1. The Vascular Quality Initiative
Improving Quality
Through Regional Collaboration
Jack L. Cronenwett, M.D.
Dartmouth-Hitchcock Medical Center
2. Vascular Study Group of New England
Initial Planning Meetings in 2001
Agreed on mission and method:
To improve the care of patients with
vascular disease by collecting and
exchanging information.
Focus = quality improvement
3. Unique Aspects of VSGNE Database
One year follow-up for key outcomes
• Completed in surgeon’s office
Prospective, consecutive cases
• Audited against claims data
Surgeon and center level reports
• Benchmark comparison with others
Detailed clinical data
• Pre-, intra-, and post-op variables
Academic and community hospitals
• Real world practice
4. Focus on Quality Improvement
Track key procedures
• CEA, CAS, EVAR, TEVAR, open AAA, lower
extremity bypass/interventions, access
• Key procedures provide overall insight
• Consecutive procedures must be entered
Semi-annual Meetings
• Critical to success, durability of group
• Stimulate cooperative quality projects
• Overcome insular nature of practice with
granular conversations about quality
5. Recent Meeting Agenda
• AV access and TEVAR working groups report
• CLI treatment preference survey results
• Panel: Lower extremity bypass: Techniques that work
• VSG CRI cardiac risk online prediction tool
• Predicting respiratory failure after elective OAAA repair
• Carotid patch and re-stenosis update
• Intensive glucose management in LEB patients
• Outcomes of LEB after previous interventional treatment
• MI rates in diabetics after LEB
• Clinical improvement vs. graft patency in LEB
• Impact of increased beta blocker usage
• Statin use working group report
• New QI projects and clinical uses for registry
• Variation in complication rates by center and procedure
6. Dartmouth-Hitchcock
Medical Center
Fletcher Allen
Health Care
Eastern Maine Medical Center
Maine Medical Center
Catholic Medical Center
Concord Hospital
Lakes Region
Hospital
Cottage
Hospital
Central Maine Medical Center
VSGNE 2003
9 Participating Hospitals
7. Dartmouth-Hitchcock
Medical Center
Fletcher Allen
Health Care
Eastern Maine Medical Center
Maine Medical Center
Concord Hospital
Lakes Region
Hospital
Cottage
Hospital
Central Maine Medical Center
Mercy Hospital
U. Mass. Medical Center
Elliot Hospital
Tufts Medical Center
Boston Medical Center
St. Francis Hospital
Massachusetts General Hospital
MaineGeneral Medical Center
Caritas St. Anne’s Hospital
Yale-New Haven Hospital
Baystate Medical Center
VSGNE 2013
30 Participating Hospitals
Berkshire Medical Center
15 Community - 15 Academic
Hartford Hospital
St. Luke’s Hospital
Charlton Memorial Hospital
Beth Israel Deaconess Medical Center
Hospital of St. Raphael
Cardiothoracic Surgical Associates
Brigham & Women’s Hospital
Danbury Hospital
St. Elizabeth’s
Hospital
Center
Miriam Hospital
Rhode Island Hospital
“Real World Practice”
9. Regional Quality Improvement
Use registry as a tool for regional QI
• Different from most national registries
Analyze natural variation in process
and outcomes across centers
Develop QI projects in areas where
substantial variation exists
Provide benchmark comparisons that
stimulate everyone to improve
10. Regional Quality Improvement
Can benchmarking change practice?
Can we improve outcomes?
Can we improve patient selection?
Can we generate new knowledge?
