OSHPD Public Reporting Program

1
OSHPD Public Reporting Program
Zhongmin Li, Ph.D. UCD K-30 Methods in Clinical
Research July 26, 2006
2
Overview

• What is OSHPD?
• What is CCORP?
• How the CABG report cards were produced?

3
What is the Office of Statewide Health Planning
and Development (OSHPD)?

• Facilities Development
• Seismic safety
• Code compliance
• Cal-Mortgage
• Facility financing
• Workforce and Community Development
• Health professions training
• Nursing initiative

• Healthcare data
• PDD, utilization, financial
• Inpatient, ED and ASC
• Healthcare Quality and Analysis Division
• Hospital outcome reports (HOC)
• Data dissemination - HIRC

4
Healthcare Outcomes Center (HOC)

• 2 Programs within HOC
• Clinical Data (Heart bypass surgery ICU
  outcomes)
• Patient Discharge Data (heart attack, pneumonia)
• 12 professional/administrative staff with 3
  Senior Ph.D.-level research scientists
• Contracts with University of California
  researchers

5
OSHPDs Outcome Programs
Studies Using Clinical Data
Studies Using Patient Discharge Data

• Coronary Artery Bypass Graft (CABG) Surgery (2006
  hospitals and surgeons)
• Intensive Care Outcomes (Final study and
  recc.-2005)

• Acute Myocardial Infarction (2002)
• Community-Acquired Pneumonia (2004)
• Maternal Outcomes (2006)
• Hip Fracture surgery (2006)

Gold Reports that have been or will shortly be
released
6
OSHPD Reports and Report Cards
7
What is the California CABG Outcomes Reporting
Program (CCORP)?

• Started as voluntary program in 1995 by OSHPD
  PBGH
• Mandatory data reporting for all Coronary Artery
  Bypass Graft (CABG) surgery
• Risk-adjusted operative mortality reported
  publicly reported for
• Hospitals (annually starting 2003 data)
• Surgeons (bi-annually starting 2003-04 data)
• A Clinical Advisory Panel of heart surgeons,
  consumers, cardiologists, and university
  researchers

8
What is CCORP (Contd)?

• Clinical risk-adjustment to account for
  differences in patient severity of illness
  (case-mix)
• Yearly patient medical records audit to ensure
  data integrity
• First data year Jan-Dec 2003, received from 121
  hospitals
• Approximately 300 California surgeons
• UCDMC has been participating since 1997

9
Rationale for CCORP

• Senate Bill 680 (Figueroa, Statutes of 2001)
• CABG is one of the most expensive hospital
  surgeries and ranks among top 10 procedures for
  volume and mortality rate
• Limitations of the voluntary OSHPD/PBGH program
• Comparative quality information informs
• Consumer choice
• Healthcare purchasing
• Quality improvement by physicians and hospitals
• Salience of surgeon-level outcomes

10
CCORP Clinical Advisory Panel (CAP) Members

• Robert Brook, M.D. (Chair) RAND and UCLA
• Andrew Bindman, M.D. UCSF and S.F. General
• Ralph Brindis, M.D.Kaiser Northern California
• Cheryl Damberg, Ph.D. Pacific Business Group on
  Health
• Timothy Denton, M.D. Cedars Sinai

• Coyness Ennix, M.D.Alta Bates Medical Center
• Keith Flachsbart, M.D. Kaiser San Francisco
• Fred Grover, M.D.University of Colorado, STS
• James MacMillan, M.D. Valley Heart Surgeons,
  Modesto

11
Unadjusted Isolated CABG Inpatient Mortality
Rates for CA and Other States (1996-2004)
NOTE MA 30-day mortality
12
Hospital Distribution of Unadjusted In-Hospital
Mortality Rate Avg. 2.4 (2003)
13
Hospital Distribution of Risk Adjusted Operative
Mortality Avg. 2.9 (2003)
14
CCORP 2003 Hospital Performance Ratings for
Operative Mortality
2003 CCORP public report was released in March
2006 http//www.oshpd.state.ca.us
15
How CABG Report Cards Were Produced?

• Data Collection
• Source
• Data elements (59 not exact same as STS)
• Hospital reporting tools
• Acceptance criteria
• Deadline and Extension
• OSHPD staff support
• Data abstraction manual
• Training of coders

16
How CABG Public Report Was Produced (Contd)?

• Data Cleaning and Validation
• Data quality report (DQR)
• Data Linkage/discrepancy report (DDR)
• On-site data audit
• Audit strategy
• Sample selection
• As good as you can get?

