percent of the US population will
have the disease by 20303.
It is no wonder, then, that cardiovascular services are
among the highest-profile service lines in healthcare.
With the stakes so high, it is important that hospitals
provide high-quality, highly efficient cardiac care and
that they look for ways to improve. The 50 Top
Cardiovascular Hospitals study attempts to answer
that need each year.
4The 50 Top Cardiovascular Hospitals study is also A measure of leadership excellence and its
unique for the 100 Top Hospitals program. The effect on service line performance
program’s research series publishes only this one
clinical service line study. Only the cardiovascular For more than 20 years, the 100 Top Hospitals
service line has consistently had both the inpatient program has collaborated with academics on a wide
volume and supplemental clinical process metrics range of topics to dig deeper into the leadership
from the Centers for Medicare & Medicaid Services practices of the nation’s top healthcare
(CMS) Hospital Compare initiative to support the organizations.
publication of scorecard-based benchmarks for a As such, the 100 Top Hospitals studies not only
service line. And with each annual 50 Top provide a distinctive approach to measuring the
Cardiovascular Hospitals study, more is learned, as performance of hospitals, health systems, and
the transparency and depth of inpatient and cardiovascular service lines, but also deliver insights
continuum-of-care data grow and evolve. into the effectiveness of hospital leadership. Higher
composite scores on our national balanced scorecard
reflect the effectiveness of the leadership team in
Objective, real-world assessment executing both short-term and long-term strategies
across the organization.
To maintain the study’s level of integrity, only public
data sources are used for calculating study metrics. The leadership of today’s hospitals, including the
This helps eliminate bias while including as many board, executive team, and medical staff leadership,
hospitals as possible, and facilitates consistency of is responsible for ensuring all facets of a hospital and
definitions and data. In turn, this allows us to its cardiovascular service line are performing at the
produce national norms and benchmarks that are same high level. The 50 Top Cardiovascular Hospitals
useful for assessing clinical outcomes and study and analytics provide a view of that enterprise
operational efficiency in an objective, independent, performance alignment. And that information can be
and meaningful way. In addition, we report rate of helpful in assessing the strategic intersection among
improvement compared to peers, which enables cost, quality, efficiency, and community value.
clinical leadership and service line management to
determine their real-world progress toward
consistent top performance within and across the The performance of this year’s
cardiovascular patient groups profiled. 50 Top Cardiovascular Hospitals
The 50 Top Cardiovascular Hospitals study identifies
US hospitals that have achieved the highest
performance on a balanced scorecard of
performance measures.
This year, based on comparisons between the study
winners and a peer group of similar hospitals that
were not winners, we found that our study winners
delivered better outcomes while operating more
efficiently and at a lower cost.
5Compared to nonwinning cardiovascular hospitals, We based this analysis on the Medicare patients
this year’s winners had: included in this study. If the same standards were
applied to all inpatients, the impact would be even
• Significantly higher inpatient survival (28.7% to greater.
47.4% higher)
• Fewer patients with complications (6.3% to 27.7%
fewer) Trends in cardiovascular care
• Higher 30-day survival rates for acute myocardial An analysis of trends in cardiovascular care over the
infarction (AMI), and coronary artery bypass five years ending in this study’s data year, revealed:
grafting (CABG) patients (0.3 to 0.5 percentage • Readmission rates for AMI and CABG patients
points higher)* showed statistically significant improvement in a
• Lower readmission rates for AMI, HF, and CABG large percentage of hospitals (38.2% and 49.9%
improvement, respectively)
patients (0.5 to 0.8 percentage points lower)
• HF wage- and severity-adjusted average cost per
• Average lengths of stay (ALOS) for CABG patients
case also showed statistically significant
that were over one day lower than nonwinners and
improvement for a large percentage of hospitals
on average a half a day lower for AMI, HF, and
(32.2%), which was more than twice the
percutaneous coronary intervention (PCI)**
improvement than in the other patient groups
• $1,485 to $6,704 less in total costs per patient
• As in our prior study, a majority of hospitals (from
case (the smallest dollar-amount difference was 67% to 86%) continue to hold the cost of
for HF, and the largest was for CABG) delivering care to AMI, HF, CABG, and PCI patients
• Lower average 30-day episode of care payments stable from 2014 to 2018, with no statistically
for AMI and HF ($1,323 and $706 less per significant change, at 95% confidence
episode, respectively)
Further, our study indicated that if all cardiovascular Additional findings
hospitals performed at the same level of this year’s
winners: For more details about the 50 Top Cardiovascular
Hospitals study findings, including complete hospital
• More than 11,000 additional lives could be saved reporting data on this year’s winning cardiovascular
hospitals, please see the Findings section of this
• More than 2,800 heart patients could be
document.
complication-free
• Over $1.5 billion could be saved
* An AMI is a heart attack, which happens when the arteries leading to the heart become blocked and blood supply is slowed or stopped. Heart failure is a
weakening of the heart's pumping power, leading to the body not receiving enough oxygen and nutrients to work properly. A CABG is a type of surgery that
improves blood flow to the heart by moving or redirecting a blood vessel to bypass blockages.
** A PCI is a procedure that uses a small stent to open up blood vessels in the heart that have narrowed from a buildup of plaque.
6Study integrity The value of 50 Top
Cardiovascular Hospitals
Organizations are included in the 100 Top Hospitals® benchmarks
program studies based solely on availability of data
from Medicare and meeting criteria listed in the
methodology section, without regard to whether they • To improve performance,
cardiovascular hospital leaders need
are a client. They do not apply to be included in the
objective information about what is
studies, nor do winners pay to promote their award.
achievable. They need relevant
benchmarks that allow them to
To uphold the integrity of the study, it is the policy of
compare their performance to peers
IBM Watson Health to revoke a 100 Top Hospitals
and top-performing organizations.
award if hospital data is found to be inaccurate or
misleading for any 100 Top Hospitals data source. • By naming the 50 Top
Cardiovascular Hospitals in the
At the sole discretion of IBM Watson Health, the nation, the 100 Top Hospitals
circumstances under which a 100 Top Hospitals program provides hospital
executives, physicians, and
award could be revoked include, but are not limited
cardiovascular service line managers
to:
with practical targets for raising
• Inaccurate data performance.
• Agency investigations or sanctions • Information in this study, and in
separate hospital-specific reports,
provides performance levels to
We welcome your input reach for, with detailed analysis of
Since 1993, the 100 Top Hospitals program has how the winners and their
worked to ensure that the measures and nonwinning peers performed on the
methodologies used are fair, consistent, and study’s balanced scorecard of
meaningful. We continually test the validity of our measures.
performance measures and data sources.
