How A Regional Collaborative Of Hospitals And Physicians In Michigan Cut Costs And Improved The Quality Of Care

Article (PDF Available)inHealth Affairs 30(4):636-45 · April 2011with89 Reads
DOI: 10.1377/hlthaff.2010.0526 · Source: PubMed
Abstract
There is evidence that collaborations between hospitals and physicians in particular regions of the country have led to improvements in the quality of care. Even so, there have not been many of these collaborations. We review one, the Michigan regional collaborative improvement program, which was paid for by a large private insurer, has yielded improvements for a range of clinical conditions, and has reduced costs in several important areas. In general and vascular surgery alone, complications from surgery dropped almost 2.6 percent among participating Michigan hospitals-a change that translates into 2,500 fewer Michigan patients with surgical complications each year. Estimated annual savings from this one collaborative are approximately $20 million, far exceeding the cost of administering the program. Regional collaborative improvement programs should become increasingly attractive to hospitals and physicians, as well as to national policy makers, as they seek to improve health care quality and reduce costs.
By David A. Share, Darrell A. Campbell, Nancy Birkmeyer, Richard L. Prager, Hitinder S. Gurm,
Mauro Moscucci, Marianne Udow-Phillips, and John D. Birkmeyer
How A Regional Collaborative
Of Hospitals And Physicians
In Michigan Cut Costs And
Improved The Quality Of Care
ABSTRACT
There is evidence that collaborations between hospitals and
physicians in particular regions of the country have led to improvements
in the quality of care. Even so, there have not been many of these
collaborations. We review one, the Michigan regional collaborative
improvement program, which was paid for by a large private insurer, has
yielded improvements for a range of clinical conditions, and has reduced
costs in several important areas. In general and vascular surgery alone,
complications from surgery dropped almost 2.5 percent among
participating Michigan hospitalsa change that translates into 2,500
fewer Michigan patients with surgical complications each year. Estimated
annual savings from this one collaborative are approximately
$20 million, far exceeding the cost of administering the program.
Regional collaborative improvement programs should become
increasingly attractive to hospitals and physicians, as well as to national
policy makers, as they seek to improve health care quality and reduce
costs.
T
he need to improve quality of care
in US hospitals is widely recog-
nized. Potentially avoidable ad-
verse events are common among
hospitalized patients, and wide
variation in hospital performance outcomes sug-
gests that there is ample room for improve-
ment.
14
The business case for improving hospi-
tal quality is also apparent. In surgery, for
example, the true cost associated with treating
complications exceeds $10,000 per patient, the
large majority of which is passed on to payers
and purchasers.
5
Additional payments for com-
plicated hospital stays (outlier payments), un-
planned readmissions, and care following dis-
charge for patients with complications account
for approximately 20 percent of the total costs
associated with many inpatient procedures, ac-
cording to national Medicare data.
6
Background On Hospital Quality
Improvement
Despite increasing attention from payers, policy
makers, and professional organizations, large-
scale efforts to improve hospital quality have had
little effect on patient outcomes. Public report-
ing of performance data may motivate hospitals
to improve.
7
However, there remain doubts that
programs such as the Centers for Medicare and
Medicaid Services Hospital Compare website or
the Leapfrog Groups selective referral initiative
will be successful in redirecting large numbers of
patients to hospitals that have demonstrated
superior results.
810
Simply put, it hasnt been
demonstrated that patients will actually stop
going to hospitals that achieve poor results
and start going to hospitals that achieve far bet-
ter ones. Even if practical barriers to changing
these referral patterns could be addressed
such as efficient transfer of patients medical
doi: 10.1377/hlthaff.2010.0526
HEALTH AFFAIRS 3 0,
NO. 4 (2011):
©2011 Project HOPE
The People-to-People Health
Foundation, Inc.
David A. Share (dshare@
bcbsm.com) is the executive
medical director at Blue Cros s
andBlueShieldofMichigan,
in Ann Arbor.
Darrell A. Campbell is the
chief of clinical affairs in the
Department of Surger y,
University of Michigan, in Ann
Arbor.
Nancy Birkmeyer is an
associate professor in the
Department of Surger y,
University of Michigan.
Richard L. Prager is a
professor in the Department
of Surgery, University of
Michigan.
Hitinder S. Gurm is an
assistant professor in the
Department of Internal
Medicine, University of
Michigan.
Mauro Mosc ucci is a
professor of medicine in the
Cardiovascular Division of the
Miller School of Medicine,
University of Miami, in Coral
Gables, Florida.
Marianne Udow-Phillips is
director of the Center for
Healthcare Research and
Transformation, in Ann Arbor.
John D. Birk meyer is a
professor in the Department
of Surgery, University of
Michigan.
April 2011 30:4 Health Affairs 1
Quality Profiles
recordsthese initiatives are limited by a lack of
good data and measures for identifying truly
superior hospitals.
Varied Approaches In addition to not paying
for so-called never events, such as surgical pro-
cedures on the wrong site or when foreign ob-
jects are left inside a patient after surgery, both
public and private insurers have implemented
pay-for-performance programs aimed at increas-
ing the use of specific, evidence-based practices.
An example is ensuring that a patient is taking a
beta-blocker when discharged from the hospital
after a myocardial infarction, or heart attack.
11,12
Hospitals have generally improved their perfor-
mance with these process-of-care measures,
which are distinct from outcome measures that
indicate how the health status of patients has
changed. But whether such programs have con-
ferred clinically meaningful improvements in
patient outcomes is debatable.
1317
Regional Collaborations Regional collabo-
rations between hospitals and physicians may be
more effective than either selective referral or
pay-for-performance in improving the quality
of health care at the population level. Pioneered
by the Northern New England Cardiovascular
Disease Study Group, regional collaborative im-
provement programs are based upon clinical
registries containing detailed information about
patients risk status, processes of care, and out-
comes.
