Surveillance of Heart Diseases and Stroke Using Centers for Medicare and Medicaid CMS Data

1
Surveillance of Heart Diseases and Stroke Using
Centers for Medicare and Medicaid (CMS) Data
Judith H. Lichtman, PhD MPH Associate Professor
2

• Need for strong population-based data to achieve
  surveillance goals.

Circulation, 2007115127-155
3
Current Gaps

• No comprehensive national surveillance system
• Track patterns of cardiovascular disease, care,
  and outcomes over time
• Cohort Studies
• Provide valuable information, but have
  limitations
• CVD and stroke have a large impact in the elderly
  population

4
Mortality by Age, 2004
Created from CDC Mortality Data 2004.
5
(No Transcript)
6
Previous Strategies Using CMS Data

• Cross-sectional design
• Subset of national cohort
• Focus on short-term outcomes
• Limited patient-level information
• No individual follow-up over time

7
Rates and Trends of Vascular Disease

• Prevalence
• Mortality
• Morbidity
• Incidence

Age

• Vascular Diseases
• MI
• Stroke
• Heart Failure

Sex
Race
Region
8
Examples provided

• Change in AMI mortality from 1995-2006
• Trends in 30-Day mortality following AMI by age,
  gender, race/ethnicity

9
Outcomes Short and Long Term
Index Year
Follow-up
Prior Year
Recurrent Event
Selected cohort
Other Vascular Event
Medical History
Non-Vascular Event
10
Recurrent 1-Year Stroke Rates
Allen, et al 2009
11
Percentage Change in Risk Standardized 1-Year
Recurrent Stoke Rates from 1994-1996 to 2000-2002

• 1-year recurrent stroke hospitalization rates
  decreased by 5 from 1994-1996 to 2000-2002

Allen, et al 2009
12
Sex-Based Outcomes following TIA Hospitalization?
122,063 hospital discharges for TIA (ICD-9 435)
in 2002
Stroke. 2009 Feb 19. Epub ahead of print
13
Coronary Artery Disease and Mortality
Stroke. 2009 Feb 19. Epub ahead of print
14
Cost Analysis
Index Event or Procedure

• Inpatient
• Outpatient
• Home Health
• Hospice
• Skilled Nursing Facility
• Physician Visits
• Durable Goods

Selected cohort
15
CMS Medicare Health Care Quality Improvement
Programs Enriched Datasets

• Include chart abstracted information
• Sample of Medicare patients
• Cooperative Cardiovascular Project
• National Stroke Project

16
Receipt of Medications at Discharge by State,
2000-2001
Antithrombotics Among Stroke/TIA Patients
Beta-Blockers Among AMI Patients
Modified from Jencks et al. JAMA.
2003289(3)305-312.
17
Advantages of Using CMS Data

• National perspective
• Patient level and hospital level analyses
• Aerial view of disease in the elderly
• Subgroups
• Time trends
• Utilization of resources
• Complements the perspective of cohort studies and
  registries

18
Limitations

• Accuracy of diagnostic codes
• Unmeasured factors in administrative data
• symptoms, test results, medical decisions
• No information on medication use
• Restriction to hospitalized events
• Underestimate true burden in community
• but .. reflects hospital resource utilization

19
Next Steps..

• Develop expertise in using administrative data
  for disease surveillance
• Combine data resources
• Social economic status
• Behavioral / lifestyle factors
• Access and availability of care
• Medications
• Cost data
• Registries with additional clinical detail