In addition, as part of our internal performance
improvement process, we welcome comments about
our study from health systems, hospitals, and
physicians. To submit comments, visit the Contact Us
section of 100tophospitals.com.
7More about the 100 Top Hospitals program About IBM Watson Health
The 50 Top Cardiovascular Hospitals research is one Watson Health aspires to improve lives and give hope
of several studies of the Watson Health 100 Top by delivering innovation to address the world’s most
Hospitals program. To increase understanding of pressing health challenges through data and
trends in specific areas of the healthcare industry, the cognitive insights.
program includes a range of studies and reports:
Each day, professionals make powerful progress
• 100 Top Hospitals and Everest Award studies
toward a healthier future. In an industry that is
Research that annually recognizes the 100 top- fragmented and complex, there are many
rated hospitals in the nation based on a opportunities to support professionals as they work
proprietary, balanced scorecard of overall toward their goals to simplify, solve, care or cure, so
organizational performance and also identifies they can transform health for the people they serve.
those hospitals that excel at long-term rates of
At Watson Health, we see and work across the health
improvement in addition to performance.
landscape, from payers and providers to government
• 50 Top Cardiovascular Hospitals study and life sciences. With an unrivaled vantage point
An annual study identifying hospitals that across the industry, deep health expertise, and the
demonstrate the highest performance in hospital power of cognitive computing, we create intelligent
cardiovascular services. connections that shape new ways of working, drive
value, and accelerate breakthroughs.
• 15 Top Health Systems study
An annual study introduced in 2009 that provides With Watson Health at work in their organizations,
an objective measure of health system our clients can uncover, connect, and act on the
performance overall and offers insight into the insights that advance their work, and change the
ability of a system’s member hospitals to deliver world.
consistent top performance across the For more information about IBM Watson Health, visit
communities they serve, all based on our national ibm.com/ watson/health.
health system scorecard.
• 100 Top Hospitals Performance Matrix
A two-dimensional analysis, available for nearly all
US hospitals, that provides a view of how long-
term improvement and resultant current
performance compare with national peers.
• Custom benchmark reports
A variety of reports designed to help healthcare
executives understand how their organizational
performance compares to peers within health
systems, states, and markets.
You can read more about these studies and see lists
of all winners by visiting 100tophospitals.com.
8The Watson Health 50 Top Teaching hospitals with cardiovascular residency programs*
Medicare
Cardiovascular Hospitals, 2020 Hospitals Location
ID
Beaumont Hospital - Troy Troy, MI 230269
The Watson Health™ 100 Top Hospitals® program is Guthrie Robert Packer Hospital Sayre, PA 390079
pleased to present the 2020 Watson Health 50 Top Huntington Hospital Pasadena, CA 050438
Kettering Medical Center Kettering, OH 360079
Cardiovascular Hospitals.
Lahey Hospital & Medical Center Burlington, MA 220171
Lankenau Medical Center Wynnewood, PA 390195
We stratified winners by three separate peer groups:
Mayo Clinic Hospital - Saint Marys Hospital Rochester, MN 240010
teaching hospitals with cardiovascular residency OhioHealth Riverside Methodist Hospital Columbus, OH 360006
programs, teaching hospitals without cardiovascular St. Vincent Indianapolis Hospital Indianapolis, IN 150084
Summa Akron City Hospital Akron, OH 360020
residency** programs, and community hospitals
UAB Hospital Birmingham, AL 010033
UNC Rex Healthcare Raleigh, NC 340114
Please note that the order of hospitals in the
University Hospital Madison, WI 520098
following tables does not reflect performance rating. Virtua Our Lady of Lourdes Hospital Camden, NJ 310029
Hospitals are ordered alphabetically. For full details WakeMed Raleigh Campus Raleigh, NC 340069
on these peer groups and the process we used to
select the winning benchmark hospitals, please see
the Methodology section of this document.
* Order of hospitals does not reflect performance rating. Hospitals are ordered alphabetically.
** Throughout this document where we refer to ‘cardiovascular residency programs,’ we are including cardiovascular fellowship programs as well.
Please refer to the Methodology section of this document for a complete list of cardiovascular residency and fellowship programs that are used to
classify hospitals.
9Teaching hospitals without cardiovascular residency programs*
Medicare
Hospitals Location
ID
Aspirus Wausau Hospital Wausau, WI 520030
Baylor Scott & White Medical Center - Hillcrest Waco, TX 450101
CHRISTUS St. Michael Health System Texarkana, TX 450801
Halifax Health Medical Center Daytona Beach, FL 100017
Holston Valley Medical Center
Kingsport, TN 440017
Lee Memorial Hospital/HealthPark Medical Center
Fort Myers, FL 100012
(LMH/HPMC)
MacNeal Hospital Berwyn, IL 140054
Mercy General Hospital Sacramento, CA 050017
MercyOne Des Moines Medical Center Des Moines, IA 160083
Mission Hospital Asheville, NC 340002
MultiCare Tacome General Hospital Tacoma, WA 500129
North Mississippi Medical Center Tupelo, MS 250004
PIH Health Hospital - Whittier Whittier, CA 050169
Sacred Heart Hospital Pensacola, FL 100025
St. Joseph Mercy Ann Arbor Ann Arbor, MI 230156
Community hospitals*
Medicare
Hospitals Location
ID
Asante Rogue Regional Medical Center Medford, OR 380018
Baylor Scott & White Medical Center - Round Rock Round Rock, TX 670034
Bellin Health Green Bay, WI 520049
Carolinas Medical Center Mercy-Pineville Charlotte, NC 340098
Columbus Regional Hospital Columbus, IN 150112
Doylestown Hospital Doylestown, PA 390203
Eastern Idaho Regional Medical Center Idaho Falls, ID 130018
Harlingen Medical Center Harlingen, TX 450855
Fredericksburg,
Mary Washington Hospital 490022
VA
McLaren Northern Michigan Hospital Petoskey, MI 230105
Saint Mary's Regional Medical Center Reno, NV 290009
Salem Hospital Salem, OR 380051
San Antonio Regional Hospital Upland, CA 050099
Sentara RMH Medical Center Harrisonburg, VA 490004
Shasta Regional Medical Center Redding, CA 050764
St. David's Medical Center Austin, TX 450431
St. Vincent Heart Center of Indiana Indianapolis, IN 150153
Thibodaux Regional Medical Center Thibodaux, LA 190004
University of Maryland St. Joseph Medical Center Towson, MD 210063
Wake Forest Baptist Health High Point Medical
High Point, NC 340004
Center
* Order of hospitals does not reflect performance rating. Hospitals are ordered alphabetically.