18
Hospitals and physicians receive regu-
lar and (usually) confidential feedback on their
performance from their registry coordinating
centerfor example, risk-adjusted mortality
rates for cardiac surgery. Hospital officials and
physicians convene regularly to review and in-
terpret their data, often focusing on areas of
variation in practice or outcomes. Best practices
are then identified and implemented across the
region, which may be an area within a large state
or a group of one or more states.
Despite the conceptual appeal of this model
and its success in northern New England, it
has not been widely adopted in other parts of
the United States. However, an ambitious pro-
gram in Michigan now provides the first oppor-
tunity to assess the value and practicality of
regional collaborative improvement programs
on a much larger scale.
After early success with a program focusing on
percutaneous coronary interventionscom-
monly known as heart angioplastiesBlue Cross
and Blue Shield of Michigan/Blue Care Network
decided to make regional collaborative improve-
ment a major component of its statewide Value
Partnership program. Then, in 2004, the insurer
began implementing similar programs in other
clinical areas.
19,20
This insurer currently invests
almost $30 million annually in nine programs,
which collectively focus on the care of almost
200,000 Michigan patients annually.
Five of the programsin breast cancer, cardiac
computed tomography, peripheral vascular in-
terventions, trauma care, and hospital-based
medical carehave not been established long
enough to enable the judging of results. How-
ever, results from the other four, more mature
regional collaborative improvement programs
targeting percutaneous coronary interventions,
cardiac surgery, bariatric surgery for obesity,
and other types of general and vascular
surgeryare now emerging.
Focus On Michigan We review the Michigan
regional collaborative improvement program
and its success to date in improving clinical out-
comes. Given the substantial cost of these im-
provement programs, we also consider savings
accrued to payers as a result of fewer adverse
outcomes or other efficiency gains and thus
the return on investment from the payer perspec-
tive. Finally, we review lessons learned from the
first five years of the Michigan program and po-
tential challenges associated with scaling up this
model nationwide.
Overview Of The Program
Participa nts Blue Cross and Blue Shield of
Michigan/Blue Care Network is the dominant
private insurer in Michigan, insuring approxi-
mately 47 percent of the ten million residents of
the state. Based on the assessment of the lead
author of this article, David Share, approxi-
mately 5 percent of its total reimbursements to
hospitals ($160 million annually) are currently
reserved for its Participating Hospital Agree-
ment Incentive Program. This program includes
elements of traditional pay-for-performance
plans. However, 20 percent of the programs
overall budget is devoted to nine regional col-
laborative improvement programs, whose an-
nual costs range from $1.2 million to more than
$5 million each, according to financial docu-
ments from fiscal year 2010.
Each regional collaborative improvement pro-
gram is administered by a coordinating center
staffed by one of the participating hospitals
(mostly university-based), not by Blue Cross
and Blue Shield of Michigan/Blue Care Network.
Although staff composition varies by program,
most coordinating centers have a physician-
director, program epidemiologist or statistician,
data analyst, data auditor, quality improvement
nurse, and administrative support.
Costs And Payments Based on financial re-
ports from fiscal year 2010, payments to hospi-
tals account for most of the costs of the regional
collaborative improvement programs. Hospitals
Quality Profiles
2 Health Affairs April 2011 30:4
are compensated for each improvement program
in which they participate, regardless of their per-
formance relative to other centers. Payment for-
mulas were originally designed to cover the di-
rect costs of participation, but they are now
based on a fixed percentage of each hospitals
total payments from Blue Cross and Blue Shield
of Michigan/Blue Care Network. In 2007 these
payments to hospitals ranged from $11,000 to
more than $1 million across the forty-four hos-
pitals participating in at least one regional col-
laborative improvement program.
For most hospitals, payments exceed the true
costs of participation, according to a financial
analysis conducted by John Birkmeyer, one of
this papers authors. Participating hospitals are
expected to collect and submit data to the pro-
gram registries on a timely basis and allow regu-
lar site visits from data auditors. To receive pay-
ments, hospitals must send at least one
physician-representative and a program co-
ordinator to the quarterly meetings of each
regional collaborative improvement program
and participate actively in statewide and hospi-
tal-specific quality improvement interventions.
Targeted Conditions The improvement pro-
grams target clinical conditions and procedures
that are relatively common and that are associ-
ated with high costs per episode. They also tend
to focus on procedures that are technically com-
plex, evolving rapidly, and associated with wide
variation in hospital practice and outcomes.
Although the programs all administer detailed
clinical registries, they vary in several aspects of
data collection and measurement (Exhibit 1).
Outcomes are measured using established na-
tional registries administered by professional
organizations, locally developed databases, or
some combination of the two.
Data To help hospitals target and monitor
their local improvement activities, all of the
regional collaborative improvement programs
provide participating hospitals with hospital-
and physician-specific outcome data, relative
to Michigan and (in some cases) national bench-
marks. These data are confidential and not acces-
sible by Blue Cross and Blue Shield of Michigan/
Blue Care Network. Although most of the pro-
grams focus on short-term morbidity and mor-
tality, some track longer-term measures of effec-
tiveness, such as weight loss and patients
functional status after bariatric surgery. Several
of the programs link to the insurers claims data
to track use of health care services and spending.
Clinical Improvements
The regional collaborative improvement pro-
grams vary widely with respect to their primary
outcome measures, risk-adjustment models and
statistical techniques, and use of external bench-
marks for assessing comparative improvements.
In general, however, the success of the programs
is judged by trends in statewide rates of use and
adverse outcomes, which are assessed for both
clinical and statistical significance. The latter is
determined by regression-based time-series
analyses, which adjust for any measurable
changes in patient characteristics over time.
