Linking Quality of AMI Care and Outcomes

1
Linking Quality of AMI Care and Outcomes

• April 30, 2003
• 2-3 pm ET

2
AHA/ACC MI Guideline RecommendationsProcess
Quality Indicators
Acute Care
Discharge Care

• ECG within 10
• Aspirin lt 24
• Beta-Blocker lt 24
• Any Reperfusion (STEMI)
• If lytic, door-drug lt30
• If PCI, door-balloon lt90
• Ace-Inhibitor lt24hr (STEMI)
• IV Heparin use (NSTEMI)
• GP IIb-IIIa Inhibitor (NSTEMI)

• D/C Aspirin
• D/C Beta-Blocker
• D/C ACE-Inhibitor (low EF, DM)
• D/C Statin ( lipids)
• Smoking Cessation(smokers)
• Cardiac Rehab

JACC 200036970-1062 JACC 199934890-911
3
The Cycle of Clinical Therapeutics
Clinical Trials
Discovery
Guidelines
Patient Outcomes
Performance Indicators
???
Caregiver Performance
Califf, R et al JACC 2002 in press
4
Variation in AMI Care Across 1,247 US Hospitals
and its Association with Hospital Mortality

• Eric D Peterson, MD, MPH Lori S Parsons, BS
  Charles V Pollack, Jr., MD, MA L Kristin Newby,
  MD, MA Katherine A Littrell, PhD, RN, for the
  National Registry of Myocardial Infarction (NRMI)
  4 Investigators

Drs Peterson and Pollack are on the NRMI advisory
board and are on the speakers bureau for
Millennium Pharmaceuticals Dr Littrell is an
employee of Genentech, Inc, sponsor of NRMI
Duke Clinical Research Institute, Duke
University, Durham NC Pennsylvania Hospital,
Philadelphia PA, Genentech, Inc., San Francisco,
CA
Peterson Circulation 2002106(19)II-722
5
Questions for Hospital Quality Assessment and
Quality Improvement

• How does care for MI patients vary across US
  hospitals?
• What is the gap between leading and lagging
  centers?
• To what degree are specific performance
  indicators correlated to one another?
• Does one need only measure a few metrics to
  assess hospitals quality of care? Or, do ratings
  shift depending on the metric?
• Are process performance indicators associated
  with patient outcomes?
• Do centers with greater adherence to the AHA/ACC
  MI Guidelines have lower mortality rates than
  those not?

6
Study Design

• Database National Registry of Myocardial
  Infarction (NRMI) IV June 2000 thru June 2002
• Patients All confirmed MI pts (troponin or
  CK/MB)
• Exclusions
• Transfers out (incomplete mortality)
• Those with malignancy
• Those who died lt 24 hrs of admission
• Total Hospitals N 1,247
• Total Patients N 257,482

7
Methods Process Measures of Quality

• Individual measures AHA/ACC guideline-based
  acute and discharge care processes (n15)
• Denominator Specific to measure (e.g., ST vs
  NSTEMI)
• Excluded those with treatment contraindications
• Composite Quality
• Each patients MI care assessed for adherence to
  AMI guidelines for up to 15 care process
• Calculated hospitals correct care out of
  total care opportunities

8
Methods Statistics

• Hierarchical models to account for pt clustering
  within hospitals
• Models include Clinical risk factors (fixed
  effect)
• Model C-index 0.76
• Hospital-level components (random effects)
• Region, bed size, academic, type (cath/PCI/CABG)
• Process performance measures
• Individual or composite

9
Results Baseline Characteristics
10
Results Hospital Characteristics
11
Distribution of Hospital Process Indicators

• Mean of MIs per hospital per year
• 105 (70 NSTE MI)
• Mean eligible for ASA lt24 hrs measure
• 70
• Mean eligible for acute reperfusion measure
• 18
• Mean Composite Quality opportunities
• 805

12
Results Overall MI Care Correct Care
13
Results Leading and Lagging Hospital Quartiles
Acute Care (1)
14
Results Leading and Lagging Hospital Quartiles
Acute Care (2)
15
Results Gap between Leading and Lagging
Hospitals Quartiles Discharge Care
LVEF lt 40 Known hyperlipidemia
16
Correlation between Hospital Performance Measures
17
Relationship between Overall Composite Quality
and In-Hospital Mortality
18
NRMI Quality Conclusions

• Large and persistent gap between AHA/ACC MI
  guidelines and current community MI care
• Wide variability between leading and lagging
  centers
• Hospitals who performed well on one individual
  performance indicator may not do well on another
• Need to look across care spectrum and composite
  metrics most stable
• Hospitals adherence with AHA/ACC guidelines
  strongly correlated with patient outcomes
• MI mortality rates 40 lower at leading adherent
  hospitals relative to those lagging

19
CRUSADE A National Quality Improvement
Initiative
Can Rapid Risk Stratification of Unstable Angina
Patients Suppress ADverse Outcomes with Early
Implementation of the ACC/AHA Guidelines
20
CRUSADE Site Distribution
Active sites 420 Total ACS Patients 49,860
21
Gap between Leading and Lagging Hospital
Quartiles Acute Care
22
Gap between Leading and Lagging Hospital
Quartiles Discharge Care

LVEF lt 40 Known hyperlipidemia
23
Performance Matters!Relationship between Process
and Outcome
5.9
5.0
4.6
3.6
Peterson ED 2002 AHA
24
CRUSADE Practical Steps to Improve the Use of
Evidence-Based Therapies

• Identify local physician champions
• Secure institution wide commitment to improve
  care
• Promote collaboration b/t ED, primary care
  cardiology
• Develop educational and QI tools to promote
  standard use of ACC/AHA guidelines
• Provide site-specific reports with national
  benchmarks
• identify areas for QI
• Initiate rapid QI cycles and track improvement
  overtime

25
CRUSADE Trends in Acute Therapy
26
CRUSADE Trends in Discharge Therapy
27
CRUSADE Trends in Discharge
Recommendations