10Findings • The 2020 cardiovascular study winners had
27.7% and 6.3% lower complications observed-
This year’s Watson Health™ 50 Top Cardiovascular to-expected index values for CABG and PCI,
Hospitals provided better clinical care and were more respectively, when compared to their peers.
efficient than their peers. If all United States
• Long-term outcomes were better at winning
hospitals’ cardiovascular service lines performed at
hospitals, with the exception of Heart Failure (HF)
the level of these study winners, more than 11,000 30-day mortality, for which winning hospitals had
additional lives and over $1.5 billion could be saved, a slightly higher median rate than nonwinners
and nearly 2,800 additional bypass and angioplasty (11.4% v. 11.3%).
patients could be complication-free.
• 30-day heart attack (AMI) and CABG mortality
These findings are based on the Medicare patient rates were all lower among winning hospitals than
data included in this study and analysis of study peers, meaning a smaller percentage of patients
winners versus nonwinners. If the same standards died, of any cause, within 30 days after inpatient
were applied broadly to all inpatients, the impact admission. The difference was greatest among
AMI patients, with a 30-day mortality rate of
would be even greater.
12.2% for winners versus 12.7% for nonwinners.
One of the goals of the Watson Health 100 Top
• The winning hospitals also had lower readmission
Hospitals® program is to provide action-driving
rates, with a smaller percentage of patients
benchmarks that can help all hospitals improve their returning to the hospital, for any cause, within 30
performance. This section highlights winner days of discharge. AMI and CABG patient
(benchmark) versus nonwinner differences in all readmissions showed the same difference, with
study hospitals as a group and by hospital type rates of 14.9% and 12%, respectively, which was
(residency program and teaching status). nearly a full percentage point better than
nonwinning peers.
Benchmark hospitals outperformed peers
• Winning hospitals were more efficient, releasing
Comparisons between this year’s 50 Top patients sooner than their peers. The typical
Cardiovascular Hospitals and their peers showed that winning hospital released CABG patients more
room for improvement still exists (See Table 1). than a full day (1.2) earlier, and in this year’s study
results, AMI, HF, and PCI patients were released
• Survival rates were markedly better at benchmark
0.5 – 0.6 days sooner than patients getting care at
(winning) hospitals, particularly for patients nonwinning peers.
receiving coronary artery bypass graft surgeries
(CABGs). The median benchmark hospital had a • The 50 Top Cardiovascular Hospitals maintained
risk-adjusted CABG inpatient mortality index of high clinical performance while keeping inpatient
0.5, meaning there were 50% fewer deaths than costs lower. The typical winning hospital spent
about $6,700 less per CABG patient and $2,900
would be expected, given patient severity. With an
less per PCI case.
index of 0.95, peer (nonwinning) hospitals had
only 5% fewer CABG deaths than expected. • Benchmark hospitals also showed stronger
performance on measures of total Medicare claims
• Notably, in the 2020 CABG patient group we also
payment across 30-day episodes of care for AMI
saw the most pronounced difference in severity- and HF patients ($1,323 less per AMI episode and
and wage-adjusted cost per case, with winners $706 less per HF episode) compared to
having an average cost of $35,197, versus peers at nonwinning peers.
$41,901 – a difference of over $6,000.
11Table 1: National performance comparisons (all hospitals in study)
Benchmark compared with peer group
Benchmark Peer
Performance measure How winning benchmark
median median Percent
Difference hospitals outperform
difference
nonwinning peer hospitals
AMI mortality 0.72 1.01 -0.29 -28.7 Lower mortality
HF mortality 0.67 1.01 -0.34 -33.7 Lower mortality
Risk-adjusted mortality
index
CABG mortality 0.50 0.95 -0.45 -47.4 Lower mortality
Clinical outcome
measuresa PCI mortality 0.68 1.00 -0.32 -32.0 Lower mortality
CABG complications 0.68 0.94 -0.26 -27.7 Fewer complications
Risk-adjusted
complications index
PCI complications 0.89 0.95 -0.06 -6.3 Fewer complications
Clinical process CABG patients with internal mammary artery (IMA)
97.8 95.9 2.0 n/a Higher IMA use
measuresa,c use (%)
AMI 30-day mortality (%) 12.2 12.7 -0.5 n/a Lower 30-day mortality
HF 30-day mortality (%) 11.4 11.3 0.1 n/a Higher 30-day mortality
CABG 30-day mortality (%) 2.7 3.0 -0.3 n/a Lower 30-day mortality
Extended outcome
measuresb,c AMI 30-day readmission (%) 14.9 15.7 -0.8 n/a Fewer 30-day readmissions
HF 30-day readmission (%) 20.8 21.3 -0.5 n/a Fewer 30-day readmissions
CABG 30-day readmission (%) 12.0 12.8 -0.8 n/a Fewer 30-day readmissions
AMI severity-adjusted average length of stay (ALOS) 3.7 4.1 -0.5 -11.2 Shorter ALOS
Process efficiency HF severity-adjusted ALOS 4.3 4.9 -0.6 -11.8 Shorter ALOS
CABG severity-adjusted ALOS 8.1 9.2 -1.2 -12.6 Shorter ALOS
PCI severity-adjusted ALOS 3.2 3.6 -0.5 -12.5 Shorter ALOS
AMI wage- and severity-adjusted average cost per
$8,339 $9,974 -$1,635 -16.4 Lower cost per case
case
HF wage- and severity-adjusted average cost per
$8,055 $9,540 -$1,485 -15.6 Lower cost per case
case
Cost efficiency
CABG wage- and severity-adjusted average cost per
$35,197 $41,901 -$6,704 -16.0 Lower cost per case
case
PCI wage- and severity-adjusted average cost per
$15,511 $18,432 -$2,921 -15.8 Lower cost per case
case
AMI 30-day episode payment $23,671 $24,994 -$1,323.50 -5.3
Lower 30-day payment
Extended efficiency HF 30-day episdoe payment $17,079 $17,785 -$706.50 -4.0
measuresb Lower 30-day payment
a. Medicare Provider Analysis and Review (MEDPAR) 2017 and 2018, combined
b. CMS Hospital Compare July 1, 2015 – June 30, 2018
c. We do not calculate percentage difference for measures already expressed as a percent
12Better performance at benchmark teaching • Unlike last year’s results, winners performed
hospitals with cardiovascular residency better than the nonwinning group on all 30-day
programs mortality and readmission measures, with the
greatest difference in the HF 30-day readmission
Teaching hospitals with specialized cardiovascular rate (1.0 percentage point lower, 20.5% versus
residency and fellowship programs are generally 21.5%). The difference between winners and
believed to treat more complex patients, have a more national peers on the CABG 30-day readmission
complex staffing mix, and incur higher costs than measure was similar, at 0.8 percentage points
community hospitals and those without specific lower (11.9% versus 12.7%).