General And Vascular Surgery The largest
of the regional collaborative improvement pro-
grams is the Michigan Surgical Quality Collabo-
rative, which targets general and vascular sur-
gery. Given the broad range of procedures
included in this program, it tends to focus its
quality improvement efforts on aspects of peri-
operative carecare before, during, and after
surgery that is common to almost any type of
inpatient surgery, including practices aimed at
preventing common complications such as sur-
Exhibit 1
Overview Of Four Regional Collaborative Improvement Programs In Michigan
Characteristic
Percutaneous
coronary interventions Cardiac surgery Bariatric surgery
Major general and
vascular surgery
Program start 1998 2006 2006 2005
Current number of hospitals
(percent eligible)
31 (100%) 33 (100%) 27 (96%) 34 (94%)
Approximate number of patients per
year
a
32,000 10,000 7,000 50,000
Cost to BCBSM/BCN per year $3.2 million $3.0 million $2.7 million $5.0 million
Registry Locally developed STS registry with
local enhancements
Locally developed ACS-NSQIP with
local enhancements
SOURCE Blue Cross and Blue Shield of Michigan. NOTES BCBSM/BCN is Blue Cross and Blue Shield of Michigan/Blue Care Network. STS is Society of Thoracic Surgeons.
ACS-NSQIP is American College of Surgeons National Surgical Quality Improvement Program. Although approximately 100,000 Michigan patients each year undergo
general and vascular procedures targeted by ACS-NSQIP, this registry collects data on a random subset.
a
Patients per most recent year (201 0).
April 2011 30:4 Health Affairs 3
gical site infection or venous thromboembolism
(a blood clot forming in a vein).
The Michigan Surgical Quality Collaborative
shares the same measurement platform as the
American College of Surgeons National Surgical
Quality Improvement Program. It collects addi-
tional data on selected procedures, including
colorectal surgery and lower-extremity revascu-
larization. The National Surgical Quality Im-
provement Program collects very detailed clini-
cal information about patient characteristics
(for purposes of risk adjustment) and postoper-
ative complications. Between 2005 and 2009 the
national program included approximately 200
hospitals nationwide, a group in which large
academic centers tend to be overrepresented.
Although it hosts an annual national meeting
where hospitals share their experiences and im-
provement work, the program does not itself
direct improvement interventions or coordinate
collaborations across hospitals.
To assess the added value of the regional col-
laborative improvement model, we used the Na-
tional Surgical Quality Improvement Program
registry to compare surgical outcomes in hospi-
tals within Michigan to those outside the state.
For the entire study period, Michigan patients
could be identified directly using the Michigan
Surgical Quality Collaborative database. Other
patients undergoing surgery between 2005
and 2007 could be identified directly from the
National Surgical Quality Improvement Pro-
gram public-use database.
In 200809, however, the public-use file no
longer contained hospital identifiers. For this
reason, we identified patients outside of Michi-
gan by using a matching algorithm based on
patient characteristics, primary procedure code,
and other variables. This algorithm matched
more than 95 percent of patients.
When comparing the performance of hospitals
in and outside of Michigan, we focused on thirty-
day morbidity rates, which is the primary out-
come measure of the National Surgical Quality
Improvement Program. To ensure fair compar-
isons between the two groups, morbidity rates
were adjusted for patients risk factors, including
preoperative albumin, creatinine, functional sta-
tus, sepsis, inpatient and emergency surgery sta-
tus, illness severity (using the American Society
of Anesthesiologists score), work relative value
units, and surgical specialty (peripheral vascular
versus general surgery).
In addition to cross-sectional comparisons, we
used logistic regression to assess time trends in
morbidity rates in both groups of hospitals after
adjusting for the above covariates. Relative im-
provements in outcomes between the Michigan
hospitals and the others were formally compared
using a likelihood ratio test for interaction be-
tween time and site (in Michigan versus not in
Michigan) based on the logistic regression
model. In essence, this analysis examined
whether the slopes in morbidity rate trends over
time were significantly different between hospi-
tals in Michigan and those not in Michigan.
As seen in Exhibit 2, risk-adjusted morbidity
rates in Michigan hospitals fell from 13.1 percent
in 2005 to 10.5 percent in 2009 (p<0:001). In
contrast, morbidity rates in hospitals outside of
Michigan participating in the National Surgical
Quality Improvement Program remained essen-
tially flat between 2005 and 2008, before dip-
ping slightly in 2009. Although trends toward
improvement in the two populations were both
statistically significant, improvement occurred
at a faster rate in Michigan hospitals
(p<0:001). In 2009 (the latest year for which
complete data were available), overall morbidity
in Michigan hospitals was significantly lower
than in the other hospitals (10.5 percent versus
11.5 percent, p<0:001).
Although hospital-specific morbidity rates are
less precise, some Michigan hospitals improved
more than others. Of the thirty-two hospitals
participating by the end of 2008, eight hospitals
(25 percent) showed statistically significant
(p<0:05) reductions in their morbidity rates
by the end of 2009. Another eight hospitals
(25 percent) had achieved trends toward declin-
ing morbidity (p<0:20). There were no signifi-
cant improvements in morbidity rates at the re-
maining hospitals.
Bariatric Surgery The Michigan Bariatric
Surgery Collaborative, which enrolls more than
95 percent of patients undergoing bariatric sur-
gery in the state, has to date focused its improve-
ment activities on reducing technical complica-
tions and rates of venous thromboembolism.
Overall complication rates declined from 8.7 per-
cent to 6.6 percent between 2007 (the first year
for which complete data were available) and
Although hospital-
specific morbidity
rat es a re le ss p rec is e,
some Michigan
hospitals improved
more than others.
Quality Profiles
4 Health Affairs April 2011 30:4
2009.
Because the Michigan Bariatric Surgery Col-
laborative and the National Surgical Quality Im-
provement Program rely on separate registries
with different outcome measures and defini-
tions, improvements in complication rates in
Michigan cannot be assessed against that na-
tional benchmark. However, we did compare
surgical mortality in our two hospital popula-
tions, adjusting for variables common to both
registries, including age, sex, body mass index,
and procedure type.