11. Changing Practice:
Medical Risk Reduction
Statin treatment pre-operatively
• Discussed evidence for benefit at
semi-annual meetings
• Selected pre-op statin use as a quality
measure
• Reported benchmarked results to centers
and surgeons
15. Changing Practice to Change Outcome
Patching conventional CEA
• Level I evidence shows reduced stroke
risk and less re-stenosis
• Discussed evidence for benefit at
semi-annual meeting
• Selected as a quality measure
• Reported benchmarked results to
centers and surgeons
16. 3,427 CEAs in 3,304 Patients
>80% Stenosis at One Year
4.2
1.4
0
1
2
3
4
5
No Patch Angioplasty Patch Angioplasty
OneYearStenosisRate(%)
Patch:
3-Fold
Reduction
p=0.001
%
%
Multivariate
Predictor of
80-100%
Stenosis
-Goodney et al, SAVS 2010
17. 13%
12%
10%
6%
5%
4%
0%
4%
8%
12%
2003 2004 2005 2006 2007 2008
Year
RestenosisRate
Conventional
CEA without
Patch
Percentage of Patients Not
Patched Decreased over Time
p<0.003
-Goodney et al, SAVS 2010
18. 3%
2%
1%
2%
1%
0%
13%
12%
10%
6%
5%
4%
0%
4%
8%
12%
2003 2004 2005 2006 2007 2008
Year
RestenosisRate
80-99%
Stenosis
p<0.001
One Year Stenosis Rate Also
Decreased over Time
Conventional
CEA without
Patch
p<0.003
Process Improvement Outcome Improvement
-Goodney et al, SAVS 2010
20. Number of
RCRI
Risk
Factors
RCRI
Predicted
Risk (%)
VSGNE
Actual
Event
Rate (%)
0 0.4 2.6
1 0.9 6.7
2 6.6 11.6
≥ 3 11.0 18.4
Predicting Cardiac Complications
MI,CHF, arrythmia
Revised Cardiac Risk
Index – 6 factors:
• CAD, CHF, IDDM, CVA, creat
> 2, high risk surgery
• Only 20% of operations in
derivation set were vascular
Underestimates risk in
vascular surgery
patients in VSGNE
-Bertges et al, J Vasc Surg, 2010
21. Risk of Adverse Cardiac Outcome, by
VSG-CRI Score
2.6
3.5
6.0 6.6
8.9
14.3
0
4
8
12
16
0-3 4 5 6 7 8 or
More
VSG-CRI Score
RiskofAdverseCardiac
Outcome(%)
Step 1:
Calculate VSG-RCI Score
Step 2:
Use VSG-CRI Score To Predict Risk
of Adverse Cardiac Outcome
Example patient: 80 yr-old smoker with history of CAD.
VSG-CRI score = 4 + 1 + 2 = 7
Vascular Study Group Cardiac Risk Index (VSG-CRI)
VSG-CRI Risk Factors # Points
Age ≥ 80 4
Age 70-79 3
Age 60-69 2
CAD 2
CHF 2
COPD 2
Creatinine > 1.8 2
Smoking 1
Insulin Dependant Diabetes 1
Chronic β-Blockade 1
History of CABG or PCI -1
(Based on 10,000 Patients)
www.VSGNE.org
22.
23. New Knowledge Practice Change
Does protamine reduce re-operation
for bleeding after CEA?
Re-operation for bleeding: 1.2%
N= 4587 CEAs in VSGNE
-Stone et al, J Vasc Surg, 2010
24. VSGNE Surgeon Practice
4587 Total CEAs
2087 (46%)
Protamine
2500 (54%)
No Protamine
-Stone et al, J Vasc Surg, 2010
27. Value of a Detailed Clinical Registry
Only a very large registry with detailed
clinical information could answer a
question where the event rate is low.
• Clinical trial too expensive
• Administrative databases lack clinical detail
Will this information change protamine use
in our region?
Will this change reduce bleeding
complications?
28. VSGNE Protamine Use during CEA
Protamine use
increased from 46%
before 2009 to 61%
after 2009 (P<.001).