cardiovascular teaching programs. Evaluating
performance among teaching hospitals with • Medicare 30-day episode payment measures
cardiovascular programs as a unique group helps to showed AMI patients at winning cardiovascular
produce valid quantitative comparisons. (See Table teaching hospitals having the greatest difference
2.) between winner and nonwinner performance
among all three comparison groups. At winning
• Continuing to set the standard bar at a very high cardiovascular teaching hospitals, 30-day AMI
mark, cardiovascular teaching winners’ inpatient payments were 5.9% less than those at nonwinner
mortality rates were 54% and 34% lower than peer hospitals ($23,480 versus $24,962).
peers for CABG and PCI patients, respectively.
• These benchmark hospitals were also leaders for
treating PCI patients (6.5% fewer complications
than peers). However their superior performance
did not extend to CABG patients where nonwinning
hospitals outperformed with 2% fewer
complications.
• Cardiovascular teaching benchmark hospitals were
also much more efficient than their peers, with
severity-adjusted costs among all patient groups
being on average almost 14% lower than costs
calculated for peer facilities. The greatest absolute
difference in cost was found for CABG patients at
$4,196 less per bypass surgery case. In addition,
winners had 13% lower cost per case for HF
patients and about 17% lower cost for both AMI
and PCI inpatients.
13Table 2: Performance comparisons for teaching hospitals with cardiovascular residency programs
Benchmark compared with peer group
Benchmark Peer
Performance measure How winning benchmark
median median
Difference Percent difference hospitals outperform
nonwinning peer hospitals
AMI mortality 0.76 1.01 -0.25 -24.8 Lower mortality
Risk-adjusted HF mortality 0.82 0.97 -0.15 -15.5 Lower mortality
mortality index CABG mortality 0.44 0.96 -0.52 -54.2 Lower mortality
Clinical outcome
measuresa PCI mortality 0.68 1.03 -0.35 -34.0 Lower mortality
Risk-adjusted CABG complications 0.95 0.93 0.02 2.2 More complications
complications index
PCI complications 0.87 0.93 -0.06 -6.5 Fewer complications
Clinical process CABG patients with internal mammary artery
97.9 96.8 1.1 n/a Higher IMA use
measuresa,c (IMA) use (%)
AMI 30-day mortality (%) 11.8 12.4 -0.6 n/a Lower 30-day mortality
HF 30-day mortality (%) 10.1 10.5 -0.4 n/a Lower 30-day mortality
CABG 30-day mortality (%) 2.4 2.8 -0.4 n/a Lower 30-day mortality
Extended outcome
Fewer 30-day
measuresb,c AMI 30-day readmission (%) 15.0 15.7 -0.7 n/a
readmissions
Fewer 30-day
HF 30-day readmission (%) 20.5 21.5 -1.0 n/a
readmissions
Fewer 30-day
CABG 30-day readmission (%) 11.9 12.7 -0.8 n/a
readmissions
AMI severity-adjusted average length of stay
3.7 4.1 -0.4 -10.5 Shorter ALOS
(ALOS)
HF severity-adjusted ALOS 4.3 5.0 -0.7 -13.8 Shorter ALOS
Process efficiency
CABG severity-adjusted ALOS 8.5 9.1 -0.6 -7.0 Shorter ALOS
PCI severity-adjusted ALOS 3.1 3.7 -0.6 -15.8 Shorter ALOS
AMI wage- and severity-adjusted average cost
$8,130 $9,775 -$1,646 -16.8 Lower cost per case
per case
HF wage- and severity-adjusted average cost
$8,302 $9,529 -$1,228 -12.9 Lower cost per case
per case
Cost efficiency
CABG wage- and severity-adjusted average
$37,154 $41,350 -$4,196 -10.1 Lower cost per case
cost per case
PCI wage- and severity-adjusted average cost
$15,668 $18,837 -$3,169 -16.8 Lower cost per case
per case
AMI 30-day episode payment $23,480 $24,962 -$1,482 -5.9
Lower 30-day payment
Extended efficiency
measuresb HF 30-day episdoe payment $17,104 $17,892 -$788 -4.4
Lower 30-day payment
a. Medicare Provider Analysis and Review (MEDPAR) 2017 and 2018, combined
b. CMS Hospital Compare July 1, 2015 – June 30, 2018
c. We do not calculate percentage difference for measures already expressed as a percent
14Better performance at benchmark teaching • The greatest difference between winning and
hospitals without cardiovascular teaching nonwinning hospitals in the extended outcome
programs measures was found in the 30-day readmission
measure for CABG patients where there was
Winning teaching hospitals without cardiovascular nearly a full percentage point difference (0.9) with
teaching programs were much more efficient than
rates of 11.8% versus 12.7%.
their peers, with large differences found in a number
of measures. (See Table 3.) • On the Medicare 30-day episode payment
measures, winning teaching hospitals without
• This difference was most notable in the inpatient
cardiovascular residency programs outperformed
mortality measure across all patient groups, with
their peers, with lower median AMI and HF 30-day
HF and CABG showing the greatest differences payment values (1.1% and 3.2% lower,
between winning and nonwinning hospitals: HF respectively).
with 43.6% fewer deaths and CABG with 52%
fewer. There were two measures in which nonwinning
hospitals outperformed the winning hospitals: PCI
• These benchmark hospitals also treated AMI, HF,
complications and HF 30-day mortality. Most
CABG, and PCI cases at a lower cost, 19.2%, noticeable is the difference found in performance on
28.8%, 20.1% and 14.5% less, respectively,
the PCI complications index: Peer hospitals did
saving $8,473 per CABG case and $1,967 per AMI better than benchmark hospitals by a margin of 26%,
case.
with a median index value of 0.94 compared to 1.19
• Most 30-day extended outcome measures were among winners.
also better at winning teaching hospitals without
cardiovascular teaching programs, with winners
having median AMI and CABG 30-day mortality
rates 0.4 and 0.5 percentage points lower than
those of peers (AMI – 12.5% v. 12.9%, CABG –
2.6% v. 3.1%).