As seen in Exhibit 3, risk-adjusted thirty-day
mortality with bariatric surgery in Michigan hos-
pitals dropped significantly from 2007 to 2009
(p ¼ 0:004). Bariatric surgery mortality at hos-
pitals outside of Michigan participating in the
National Surgical Quality Improvement Pro-
gram also declined during the same time period,
although this improvement was not statistically
significant. Based on analysis of interaction
terms in the mortality model, the rate of im-
provement at Michigan hospitals exceeded that
of the other hospitals (p ¼ 0:045).
Interventional Cardiology The main out-
come measure of the percutaneous coronary in-
tervention program is not a single endpoint,
such as whether or not the patient dies, but
rather a so-called composite endpoint of serious
complications, including emergency coronary
artery bypass graft surgery, repeat of the pro-
cedure, stroke, and death. Between 1998 and
2002,seriouscomplicationsfell from 3.8 percent
to 2.3 percent among Michigan hospitals partici-
pating in the regional collaborative improve-
ment program (p<0:001).
21
In 2002, participat-
ing hospitals had substantially fewer serious
complications than Michigan hospitals not par-
ticipating at that time (2.3 percentversus 3.2 per-
cent, p<0:001), according to our analysis.
Those latter hospitals joined the program shortly
thereafter, and their outcomes have since caught
up to those of the original cohort.
Cardiac Surgery For coronary artery bypass
graft surgery, the regional collaborative im-
provement program rates hospital performance
in terms of an eleven-item composite quality
measure, which includes risk-adjusted mortal-
ity; complications; use of a section of the internal
mammary artery that serves the chest wall and
breasts as a graft; and several other important
processes of care, as defined by the Adult Cardiac
Surgery Registry of the Society of Thoracic Sur-
geons.
22
The Society of Thoracic Surgeons coor-
dinating center conducts most of the analyses for
the Michigan program and provides it with regu-
lar reports on hospital-specific and statewide
performance.
During its initial reporting periods (200607
and 200708), composite quality scores for
Michigan hospitals as a whole were statistically
indistinguishable from national benchmarks,
according to reports provided by the Society of
Thoracic Surgeons. By 200809, however, Mich-
igan hospitals as a whole had achieved a three-
star rating from the society, indicating that their
aggregate performance exceeded national
norms (with 99 percent probability) and fell
Exhibit 3
Thirty-Day Mortality After Bariatric Surgery: Hospitals In Michigan Versus Hospitals
Outside Of Michigan, 200709
Percent
Non-Michigan hospitals
Michigan hospitals
SOURCE M ichigan Surg ical Quality C ollaborative and Nationa l Surgical Qua lity Improvement Prog ram
registries, 200709. NOTES Thirty-day mortality rates declined faster in Mich igan hospital s than in
other hospitals participating in the National Surgical Quality Improvement Program (p ¼ 0:045) .
Exhibit 2
Risk-Adjusted Morbidity With General And Vascular Surgery: Hospitals In Michigan Versus
Hospitals Outside Of Michigan, 200509
Percent
Non-Michigan hospitals
Michigan hospitals
SOURCE Michigan Surgi cal Quality C ollaborative and Nation al Surgical Quality I mprovement Program
registries, 200509. NOTES Morbi dity rates de clined fas ter in Mi chigan hos pitals (p < 0:001) an d, by
2009, were lower th an in oth er hospitals participating in the Nati onal Surgical Quality Improvement
Program (p < 0:001).
April 2011 30:4 Health Affairs 5
within the top tenth percentile of hospitals na-
tionwide.
Return On Investment
The most persuasive return-on-investment
analysis of the regional collaborative improve-
ment programs would require linking the clini-
cal outcome registries to claims databases and
demonstrating the extent to which measured
improvements lead directly to less cost to insur-
ers. Although this work is ongoing, there is rea-
son to believe that the programs more than pay
for themselves.
For example, in general and vascular surgery
alone, the approximately 2.5 percent drop in
surgical morbidity rates observed by the Michi-
gan Surgical Quality Collaborative translates to
2,500 fewer Michigan patients with surgical
complications each year, based on our analyses.
One studywhich used resource-based cost ac-
counting methodsfound that the average cost
of such complications is $11,000, of which 75 per-
cent is passed along to insurers.
5
If these esti-
mates are correct, the Michigan Surgical Quality
Collaborative reduces payments associated with
adverse outcomes by approximately $20 million
annuallyfar exceeding the $5 million annual
cost of administering the program.
The business case for the regional collabora-
tive improvement programs can be made with far
less extrapolation. For example, in 2007 almost
10 percent of patients in Michigan hospitals
undergoing gastric bypass surgery received
inferior vena cava filters to prevent postopera-
tive pulmonary embolism. In this procedure, a
filter is placed in the large abdominal vein that
returns blood to the heart, in order to trap clot
fragments and prevent them from traveling
through the vein to the heart and lungs and caus-
ing blockage of circulation.
The use of these filters varied widely across
hospitals, from 0 percent to more than 40 per-
cent.
23
Six of the twenty-four hospitals were plac-
ing the large majority of the filters being placed
statewide. Analysis of outcome data from the
Michigan Bariatric Surgery Collaborative re-
vealed that the use of inferior vena cava filters
was not protective, but instead was associated
with markedly higher risks of serious complica-
tions, many of which were directly related to
complications from the filter itself. Following
feedback of this information to surgeons and
implementation of statewide guidelines, the
use of the filters dropped to fewer than 2 percent
of patients in a one-year period, according to
Michigan Bariatric Surgery Collaborative data.
The average payment associated with placing
the filter is $13,000 (in 2007 dollars), so this
single change in practice saves payers more than
$4 million annuallyconsiderably more than the
cost of administering the regional collaborative
improvement program in bariatric surgery.
Several other specific quality improvement in-
terventions have also generated substantial sav-
ings. The use of two very expensive therapies in
cardiac surgeryintra-aortic balloon pumps and
prolonged mechanical ventilationhas fallen
substantially.