17
Risk Adjustment

• Central to the report cards
• Why Isolated CABGs only
• Inpatient vs. operative mortality
• Risk model
• Development
• Validation
• Testing (c-statistics calibration)
• CAP approval
• Identification of quality outliers
• 95CI
• Exact Poisson Probability

18
Major Risk Factors

• Demographic (Age/Gender/Race/BMI)
• Operative status
• Pre-operative comobidities (Creatinine/Dialysis/Di
  abetes/CVD/PVD/CLD/Hypertension/Hepatic failure,
  etc.)
• Cardiac (Arrhythmia type/MI timing/CHF/Shock)
• Previous intervention (Prior CABG/PCI)
• Hemodynamic status (EF/Left main/ of diseased
  vessels/Mitral insufficiency)

19
Risk Adjustment Model

• Expected rate Intercept XB, where
• B-Coefficients estimated on CCORP data using
  logistic regression (PROC LOGISTIC in SAS)
• X includes age, gender and each of risk factors
• Risk-adjusted rate
• Observed rate
• (Population rate/Expected rate)
• Alternative formula
• Risk-adjusted rate
• Population rate
• (Observed rate/Expected rate)

20
Risk Adjustment Example

• Risk adjustment adjusts the observed rate of
  Hospital A to account for differences between the
  case-mix of Hospital A and the reference
  population
• The larger the difference between the rate was
  expected for Hospital A and the population rate,
  the larger the adjustment
• If a hospital has the same case-mix as the
  reference population, then no adjustment is made

21
Risk Adjustment Example

• Population rate gt Hospital A Expected rate
• Hospital A has a less severe case-mix than the
  population
• adjustment factor is gt1
• Risk Adjusted rate is higher then observed rate
• If hospital A had the same case-mix as the
  population, we would expect their observed rate
  to be higher

22
Risk Adjustment Example

• Population rate lt Hospital A Expected rate
• Hospital A has a more severe case-mix than the
  population
• adjustment factor is lt1
• Risk Adjusted rate is lower then observed rate
• If hospital A had the same case-mix as the
  population, we would expect their observed rate
  to be lower

23
Report Review and Appeal Process

• Prior to public release
• Hospital review (60-day)
• Surgeon review (30-day)
• Surgeon can appeal for
• Flawed data
• Flawed risk adjustment
• 1st surgeon level report Winter 2006

24
Whats Next for CCORP? NQF National Voluntary
Cardiac Surgical Measures

• Gold currently collected by CCORP or will be
  collected in 2006
• 1. Participation in a systematic database for
  cardiac surgery
• 2. Surgical volume for isolated CABG surgery,
  valve surgery, and CABG valve surgery
• 3. Timing of antibiotic administration for
  cardiac surgery patients
• 4. Selection of antibiotic administration for
  cardiac surgery patients
• 5. Pre-operative beta blockade
• 6. Use of internal mammary artery
• 7. Duration of prophylaxis for cardiac surgery
  patients
• 8. Prolonged intubation
• 9. Deep sternal wound infection rate
• 10. Stroke/cerebrovascular accident
• 11. Post-operative renal insufficiency

25
NQF National Voluntary Cardiac Surgical Measures
(Contd)

• 12. Surgical re-exploration
• 13. Anti-platelet medications at discharge
• 14. Beta blockade at discharge
• 15. Anti-lipid treatment at discharge
• 16. Risk-adjusted inpatient operative mortality
  for CABG
• 17. Risk-adjusted operative mortality for CABG
• 18. Risk-adjusted operative mortality for AVR
• 19. Risk-adjusted operative mortality for MVR
• 20. Risk-adjusted operative mortality for
  MVRCABG
• 21. Risk-adjusted operative mortality for
  AVRCABG

26
IMA Usage Rates as a Reportable Process Measure
of Quality

• Internal Mammary Artery Graft
• Standard conduit for CABGs
• Supported by extensive clinical trials and
  research
• Priority research area of performance - NQF
• Process of Care element - STS
• Reported by Leap Frog Group
• IMA data and clinical info. available - CCORP

27
Hospital Variation in Left Internal Mammary
Artery (LIMA) Use CCORP 2003

• Number of Hospitals ()
• 1 (1)
• 11 (9)
• 17 (14)
• 48 (40)
• 44 (36)
• 121 total hospitals
• UCDMC - 92.5

• Percent LIMA Use
• lt60 -
• 60-74 -
• 75-84 -
• 85-94 -
• ?95 -
• Avg. LIMA Use 89.1 -
• Rates are after exclusions for certain patient
  subgroups

28
CPB Used vs. OPCABG

• Operative Mortality
• Off Pump CABGs
• 9,025 (22.3) 2.17
• On Pump CABGs
• 31,380 (77.7) 3.36
• UCDMC - lt 2

29
How Are Hospital CABG Report Results Used?

• Hospital tiered networks Blue Cross Cardiac
  Centers of Expertise and Blue Shield Cardiac
  Quality Initiative
• Private Healthcare data providers such as Subimo
• Purchaser coalitions such as Leapfrog, PBGH
  (Healthscope consumer website)
• Public Access through OSHPD website
• http//www.oshpd.state.ca.us/HQAD/Outcomes/index.h
  tm
• CHART
• Hospital Promotion

30
Questions

• OSHPD web site (www.oshpd.ca.gov)
• Contact HIRC at [phone] to obtain copy of
  the report