15Table 3: Performance comparisons for teaching hospitals without cardiovascular residency programs
Benchmark compared with peer group
Benchmark Peer How winning
Performance measure benchmark hospitals
median median
Difference Percent difference
outperform nonwinning
peer hospitals
AMI mortality 0.73 1.01 -0.28 -27.7 Lower mortality
Risk-adjusted HF mortality 0.57 1.01 -0.44 -43.6 Lower mortality
mortality index CABG mortality 0.48 1.00 -0.52 -52.0 Lower mortality
Clinical outcome
measuresa PCI mortality 0.65 0.97 -0.32 -33.0 Lower mortality
CABG complications 0.92 0.94 -0.02 -2.1 Fewer complications
Risk-adjusted
complications index PCI complications 1.19 0.94 0.25 26.6 More complications
Clinical process CABG patients with internal mammary artery
97.4 95.8 1.7 n/a Higher IMA use
measuresa,c (IMA) use (%)
AMI 30-day mortality (%) 12.5 12.9 -0.4 n/a Lower 30-day mortality
HF 30-day mortality (%) 11.6 11.5 0.1 n/a Higher 30-day mortality
CABG 30-day mortality (%) 2.6 3.1 -0.5 n/a Lower 30-day mortality
Extended outcome
Fewer 30-day
measuresb,c AMI 30-day readmission (%) 14.9 15.5 -0.6 n/a
readmissions
Fewer 30-day
HF 30-day readmission (%) 21.2 21.3 -0.1 n/a
readmissions
Fewer 30-day
CABG 30-day readmission (%) 11.8 12.7 -0.9 n/a
readmissions
AMI severity-adjusted average length of stay
3.8 4.2 -0.5 -10.7 Shorter ALOS
(ALOS)
HF severity-adjusted ALOS 4.4 5.0 -0.6 -12.2 Shorter ALOS
Process efficiency
CABG severity-adjusted ALOS 8.1 9.3 -1.2 -12.4 Shorter ALOS
PCI severity-adjusted ALOS 3.1 3.6 -0.5 -13.9 Shorter ALOS
AMI wage- and severity-adjusted average cost
$8,267 $10,234 -$1,967 -19.2 Lower cost per case
per case
HF wage- and severity-adjusted average cost
$6,998 $9,834 -$2,836 -28.8 Lower cost per case
per case
Cost efficiency
CABG wage- and severity-adjusted average
$33,656 $42,129 -$8,473 -20.1 Lower cost per case
cost per case
PCI wage- and severity-adjusted average cost
$15,646 $18,310 -$2,663 -14.5 Lower cost per case
per case
AMI 30-day episode payment $24,430 $24,691 -$261 -1.1
Lower 30-day payment
Extended efficiency
HF 30-day episdoe payment $17,137 $17,696 -$559 -3.2
measuresb Lower 30-day payment
a. Medicare Provider Analysis and Review (MEDPAR) 2017 and 2018, combined
b. CMS Hospital Compare July 1, 2015 – June 30, 2018
c. We do not calculate percentage difference for measures already expressed as a percent
16Better performance at benchmark community
hospitals
Benchmark community hospitals again outperformed
their peers on the inpatient risk-adjusted mortality
measure, with the most observable performance
difference being in the CABG patient group. Winning
community hospitals had a median risk-adjusted
mortality index value of 0.50, compared to the
median index value of 0.91 at peer hospitals (a
45.1% gap, with fewer patients dying at hospitals
named winners). (See Table 4.)
• The winning community hospitals were also much
more efficient than their peers. They discharged
CABG patients almost a day and a half sooner
(1.4), HF patients almost a day sooner (0.9) and
AMI patients a half a day sooner (0.5).
• Cost-per-case medians in all patient groups were
also much lower for benchmark community
hospitals, with the largest difference, in both
absolute and percentage terms (23% less), being
observed for CABG patients, at $9,642 less per
case than peer hospitals.
• Notably, benchmark community hospitals had the
most marked contrast in performance between
winning and nonwinning hospitals on the risk-
adjusted complications measure for both CABG
and PCI patients, with a median observed-to-
expected ratio for winners at 0.60, versus 0.99 for
nonwinning peers in the CABG patient group (a
difference of 39%), and a median observed-to-
expected ratio for winners at 0.67, versus 0.98 for
nonwinners in the PCI patient group (a difference
of 31%).
17Table 4: Performance comparisons for community hospitals
Benchmark compared with peer group
Benchmark Peer How winning
Performance measure benchmark hospitals
median median
Difference Percent difference
outperform nonwinning
peer hospitals
AMI mortality 0.69 0.98 -0.29 -29.6 Lower mortality
Risk-adjusted HF mortality 0.66 1.02 -0.36 -35.3 Lower mortality
mortality index CABG mortality 0.50 0.91 -0.41 -45.1 Lower mortality
Clinical outcome
measuresa PCI mortality 0.73 1.00 -0.27 -27.0 Lower mortality
CABG complications 0.60 0.99 -0.39 -39.4 Fewer complications
Risk-adjusted
complications index PCI complications 0.67 0.98 -0.31 -31.6 Fewer complications
Clinical process CABG patients with internal mammary artery
98.1 95.5 2.6 n/a Higher IMA use
measuresa,c (IMA) use (%)
AMI 30-day mortality (%) 12.3 12.7 -0.4 n/a Lower 30-day mortality
HF 30-day mortality (%) 11.3 11.6 -0.4 n/a Lower 30-day mortality
CABG 30-day mortality (%) 2.8 3.1 -0.4 n/a Lower 30-day mortality
Extended outcome
measuresb,c AMI 30-day readmission (%) 14.8 15.6 -0.9 n/a Fewer 30-day readmissions
HF 30-day readmission (%) 20.4 21.3 -0.9 n/a Fewer 30-day readmissions
CABG 30-day readmission (%) 12.4 13.0 -0.6 n/a Fewer 30-day readmissions
AMI severity-adjusted average length of stay
3.7 4.1 -0.5 -11.1 Shorter ALOS
(ALOS)
HF severity-adjusted ALOS 4.2 5.0 -0.9 -17.5 Shorter ALOS
Process efficiency
CABG severity-adjusted ALOS 7.9 9.3 -1.4 -14.8 Shorter ALOS
PCI severity-adjusted ALOS 3.3 3.6 -0.3 -8.9 Shorter ALOS
AMI wage- and severity-adjusted average cost
$8,392 $10,033 -$1,641 -16.4 Lower cost per case
per case
HF wage- and severity-adjusted average cost
$8,158 $9,574 -$1,415 -14.8 Lower cost per case
per case
Cost efficiency
CABG wage- and severity-adjusted average cost
$32,295 $41,937 -$9,642 -23.0 Lower cost per case
per case
PCI wage- and severity-adjusted average cost
$15,291 $18,126 -$2,835 -15.6 Lower cost per case
per case
AMI 30-day episode payment $23,825 $25,179 -$1,354 -5.4 Lower 30-day payment
Extended efficiency
HF 30-day episdoe payment $16,913 $17,794 -$881 -5.0
measuresb Lower 30-day payment
a. Medicare Provider Analysis and Review (MEDPAR) 2017 and 2018, combined
b. CMS Hospital Compare July 1, 2015 – June 30, 2018
c. We do not calculate percentage difference for measures already expressed as a percent
18Additional measures for informational The measures report the difference (“excess”)
purposes between each hospital’s average days in acute care
(“predicted days”) and the number of days in acute
Every year, we publish new measures that may be of
care that each hospital’s patients would have been
interest to the leaders of hospitals and health
expected to spend if discharged from an average-
systems. For this study edition, we continue to
performing hospital (“expected days”).