22,24
Implementation of risk-predic-
tion tools and practice guidelines has reduced
the incidence of contrast-induced nephropathy
(acute kidney failure triggered by the use of con-
trast dye in the procedure) and the need for
dialysis after percutaneous coronary interven-
tion.
21
Between 2007 and 2009, rates of thirty-
day emergency department visits after bariatric
surgery fell from 8 percent to 5 percent, with
associated savings approaching $1 million an-
nually.
Lessons Learned And Challenges For
Dissemination
Hospitals have options for improving quality
and efficiency that do not require them to col-
laborate with competing hospitals and physi-
cians. Internal quality improvement activities
can include the implementation of protocols
and clinical pathways that reduce unwanted
variation and incorporate evidence-based prac-
tices and guidelines. Hospitals can also establish
checklists to minimize mistakes and improve
communication and teamwork among providers
and staff.
25,26
Unfortunately, although protocols and check-
lists help ensure that processes known to be ef-
fective (for example, timely administration of
perioperative antibiotics) are implemented,
such evidence-based practices represent only a
small proportion of the overall care delivered to
hospitalized patients. Such efforts do not teach
hospitals and physicians how to improve other
aspects of care.
Benefits Of Regional Collaboration Re-
sults from the Michigan initiative suggest that
hospitals participating in regional collaborative
improvement programs improve far more
quickly than they can on their own. Practice
variation across hospitals and surgeons creates
innumerable natural experiments for identify-
ing what works and what doesnt.
The large sample sizes and statistical power
associated with regional collaborative improve-
ment program registries allow for more robust,
rapid assessment of relationships between proc-
ess and outcomes and of the effects of quality
improvement interventions than can be
achieved by hospitals examining their own prac-
Quality Profiles
6 Health Affairs April 2011 30:4
tice in isolation. Although identification and im-
plementation of best practices are cornerstones
of the regional collaborative improvement
model, we believe that these programs also have
salutary but immeasurable effects on the local
safety culture. In our experience, participating
hospitals and physicians simply start paying
more attention to their practices and how to
improve them.
Differences Among Programs It is difficult
to identify which specific components of the
regional collaborative improvement model are
most important. Each program involves numer-
ous, concurrent interventions including perfor-
mance feedback, site visits, collaborative learn-
ing, and targeted interventions aimed at specific
clinical problems. Their cumulative effects are
not readily disentangled.
The programs also use different approaches to
identifying and disseminating best practices.
Some are more evidence based than others, rely-
ing primarily on empirical analyses that link
specific processes of care to clinical outcomes
data. Others place a greater emphasis on hospital
site visits and benchmarking, examining organi-
zational factors and safety culture as well as spe-
cific processes of care. The comparative effective-
ness of these different strategies is difficult to
assess.
We believe that improvements in Michigan
hospitals are largely attributable to the programs
themselves, not to trends toward improvement
occurring everywhere. First, many of the im-
provements in overall outcome measures can
be directly attributed to specific interventions
initiated by the programs. For example, our
analysis indicates that mortality rates associated
with bariatric surgery fell in large part because of
declining rates of fatal pulmonary embolism,
which were temporally related to statewide im-
plementation of a protocol for increased preven-
tion of this complication. Similar examples in-
clude the effects of comprehensive interventions
targeting surgical site infection in the Michigan
Surgical Quality Collaborative and contrast-
related nephropathy in percutaneous coronary
intervention.
Second, as described earlier, Michigan hospi-
tals had more substantial improvements in rates
of morbidity and mortality than other hospitals
participating in national data feedback pro-
grams administered by the Society of Thoracic
Surgeons and the American College of Surgeons.
Such data suggest that results in Michigan can-
not be attributed simply to secular trends toward
improving technical quality. Because most of the
regional collaborative improvement programs
are based on clinically detailed, well-validated
national outcomes registries, results in Michi-
gan cannot be attributed to differences in data
collection techniques or outcomes definitions.
It is also important to note that hospitals par-
ticipating in the Adult Cardiac Surgery Registry
of the Society of Thoracic Surgeons or the Ameri-
can College of Surgeons National Surgical Qual-
ity Improvement Program may represent a high
bar for purposes of benchmarking. These pro-
grams are voluntary and may attract hospitals
most committed to quality improvement. At least
with the National Surgical Quality Improvement
Program, large teaching centers are overrepre-
sented among participating hospitals and, based
on our own (unpublished) analyses of national
Medicare data, have notably lower surgical mor-
tality rates than nonparticipating US hospitals.
As currently implemented, the Michigan
regional collaborative improvement programs
are evaluated for their effect on cost and out-
comes in specific, clinically defined patient pop-
ulations, not for their cumulative effect on the
health of the entire population. Nonetheless, be-
cause these programs target clinical conditions
and procedures that are common, expensive,
and associated with substantial morbidity, we
believe that their benefits at the population level
would compare favorably to weaker interven-
tions aimed at much broader populations, such
as employee wellness programs and other pre-
ventive strategies.
Role Of Dominant Insurer Although suc-
cessful regional collaborative improvement pro-
grams do not necessarily require payer involve-
ment, the programs in Michigan would not have
occurred had the states largest private insurer
not underwritten their substantial costs, offered
additional financial incentives for hospitals to
participate, and provided a neutral meeting
ground for collaborating hospitals and physi-
cians. Although large private insurers are ob-
vious candidates for leading the dissemination
of regional collaborative improvement programs
nationwide, this model has challenges.
The insurer had the
confidence that
benefits would accrue
primarily to its
beneficiaries and
purchasers.