29. Protamine Use and Bleeding
46%
61%
0%
10%
20%
30%
40%
50%
60%
70%
Protamine Use Before 2009
After 2009
P<.001 1.2%
0.6%
0.0%
0.2%
0.4%
0.6%
0.8%
1.0%
1.2%
1.4%
Re-operation for
Bleeding
P=.003
30. VSGNE- Lessons Learned
Many registries have failed
• Failure to add value to clinician, hospital
Comparison reports important
• Stimulates improvement
• Valuable to hospitals and clinicians
Regional group meetings promote QI
• Maintain momentum, enthusiasm
• Develop group trust, cooperative projects
Research is an important derivative
• Stimulates academic center participation
32. Cooperative Regional Groups
Vision:
• Regional groups for local control, data
ownership and responsibility
• Central data collection, analysis
Grp
A
Grp
E
Grp
B
Grp
C
Grp
D
Mechanism for Data Sharing
Among Regional Groups
33. Launched by SVS in 2011
• Mission: To improve the quality, safety,
effectiveness and cost of vascular health care by
collecting and exchanging information.
• Procedures:
Carotid endarterectomy and stenting; open and
endovascular AAA repair; lower extremity bypass and
interventional Rx; thoracic aorta stent grafts; dialysis
access; amputation; vena cava filters; varicose veins
34. • Patient Safety Organization
– Central data collection, analysis
– Protects data from discovery or disclosure
– Prevents identification of patient, hospital, physician
• Regional Quality Improvement Groups
– Analogous to VSGNE
• Web-based data collection - reporting system
– Provided by M2S, Inc.
3 Components
38. Risk-Adjusted Benchmark Comparison
All Centers with at least 40 Procedures in SVS VQI
0%
1%
2%
3%
4%
5%
6%
7%
8%
* *
Death or Stroke Rate after Elective Primary CEA
Observed and Expected by Centers
14,182 patient procedures, 2003 to September 2012
(Excludes: previous ipsilateral CEA, concomitant CABG)
Observed Expected
Overall stroke or death rate:
VQI = 1.1%
AUC = 0.65
Centers
adjusted for: Hypertension, CABG/PTCA, ASA/Plavix,
degree of contralateral stenosis and ipsilateral ICA stenosis,
ipsilateral cortical symptoms
Significantly lower or higher than
expected:
* p<0.05
41. 0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
A* B C* D E* F G H I J K L M N O P Q R S* T* U* V* W* X* Y* Z*
Percentage of Patients with Length of Stay >1 Day after Elective
Carotid Endarterectomy : Observed and Expected by Center
8,112 CEAs, 2003 thru 2011 (Excludes inhospital deaths)
observed expected
Overall LOS > 1 day= 23%
AUC = 0.611
Medical Center
adjusted for: age, gender, prior stroke, nursing home,
prior vascular procedure, diabetes, creatinine, CHF, CAD
In 11 centers the observed % of pateints with
LOS > 1 day was significantly lower or higher
than expected (*)
42. • Achieved PSO accreditation by AHRQ
• Established PSO structure, Governing Council
• Enabled international participation
Results to Date (since February, 2011)
43. Organization
Governing Council
4 SVS Representatives
2 AVF Representatives
15 Regional Group Representatives
Arterial Research Advisory
Committee
2 SVS Representatives
10 Regional Group Representatives
Arterial Quality Committee
4 SVS Representatives
15 Regional Group Representatives
Venous Quality Committee
3 AVF + 2 SVS Representatives
15 Regional Group Representatives
Venous Research Advisory
Committee
3 AVF + 2 SVS Representatives
10 Regional Group Representatives
44. 228 Centers, 45 States + Ontario
as of 5/1/2013
0
15
30
45
60
75
90
105
120
135
150
165
180
195
210
225
Growth of ParticipatingCenters
45. Organized Regional Groups:
– New England
– Carolinas
– Florida-Georgia
– Southern California
– South
– Virginias
– New York City
– Rocky Mountains
– Illinois
– Wisconsin
– Mid-Atlantic
– Upstate New York
– Chesapeake
– Indiana
– Great Lakes
Organizing Regional Groups:
– Northern California
– Michigan
– Missouri
– Tennessee/Mississippi
– Minnesota
15 Regional Quality Groups
47. Endorsing Societies:
American Venous Forum
Canadian Society for Vascular Surgery
Eastern Vascular Society
Florida Vascular Society
Michigan Vascular Society
Midwestern Vascular Surgical Society
New England Society for Vascular Surgery
New York Society for Vascular Surgery
Peripheral Vascular Surgery Society
Rocky Mountain Vascular Society
Society for Clinical Vascular Surgery
Society of Interventional Radiology
Southern Association for Vascular Surgery
Southern California Vascular Surgical Society
Western Vascular Society
48. Very Different than ACS-NSQIP
SVS VQI ACS-NSQIP
PSO Protected Data Collection Yes No
Regional Quality Groups Yes No
Physician Benchmark Reports Yes No
Number of Procedure Modules 11 6
Selection of Cases All Sampling
Variables Recorded
Detailed, Vascular
Focused
General, Apply to
Multi-specialties
Follow-up One year 30 days
Annual Cost $10,000 $35,000
49. • Significant variation
found across VQI
participating centers
and regions
• Risk factors associated
with SSI:
– Skin prep not
chlorhexidine
– Operation > than
220 minutes
– Transfusion > 3 units
PRBC
Surgical Site Infection after Infrainguinal Bypass
0%
1%
2%
3%
4%
5%
6%
7%
8%
9%
10%
** ** **
In-Hospital Surgical Site Infection Rate after Infra-Inguinal Bypass
Procedure
Observed and Expected by Regions
9,035 patient procedures, 2003 to October 2012
Observed Expected
Overall Rate
Wound Infection
VQI = 4.5%
VQI Regional Quality Groups
adjusted for: skin preparation, ankle/brachial
systolic pressure index, transfusion, length of
procedure
Significantly lower or
higher than expected:
* p<0.05
**p<0.01
Note: This patient safety work product generated within the SVS PSO, LLC, is considered privileged and confidential.
51. • Province would function as “Regional Quality Group”
– Anonymous comparison province, country, all U.S. sites
– Control of data use for research and QI projects
• De-identified patient data transmitted to SVS PSO via M2S site
– Identification key kept by each site and province
– Data copied to province for analysis
• Data audits for consecutive cases by province
• Representation on SVS PSO Governing Council, Committees
• Access to all VQI data for research and quality improvement
Quebec Province Participation
52. • National registries offer power of large database
for research, risk-adjusting, benchmarking
–Don’t operationalize quality improvement
• Regional quality groups create local ownership,
responsibility, vehicle for QI projects
–Potential for international collaboration
• Opportunity to combine both factors in VQI
• www.svsvqi.org
Conclusions
Editor's Notes
9 of the VSGNNE risk factors were converted to a weighted point score to create a practical formula called the Vascular Study Group-Cardiac Risk Index or VSG-CRICardiac stress testing was removed from the index to provide a purely clinical risk prediction formula applicable to all patients in the preoperative settingThe Vascular Study Group-Cardiac Risk Index translates to increasing levels of risk for adverse cardiac outcomes ranging from 2.6% for lowest risk score of 0-3, 6.0-6.6% for intermediate risk score of 5-6 and 8.9-14.3% for highest risk score of 7-8For example an 80 yr old smoker with a history of CAD and prior CABG would have a VSG-CRI score of 6 with predicted risk of cardiac events of nearly 7%
A multivariable logistic model was developed which found 4 variables that were independently predictive of SSI:Patients with Ankle-Brachial Index (ABI) less than 0.35 had a higher risk of SSISkin preparation using Chlorhexidine was associated with a lower risk of SSI, especially in patients with tissue loss.Patients receiving Transfusion of 3 or more units during their admission had a higher risk of SSISurgery time of more than 220 minutes was associated with a higher risk of SSICenters received Center Opportunity Profile for Improvement (COPI) report