publish 30-day excess days in acute care (EDAC)
measures for AMI and HF patients. These The measure is reported as excess days per 100
performance measures, along with the existing discharges.
ranked extended care measures, 30-day mortality,
Comparing benchmark hospitals and peers on this
readmission and episode of payment, offer health
measure yields interesting results, as shown in Table
care leaders an additional insight into the
5 on the following page.
performance of hospitals across the continuum of
care. If you would like to provide feedback on these • The benchmark median EDAC score for AMI
informational measures, please email patients was 9.4 days less than the peer EDAC
100tophospitals@us.ibm.com. score, at -3.7 versus 5.7 for nonwinning hospitals.
30-day excess days in acute care (heart attack [AMI] • The benchmark median EDAC score for HF
and heart failure [HF]) patients was 9.8 days less than the peer EDAC
score, at -1.5 versus 8.3 for nonwinning hospitals.
In this study, we have profiled performance, for
information only, on the relatively new Centers for
Medicare & Medicaid Services (CMS) excess days in
acute care (EDAC) measures:
1. 30-day EDAC for AMI patients
2. 30-day EDAC for HF patients
As defined by CMS5, the EDAC measures capture
excess days that a hospital’s patients spent in acute
care within 30 days after discharge. These measures
summarize the number of risk-adjusted days a
hospital’s patients spend in an emergency
department (ED), a hospital observation unit, or a
hospital inpatient unit during 30 days following a
hospitalization for AMI or HF.
19Table 5: National performance comparisons for excess days in acute care (all hospitals in study)
Benchmark compared with peer group
Benchma
Peer
Performance measure rk How benchmark
median Differenc
median Percent difference hospitals differ from
e
peer hospitals
AMI 30-day excess days in acute
-3.4 6.0 -9.4 n/a Fewer days in acute care
carec
a,b
Extended efficiency measures
HF 30-day excess days in acute
-1.5 8.3 -9.8 n/a Fewer days in acute care
carec
a. CMS Hospital Compare July 1, 2015 − June 30, 2018
b. We do not calculate percentage difference for measures already expressed as a percent
c. Reported as excess days per 100 discharges
Trends in cardiovascular care • A healthy proportion of hospitals significantly
improved their 30-day readmission rates for AMI
Again in this edition of the 50 Top Cardiovascular
and CABG (38.2% and 49.9 %, respectively).
Hospitals study, we are presenting new findings on
trends in cardiovascular care delivered in the nation’s • 34.5% of all hospitals significantly improved on
teaching and community hospitals. Our intent is to the AMI 30-day mortality measure, between
provide healthcare leaders with new insights by 2014-2018.
showing the direction and magnitude of change in
• In 32.2% of all hospitals, heart failure cost per
key cardiovascular care performance indicators,
case statistically decreased, between 2014- 2018.
between 2014 and 2018.
However, there were a few notable declines in
All measures are being trended using five (5) years of
performance over time that should also be pointed
data, with data periods ending in 2014, 2015, 2016,
out. (See the right gray column in Table 6):
2017, and 2018.
• PCI cost per case statistically increased in 15.3%
of all hospitals.
Performance improvement over time: All hospitals
• AMI and HF 30-day episode of payment
By studying the direction of performance change of statistically worsened, or increased, in 50.8% and
all hospitals eligible for our study (winners and 52.2%, respectively, of all in-study hospitals.
nonwinners), we can see that US hospitals have not
been able to significantly improve performance
across the entire 50 Top Cardiovascular Hospitals
balanced scorecard: In the majority of measures (17
of 23), 71% or more of all in-study hospitals saw no
statistically significant change on any of the
scorecard measures. (See Table 6.) However, over
the years we studied there were a few notable
performance improvements for specific measures,
especially those extending beyond the acute
inpatient stay. (See the green left column in Table 6.)
20Table 6: Direction of performance change for all cardiovascular hospitals in study, 2014 - 2018
Significantly improving No statistically significant Significantly declining
performance change in performance performance
Performance measure
Count of Percent of Count of Percent of Count of Percent of
hospitals1 hospitals2 hospitals1 hospitals2 hospitals1 hospitals2
AMI mortality 10 1.0% 899 94.2% 45 4.7%
HF mortality 10 1.0% 901 94.4% 43 4.5%
Risk-adjusted mortality index
CABG mortality 19 2.0% 923 96.8% 12 1.3%
PCI mortality 16 1.7% 905 94.9% 33 3.5%
CABG complications 30 3.1% 895 93.8% 29 3.0%
Risk-adjusted complications index
PCI complications 15 1.6% 896 93.9% 43 4.5%
CABG patients with internal mammary artery (IMA) use 143 15.0% 730 76.8% 78 8.2%
AMI 30-day mortality 329 34.5% 607 63.6% 18 1.9%
HF 30-day mortality 139 14.6% 745 78.1% 70 7.3%
CABG 30-day mortality 129 13.5% 753 79.0% 71 7.5%
AMI 30-day readmission 364 38.2% 567 59.4% 23 2.4%
HF 30-day readmission 178 18.7% 684 71.7% 92 9.6%
CABG 30-day readmission 476 49.9% 472 49.5% 6 0.6%
AMI severity-adjusted average length of stay (ALOS) 47 4.9% 871 91.3% 36 3.8%
HF severity-adjusted ALOS 136 14.3% 780 81.8% 38 4.0%
CABG severity-adjusted ALOS 44 4.6% 857 89.8% 53 5.6%
PCI severity-adjusted ALOS 156 16.4% 794 83.2% 4 0.4%
AMI wage- and severity-adjusted average cost per case 122 12.8% 818 86.0% 11 1.2%
HF wage- and severity-adjusted average cost per case 306 32.2% 637 67.1% 7 0.7%
CABG wage- and severity-adjusted average cost per case 85 9.0% 800 84.3% 64 6.7%
PCI wage- and severity-adjusted average cost per case 10 1.1% 795 83.7% 145 15.3%