April 2011 30:4 Health Affairs 7
Given its dominant share of the private insur-
ance market in Michigan, Blue Cross and Blue
Shield of Michigan/Blue Care Network had the
leverage to urge hospitals to participate in the
programs and the confidence that benefits would
accrue primarily to its beneficiaries and purchas-
ers. Other states are similarly dominated by one
large insurer;
27
several, including Tennessee and
Florida, are implementing similar regional col-
laborative improvement programs. Although
private insurers have taken the lead so far,
regional collaborative improvement programs
could be similarly fostered by public payers or
regional coalitions of private payers, purchasers,
and provider systems.
Relevance F or National Efforts Evidence
that regional collaborative improvement pro-
grams can simultaneously improve quality and
reduce costs at the population level comes at an
opportune time. The regional collaborative im-
provement model is particularly relevant to the
interests of the Centers for Medicare and Medic-
aid Services as it begins to enact provisions of the
Affordable Care Act, including accountable care
organizations.
28
In that context, such programs
provide a robust data infrastructure for monitor-
ing quality as health systems work toward con-
straining their costs.
More important, such programs provide a
framework for facilitating improvement with re-
gard to both cost and quality domains. Regional
collaborative improvement programs should
also become increasingly attractive to hospitals
and physicians as they seek to improve quality
and reduce costs. As the Centers for Medicare
and Medicaid Services and other payers move
toward episode-based bundled payments for
inpatient surgery and other types of hospital-
based care,
29
providers will increasingly bear
the financial risk associated with complications
and unnecessary services.
Conclusion As other stakeholders consider
the value of the regional collaborative improve-
ment model, Blue Cross and Blue Shield of Mich-
igan/Blue Care Network and clinical leaders in
Michigan are already fully persuaded of the ben-
efits, and they continue to expand the scope of
these programs. New programs focused on total
joint replacement and interventions for atrial
fibrillation are being added in 2011. If early re-
sults from the Michigan initiative hold up, such
programs may represent a rare triple win: pro-
fessional satisfaction and preserved autonomy
for physicians; lower costs for payers; and better
outcomes for patients.
The authors gratefully acknowl edge
funding from Blue Cros s and Blue Shield
of Michigan/Blue Care Networ k and the
Blue Cross Blue Shie ld Foundation. This
work was also supporte d in part by the
Agency for Healthcare Research and
Quality.
NOTES
1 Brennan TA, Leape LL, Laird NM,
Hebert L, Localio AR, Lawthers AG,
et al. Incidence of adverse events and
negligence in hospitalized patients:
results of the Harvard Medical
Practice Study I. N Engl J Med.
1991;324(6):3706.
2 Ghaferi AA, Birkmeyer JD, Dimick
JB. Variation in hospital mortality
associated with inpatient surgery. N
Engl J Med. 2009;361(14):136875.
3 Leape LL, Brennan TA, Laird N,
Lawthers AG, Localio AR, Barnes BA,
et al. The nature of adverse events in
hospitalized patients: results of the
Harvard Medical Practice Study II.
N Engl J Med. 1991;324(6):37784.
4 OConnor GT, Plume SK, Olmstead
EM, Coffin LH, Morton JR, Maloney
CT, et al. A regional prospective
study of in-hospital mortality asso-
ciated with coronary artery bypass
grafting: the Northern New England
Cardiovascular Disease Study Group.
JAMA. 1991;266(6):8039.
5 Dimick JB, Weeks WB, Karia RJ, Das
S, Campbell DA Jr.Who pays for poor
surgical quality? Building a business
case for quality improvement. J Am
Coll Surg. 2006;202(6):9337.
6 Birkmeyer JD, Gust C, Baser O,
Dimick JB, Sutherland JM, Skinner
JS. Medicare payments for common
inpatient procedures: implications
for episode-based payment bun-
dling. Health Serv Res. 2010;
45(6 Pt 1):178395.
7 Hibbard JH, Stockard J, Tusler M.
Hospital performance reports: im-
pact on quality, market share, and
reputation. Health Aff (Millwood).
2005;24(4):115060.
8 Jha AK, Epstein AM. The predictive
accuracy of the New York State
coronary artery bypass surgery
If early results from
the Michigan initiative
hold up, such
programs may
represent a rare triple
win.
Quality Profiles
8 Health Affairs April 2011 30:4
report-card system. Health Aff
(Millwood). 2006;25(3):84455.
9 Romano PS, Zhou H. Do well-
publicized risk-adjusted outcomes
reports affect hospital volume? Med
Care. 2004;42(4):36777.
10 Scanlon DP, Lindrooth RC,
Christianson JB. Steering patients to
safer hospitals? The effect of a tiered
hospital network on hospital ad-
missions. Health Serv Res. 2008;
43(5 Pt 2):184968.
11 Rosenthal MB, Frank RG, Li Z,
Epstein AM. Early experience with
pay-for-performance: from concept
to practice. JAMA. 2005;294(14):
178893.
12 Rosenthal MB. Beyond pay for per-
formanceemerging models of pro-
vider-payment reform. N Engl J Med.
2008;18;359(12):1197200.
13 Bradley EH, Herrin J, Elbel B,
McNamara RL, Magid DJ,
Nallamothu BK, et al. Hospital
quality for acute myocardial infarc-
tion: correlation among process
measures and relationship with
short-term mortality. JAMA.
2006;296(1):728.
14 Campbell SM, Reeves D,
Kontopantelis E, Sibbald B, Roland
M. Effects of pay for performance on
the quality of primary care in
England. N Engl J Med. 2009;
361(4):36878.
15 Lindenauer PK, Remus D, Roman S,
Rothberg MB, Benjamin EM, Ma A,
et al. Public reporting and pay for
performance in hospital quality im-
provement. N Engl J Med. 2007;
356(5):48696.
16 Nicholas LH, Osborne NH,
Birkmeyer JD, Dimick JB. Hospital
process compliance and surgical
outcomes in Medicare beneficiaries.
Arch Surg. 2010;145(10):9991004.
17 Werner RM, Bradlow ET. Relation-
ship between Medicares Hospital
Compare performance measures and
mortality rates. JAMA. 2006;
296(22):2694702.