AMI 30-day episode of payment 8 0.8% 461 48.3% 485 50.8%
HF 30-day episode of payment 11 1.2% 445 46.6% 498 52.2%
1. Count refers to the number of in-study hospitals whose performance fell into the highlighted categor for the measure.
2. Percent is calculated by dividing the 'count' by the total in-study hospitals cacross all comparison groups.
Note: Total number of hospitals included in the analysis can vary by measure due to exclusion of IQR outlier data points, causing some in-study hospitals to have too few
remaining data points to calculate trend. This affects the Cost per Case measures.
21This page intentionally left blank.
22Methodology
Building the database of hospitals
The Watson Health™ 50 Top Cardiovascular Hospitals
study is based on quantitative research that uses a Like all Watson Health 100 Top Hospitals® studies,
the 50 Top Cardiovascular Hospitals study uses only
balanced scorecard approach, based on publicly
publicly available data. The data come from:
available data, to identify the top cardiovascular
hospitals in the US. This study focuses on short-term, • Medicare Provider Analysis and Review (MEDPAR)
acute care, nonfederal US hospitals that treat a broad data set
spectrum of cardiology patients. It includes patients
requiring medical management, as well as those who • Centers for Medicare & Medicaid Services (CMS)
receive invasive or surgical procedures. Because Hospital Compare data set
multiple measures are used, a hospital must provide • Medicare Cost Reports
all forms of cardiovascular care, including open heart
surgery, to be included in the study. We use MEDPAR patient-level record information to
calculate inpatient mortality, complications, and
length of stay (LOS). MEDPAR is also used for patient-
Overview level charge data in estimating average cost per case.
This data set contains information on approximately
The main steps used in the selection of the 50 Top 15 million Medicare patients who are discharged
Cardiovascular Hospitals study winners are: from the nation’s acute care hospitals annually.
• Building the database of hospitals, including Six years of MEDPAR data are used to develop the
special selection and exclusion criteria study trend database The two most recent years of
MEDPAR data available are used to identify current
• Classifying hospitals into comparison groups
performance and to select the winning hospitals. To
• Scoring hospitals on a set of weighted be included in the study, a hospital must have both
performance measures years of data available, with valid present-on-
admission (POA) coding.
• Determining the 50 hospitals with the best
overall performance by ranking relative to like We use Medicare Cost Reports to create our
comparison groups proprietary database, which contains hospital-
specific demographic information and hospital-
The following section is intended to be an overview of
specific, all-payer cost and charge data. The hospital
these steps. To request more detailed information on
cost-to-charge ratios are applied to MEDPAR patient-
any of the study concepts outlined here, please email
level claims data to estimate cost for the study’s cost
us at 100tophospitals@us.ibm.com or call 800-525-
measures. This is done at the cost-center and
9083.
charge-code levels for each patient record
23The Medicare Cost Report is filed annually by every We reference residency and fellowship program
US hospital that participates in the Medicare information from the Accreditation Council for
program. Hospitals are required to submit cost Graduate Medical Education (ACGME) and the
reports to receive reimbursement from Medicare. It American Osteopathic Association (AOA) to classify
should be noted, however, that cost report data teaching hospitals. Participation in a cardiovascular
includes services for all patients, not just Medicare fellowship program is identified and confirmed using
beneficiaries. the sources listed below.
• Electronic Residency Application Services (ERAS),
The 100 Top Hospitals program and many others in a program of the Association of American Medical
the healthcare industry have used the MEDPAR and Colleges (AAMC)
Medicare Cost Report databases for years. We • ACGME website
believe they are accurate and reliable sources for the • AOA Office of Graduate Medical Education (OGME)
types of analyses performed in this study. Medicare website
data is highly representative of the cardiovascular • Medical college websites
patients included in this study. In fact, Medicare • Hospital websites
inpatients usually represent about two-thirds of all
patients undergoing medical treatment for acute
myocardial infarction (AMI) or experiencing heart Time periods of data
failure (HF), and about half of all patients undergoing
percutaneous coronary intervention (PCI) or coronary The following table identifies the years used in this
artery bypass grafting (CABG), as found in the study for both the current and trend profiles.
Watson Health Projected Inpatient Database (PIDB)*. References made to ‘current’ year, ‘most current’
year and ‘trend’ years throughout this overview are
We use the CMS Hospital Compare data set for 30- defined below:
day mortality and 30-day readmission rate
performance measures, as well as the 30-day Table 7. Time Periods
episode-of-care payment measures for AMI and HF References in text Time Periods
patients. CMS publishes these rates as three-year Federal fiscal years (FFY) Oct - Sept
combined data values. Five data points are used to Study Year 2020
develop the study trend database for these extended
Current data year (MEDPAR) FFY 2018
care measures. We label these data points based on
Current data year (Medicare Cost Reports) Year ending in 2018
the end year of each data set. For example, July 1,
Two most current/recent years of data
2015-June 30, 2018 is named “2018.” We used the (MEDPAR / Medicare Cost Reports)
2017 and 2018
current year (most recent data set available) to
Trend data years (MEDPAR) FFY 2013 - 2018
identify current performance and to select the
Years ending in 2014 -
winning hospitals. Trend data years (Medicare Cost Reports)
2018
PIDB data used in risk model development FFY 2015
* The Watson Health Projected Inpatient Database (PIDB) is one of the largest US inpatient, all-payer databases of its kind, containing more than 23
million all-payer discharges annually. This data is obtained from approximately 5,000 hospitals, representing over 65% of all discharges from short-
term, general, nonfederal hospitals in the US.
24Present-on-admission data The effect of present-on-
admission data on risk and
Our risk-adjustment models for inpatient mortality severity adjustment
and complications, and severity-adjustment models
• Since 2008, CMS regulations
for LOS and cost per case included POA data reported
have required all Inpatient
in the MEDPAR data sets. Under the Deficit Reduction
Prospective Payment System
Act of 2005, as of federal fiscal year (FFY) 2008,
hospitals to document whether a
hospitals receive a reduced payment for cases with
patient has certain conditions
certain conditions (such as falls, surgical-site
when admitted; these are coded
infections, and pressure ulcers) that were not present
as POA.
on the patient’s admission but occurred during
hospitalization. As a result, CMS now requires all • Our complication rate
Inpatient Prospective Payment System hospitals to methodology uses this POA data.
document whether a patient has these conditions Consequently, the complication
when admitted6. rates exclude “false-positive”
complications and are more
Present-on-admission coding adjustments
accurate. In addition, our
Since 2010 we have observed a significant rise in the mortality, complications, LOS,
number of principal diagnosis (PDX) and secondary and cost-per-case risk- and
diagnosis (SDX) codes that do not have a valid POA severity- adjustment models
indicator code in the MEDPAR data files. Since 2011, develop expected values based
an invalid code of “0” has been appearing. This only on conditions that were
phenomenon has led to an artificial rise in the present on admission.
number of complications that appear to be occurring
during the hospital stay. See the Appendix for details.
To correct for this bias, we adjust MEDPAR record
processing through our inpatient mortality and
complications risk models, and LOS and cost- per-
case severity-adjustment models, as follows:
• We treat all principal diagnoses as present on
admission
• We treat all diagnosis codes on the CMS exempt
list as “exempt,” regardless of POA coding
• We treat secondary diagnoses where POA
indicator codes “Y” or “W” appeared more than
50% of the time in the all-payer database as
present on admission
25Hospitals and patient groups included Hospitals excluded
The focus of the study is on hospitals that offer both After building the database of cardiovascular hospitals,
medical and surgical treatment options for patients we exclude hospitals that reasonably might be
with two of the most common cardiovascular expected to include a different patient population or
conditions: AMI and HF. To build such a database, we population distribution, or whose data is not
included all hospitals that had, in the two most recent sufficient for analysis.
data years combined, at least 30 unique cases7 in
Excluded from the study were:
each of the four patient groups described below.
• Hospitals with fewer than 30 unique patient
• AMI patients – restricted to nonsurgical patients
records in each patient group (AMI, HF, CABG,
• HF patients – restricted to nonsurgical patients and PCI) for the two most current MEDPAR years
• CABG patients – includes all ICD-9-CM and ICD- combined
10-CM procedure codes, principal or secondary • Specialty hospitals, other than cardiac hospitals
in MS-DRGs 231 - 236 (critical access hospitals, children’s, women’s,
• PCI patients – excludes patients with open chest psychiatric, substance abuse, rehabilitation, and
coronary artery angioplasty long-term acute care hospitals)
Each patient group is mutually exclusive, by design. • Federally owned hospitals
To define patient diagnoses, older years of MEDPAR • Non-continental US hospitals (such as those in
data files in the trend profile utilize ICD-9-CM and the Puerto Rico, Guam, and the Virgin Islands)
more current data files utilize ICD-10-CM. See the • Hospitals with Medicare average LOS (ALOS)
Appendix for patient group definitions and the code- longer than 30 days
level detail. • Hospitals with no reported deaths
Patient records excluded • Hospitals that did not have Medicare claims for
the two most current years of data
The AMI and HF groups explicitly exclude patients
• Hospitals missing data for calculation of one or
who also had a PCI and/or CABG procedure. This
more performance measures
helps ensure we have exclusively medical patients in
these groups. • Hospitals for which a Medicare Cost Report was
not available for the two most current years of
Also excluded: data
• Patients who were discharged to another short- • Hospitals that did not code POA indicators on
term facility (to avoid double-counting) the two most current years of MEDPAR data
• Patients who were not at least 65 years old
26Classifying hospitals into comparison groups Participation in a fellowship program was identified
and confirmed using the following sources:
Bed size, teaching status, and residency/fellowship
• ERAS (AAMC program)
program involvement have a significant effect on the
types of patients a hospital treats and the scope of • ACGME website
services it provides. When analyzing the performance • OGME website
of an individual hospital, it is crucial to evaluate it • Medical college websites
against similar hospitals. To address this, we assign
• Hospital websites
each hospital to one of three comparison groups
according to its teaching and residency program Teaching hospitals without cardiovascular residency
status. programs
Our formula for defining the cardiovascular hospital Hospitals in this category have no involvement in a
comparison groups includes each hospital’s bed size, cardiovascular residency program. There are two
residents-to-beds ratio, and involvement in graduate ways to qualify as a teaching hospital:
medical education (GME) programs accredited by
Meet two of the following three criteria:
either the ACGME8 or the AOA9. We define the groups
as follows. • 200 or more acute care beds in service
• An intern/resident-per-bed ratio of at least 0.03
Teaching hospitals with cardiovascular residency
• Involvement in at least 3 accredited GME
programs
programs overall
Hospitals in this category must meet the definition of
Or: have an intern/ resident-per-bed ratio of 0.25 or
teaching (see teaching hospitals without
greater, regardless of bed size or AHA survey data.
cardiovascular residency programs definition) and be
involved in a cardiovascular residency program Community hospitals
accredited by the ACGME or the AOA. Cardiovascular Hospital must meet both of the following criteria:
residency programs include any of the following: • 25 or more acute care beds in service
• Cardiology • Not classified as a teaching hospital per
definitions above
• Cardiovascular disease
Bed size and number of interns/residents (full- time
• Cardiovascular medicine equivalents) are taken from each hospital’s most
• Cardiothoracic surgery current Medicare Cost Report available.
• Interventional cardiology Cardiovascular study groups
• Clinical cardiac electrophysiology The final study group counts, after exclusions, are
• Thoracic surgery listed in Table 8:
• Thoracic surgery – integrated
Table 8. Cardiovascular hospital comparison groups
• Advanced heart failure and transplant Comparison group Total
cardiology Teaching hospitals with cardiovascular residency programs 275
• Adult congenital heart disease Teaching hospitals without cardiovascular residency programs 291
Community hospitals 423
Note: Cardiovascular radiology residency programs Total in-study hospitals 989
are not included.
27You can also read