18 OConnor GT, Plume SK, Olmstead
EM, Morton JR, Maloney CT,
Nugent WC, et al. A regional inter-
vention to improve the hospital
mortality associated with coronary
artery bypass graft surgery. JAMA.
1996;275(11):8416.
19 Birkmeyer NJO, Birkmeyer JD.
Strategies for improving surgical
qualityshould payers reward
excellence or effort? N Engl J Med.
2006;354(8):86470.
20 Birkmeyer NJO, Share D, Campbell
DA, Prager RL, Moscucci M,
Birkmeyer JD. Partnering with
payers to improve surgical quality:
the Michigan plan. Surgery.
2005;138(5):81520.
21 Moscucci M, Rogers EK, Montoye C,
Smith DE, Share D, ODonnell M,
et al. Association of a continuous
quality improvement initiative with
practice and outcome variations of
contemporary percutaneous coro-
nary interventions. Circulation.
2006;113(6):81422.
22 Prager RL, Armenti FR, Bassett JS,
Bell GF, Drake D, Hanson EC, et al.
Cardiac surgeons and the quality
movement: the Michigan experi-
ence. Semin Thorac Cardiovasc Surg.
2009;21:207.
23 Birkmeyer NJO, Share D, Baser O,
Carlin AM, Finks JF, Pesta CM, et al.
Preoperative placement of inferior
vena cava filters and outcomes after
gastric bypass surgery. Ann Surg.
2010;252(2):3138.
24 Johnson SH, Theurer PF, Bell GF,
Maresca L, Leyden T, Prager RL,
et al. A statewide quality collabora-
tive for process improvement: in-
ternal mammary artery utilization.
Ann Thorac Surg. 2010;90(4):
115864.
25 Haynes AB, Weiser TG, Berry WR,
Lipsitz SR, Breizat AH, Dellinger EP,
et al. A surgical safety checklist to
reduce morbidity and mortality in a
global population. N Engl J Med.
2009;360(5):4919.
26 Pronovost P, Needham D,
Berenholtz S, Sinopoli D, Chu H,
Cosgrove S, et al. An intervention to
decrease catheter-related blood-
stream infections in the ICU. N Engl
J Med. 2006;355(26):272532.
27 Government Accountability Office.
Private health insurance: number
and market share of carriers in the
small group health insurance market
in 2004. Washington (DC): GAO;
2005. Report No. GAO-06-155R.
28 Fisher E, McClellan M, Bertko J,
Lieberman S, Lee J, Lewis J, et al.
Fostering accountable health care:
moving forward in Medicare. Health
Aff (Millwood). 2009;28(2):w2 19
31. DOI: 10.1377/hlthaff.28.2.w219.
29 Hackbarth G, Reischauer R, Mutti A.
Collective accountability for medical
caretoward bundled Medicare
payments. N Engl J Med. 2008;
359(1):35.
ABOUT THE AUTHORS: DAVID A. SHARE, DARRELL A. CAMPBELL,
NANCY BIRKMEYER, RICHARD L. PRAGER, HITINDER S. GURM,
MAURO MOSCUCCI, MARIANNE UDOW-PHILLIPS
&
JOHN D. BIRKMEYER
David A. Share is
the executive
medical director at
Blue Cross and Blue
Shield of Michigan.
1 In this issue of H ealth Affairs,
David Share, John Birkmeyer, and
their coauthor s make the cas e for
regional collaborations between
hospitalsanddoctorsasawayto
reduce health costs and improve
the quality of care. The authors
describe a Michigan-based project,
financed by Blue Cross and Blue
Shield of Michigan, that prevented
surgical complications in an
estimated 2 ,500 pa tients and saved
$20 million annually.
The researchers say that they
long ago realized that traditional
approaches to improving quality
and efficiencysuch as pay-for-
performance and outside reviews of
proposed treatmentsdidntwork
well, for various reasons. Efforts to
find another way brought them
together on this proj ect.
Share says that this papers
findings show that data-centered,
regional collaborations can
empower hospitals and doctors to
transform care in both community
and academic settings.
Share is the executive medical
director for health care qua lity at
Blue Cross Blue Shield of Michigan
and an adjunct clinical assistant
professor in the University of
Michigans Departments of Family
Medicine and Pe diatrics. He
April 2011 30:4 Health Affairs 9
received both his m edical degree
and his master of public health
degree from the University of
Michigan.
Darrell A. Campbell
is the chief of
clinical affairs in
the Department of
Surgery, University
of Michigan.
Darrell Campbell is the chief
medical officer and chief of clinical
affairs at the University of
Michigan Health System. He also is
the Henry King Ransom Professor
of Surgery at Michigan. Campbell
received his medical degree from
the George Washington University.
Nancy Birkmeyer is
an associate
professor at the
University of
Michigan.
Nancy Birkmeyer is an associate
professor in Michigans
Depa rtm ent of Su rger y and dire cto r
of the Michi gan Bar iatric Surgery
Collaborative. She received her
doctorate from DartmouthsCenter
for the Evaluative Clinical Sciences.
She and John Birkmeyer are
married.
Richard L. Prager is
a professor of
surgery at the
University of
Michigan.
Richard Prager is the director of
Michigans Cardiovascular Center
and leads the Michigan Society of
Thoracic and Cardiovascular
Surgeons Quality Collaborative
Initiative. He received his medical
degree from the State University of
New York, Downstate.
Hitinder S. Gurm is
an assistant
professor of
internal medicine at
the University of
Michigan.
Hitinder Gurm is an assistant
professor of internal medicine at
the University of Michigan Health
System and a graduate of Christian
Medical College in India.
Mauro Moscucci is
a professor of
medicine at the
University of
Miami.
Mauro Moscucc i is the chief of
the Cardiovascular Division and
vice chair of the Department of
Medicine at the Univ ersity of
Miamis Miller School of Medicine.
He received his medical degree
fromtheUniversityofRomeand
his master of business
administration degree from the
Ross School of Business, University
of Michigan.
Marianne Udow-
Phillips is the
director of the
Center for
Healthcare
Research and
Transformation.
Marianne Udow-Phillips is the
director of the C enter for
Healthcare Research and
Transfo rmation, a no nprofi t
partnership of the University of
Michigan Health System and Blue
Cross and Blue Shield of Michigan.
She holds a mastersdegreein
health services administration from
the University of Mic higan Sc hool
of Public Health.
John D. Birkmeyer
is a professor of
surgery at the
University of
Michigan.
John Birkmeyer is the Geo rge D.
Zuidema Professor of Surgery and
Chair of Health Services Research
at the University of Michigan. He
also serves as the director of the
universitysCenterforHealthcare
Outcomes and Pol icy. He rece ived
his medical degree from Harvard
Medical Scho ol.
Quality Profiles
10 Health Affairs April 2011 30:4
Queries
1. In About the Authors and on page 1 authors bios, please compare and ensure that
all details in both locations are accurate. All details do not need to be in both places,
but please be sure that there are no inconsistencies.
April 2011 30:4 Health Affairs 11
    • "Novel performance scores have been developed to forecast outcomes in cardiac surgery [62, 63], and these concepts are worthy of further consideration. To evaluate methods that optimize outcomes and to define best practices, cardiac surgery collaboratives and multicenter quality improvement programs have reported, on average, a 20% to 24% reduction in mortality rates, with one institution demonstrating a 40% reduction in risk-adjusted mortality, decreased morbidity, and increased success with early extubation and glycemic control [64][65][66]. TeamSTEPPS (Team Strategies and Tools to Enhance Performance and Patient Safety) and other training programs that incorporate human factors principles have demonstrated that surgical teams that train together develop effective leadership and communication skills, and use of briefings and debriefings can produce better outcomes [67, 68]. " Failure to rescue " patients from complications is a core quality measure endorsed by the National Quality Forum in 2012 [69]. "
    Article · Sep 2016
    • "This study analyzed data from the Michigan Bariatric Surgery Collaborative (MBSC), a statewide payor-funded consortium of hospitals and surgeons providing bariatric surgical care [12,13]. The MBSC consists of all Michigan hospitals that perform a minimum of 25 bariatric surgery cases per year (38 hospitals and 68 surgeons in 2014). "
    [Show abstract] [Hide abstract] ABSTRACT: Background: Postoperative gastroesophageal reflux is one of the most important long-term complications of sleeve gastrectomy, the most common bariatric procedure. Objective: To assess variation in hospital performance with laparoscopic sleeve gastrectomy using rates of acid-reducing medication use at postoperative 1 year. Setting: Clinical registry of bariatric surgical patients at academic and community hospitals in Michigan. Methods: We studied 2923 patients who underwent laparoscopic sleeve gastrectomy across 39 hospitals in the Michigan Bariatric Surgery Collaborative, 2007 to 2014. We compared risk- and reliability-adjusted rates of new-onset reflux-defined by new use of acid-reducing medication-across hospitals and in relation to surgical quality indicators (hospital procedure volume and 30-day complications). Results: Overall, 20% of patients were newly taking acid-reducing medication at postoperative 1 year. Hospital rates of new medication use varied 3-fold, ranging from 10% (95% CI 7-15%) to 31% (95% CI 23-40%) of patients. Of the 2 hospitals with significantly lower rates of new medication use, 1 was high volume and 1 was medium volume. The 1 hospital with significantly higher rates was medium volume. Rates of acid-reducing medication use did not correlate with hospital volume or perioperative complications. Conclusion: Across Michigan hospitals, rates of new acid-reducing medication use at 1 year after laparoscopic sleeve gastrectomy varied widely and did not correlate with traditional quality indicators. Future research could explore differences in surgical technique to better understand the factors underlying variation in long-term outcomes after sleeve gastrectomy.
    Full-text · Article · Dec 2015
    • "It should be noted that participating in these initiatives can also often involve a cost to the institution, most often including an annual fee on the order of $5,000 to $10,000, and salary associated with data entry personnel. The present study and data from the adult collaboratives suggest that these costs may be offset if participation leads to successful reduction in complications or LOS in even a handful of patients each year [16, 17]. "
    [Show abstract] [Hide abstract] ABSTRACT: Background While there is an increasing emphasis on both optimizing quality of care and reducing health care costs, there are limited data regarding how to best achieve these goals for common and resource-intense conditions such as congenital heart disease. We evaluated excess costs associated with complications and prolonged length of stay (LOS) after congenital heart surgery in a large multicenter cohort. Methods Clinical data from The Society of Thoracic Surgeons Database were linked to estimated costs from the Pediatric Health Information Systems Database (2006 to 2010). Excess cost per case associated with complications and prolonged LOS was modeled for 9 operations of varying complexity adjusting for patient baseline characteristics. Results Of 12,718 included operations (27 centers), average excess cost per case in those with any complication (versus none) was $56,584 (+$132,483 for major complications). The 5 highest cost complications were tracheostomy, mechanical circulatory support, respiratory complications, renal failure, and unplanned reoperation or reintervention (ranging from $57,137 to $179,350). Patients with an additional day of LOS above the median had an average excess cost per case of $19,273 (+$40,688 for LOS 4 to 7 days above median). Potential cost savings in the study cohort achievable through reducing major complications (by 10%) and LOS (by 1 to 3 days) were greatest for the Norwood operation ($7,944,128 and $3,929,351, respectively) and several other commonly performed operations of more moderate complexity. Conclusions Complications and prolonged LOS after congenital heart surgery are associated with significant costs. Initiatives able to achieve even modest reductions in these morbidities may lead to both improved outcomes and cost savings across both moderate and high complexity operations.
    Full-text · Article · Nov 2014
Show more