Using Microsimulation to inform Targeted Cardiovascular Disease Prevention Policy DEHMER
- 1. USING MICROSIMULATION TO INFORM TARGETED CARDIOVASCULAR DISEASE PREVENTION POLICY STEVEN DEHMER, PHD RESEARCH FELLOW, HEALTHPARTNERS RESEARCH FOUNDATION Monday, April 30th HMORN 2012 Conference, Seattle WA
- 2. Presentation Outline 1) Brief Project Background 2) Description of the CVD Prevention Policy Model 3) Brief Description of Results
- 3. Project Background Context: Need to identify priority areas for prevention (NCPP) 1) Evidence driven 2) Consistent and comparable 3) Account for disparities and population-specific effects Purpose: a) Assess USPSTF recommended cardiovascular clinical services b) Assess cardiovascular impacts of other clinical and community services
- 4. Project Background Cardiovascular Clinical Preventive Services (USPSTF): 1) Screening for lipid disorders in adults Men (35+ or 20-35 at ↑ risk); Women (20+ at ↑ risk) 2) Hypertension screening for adults (18 and older) 3) Aspirin counseling for primary prevention of CVD Men ages 45-79 with increased risk from myocardial infarction Women ages 55-79 with increased risk from stroke Balance potential CVD benefits with gastrointestinal bleeding risks
- 5. CVD Prevention Policy Model Design Overview: “Microsimulation” model: start with an individual and predict lifetime progression of health status and outcomes Simulations of many individuals can be aggregated to estimate population-wide impacts Interventions or counterfactuals tested as if in a randomized controlled trial (here: same people and all else held equal) Key benefit of complex design: sub-population effects for informing targeted policy
- 6. CVD Prevention Policy Model Design Overview: Demographic characteristics: Age 20 to 100 Sex: Male, Female Race/ethnicity: Non-Hispanic white Non-Hispanic black/African American Hispanic/Mexican American “Other” Health characteristics (change as an individual ages): Body mass index (BMI) Cholesterol: LDL, HDL Systolic blood pressure (SBP) Smoking status Disease status
- 7. CVD Prevention Policy Model Design Overview: Cardiovascular disease events: Myocardial infarction (MI) Ischemic stroke (IS) Hemorrhagic stroke (HS) Congestive heart failure (CHF) Angina pectoris (AP) Intermittent claudication (IC) Related disease: Diabetes Mortality: CVD related death Non-CVD death
- 8. CVD Prevention Policy Model
- 9. CVD Prevention Policy Model Model Initialization: NHANES (1999-2008) • Sex • Race/ethnicity • HDL, LDL • SBP BRFSS (2009) • BMI
- 10. CVD Prevention Policy Model Eligibility for Prevention: USPSTF Recommendations: 1) Lipid screening 2) BP screening 3) Aspirin counseling Delivery of untested recommendations set at contemporary rates (NHANES, 1999-2008)
- 11. CVD Prevention Policy Model Eligibility for Treatment: Based on National Clinical Guidelines: • ATP III for lipids • JNC-7 for hypertension • USPSTF for aspirin
- 12. CVD Prevention Policy Model Treatment Effects: Meta-analyses and literature reviews (evidence from major clinical trials) • Lipid treatment with statins lowers LDL, raises HDL • Hypertension treatment lowers SBP • Aspirin treatment lowers MI risk in men, ischemic stroke risk in men; raises HS and GI bleeding risk in all
- 13. CVD Prevention Policy Model CVD Event Risk: Customized 1yr risk equations estimated using Framingham Heart Study Data • Includes original and offspring cohorts • About 10,000 people • Longitudinal design from 1950-2003 • Mostly white
- 14. CVD Prevention Policy Model Risk of Death: CVD-death risk estimated using Framingham Heart Study data Death from other causes estimated using life tables (net of CVD mortality)
- 15. CVD Prevention Policy Model Progression of Risk Factors: Estimated using a two-step process: 1) Determine if there is a change 2) Determine size of change Cholesterol and BP changes from Framingham Heart Study data Changes in BMI from BRFSS (2009)
- 16. CVD Prevention Policy Model Costs: 1) Costs of disease First-year and ongoing costs estimated from MEPS (1999-2008) 2) Costs of screening/monitoring Clinic/lab fees from National Fee Analyzer (2005) 3) Costs of treatment Statin and antihypertensive treatment costs from Express Scripts Drug Trend Report (2010)
- 17. Simulation Results Evaluation of USPSTF Recommendations : Preliminary estimates (2012): 1) Screening for lipid disorders in adults C/E: $50,000 per QALY CPB: 650,000 QALYs 2) Hypertension screening for adults C/E: $65,000 per QALY CPB: 500,000 QALYs 3) Aspirin counseling for primary prevention of CVD C/E: $100 saved per person CPB: 150,000 QALYs
- 18. Simulation Results: Disparities Hypertension Reference Case, No Screening, Rates per 100k Total Men Women White Black Hispanic Other Myocardial 29,301 38,200 20,091 28,234 33,931 29,864 27,745 Infarction Ischemic Stroke 18,247 17,292 19,235 17,838 20,388 18,370 17,188 Angina Pectoris 21,355 25,080 17,499 20,948 22,656 21,722 21,256 Congestive 29,381 29,404 29,357 27,684 35,004 30,892 28,656 Heart Failure Intermittent 10,498 12,204 8,732 10,232 12,200 10,200 10,182 Claudication CVD-related 38,055 42,758 33,187 37,293 41,653 38,167 37,133 Death Life Expectancy 78.96 76.02 82.01 79.06 78.58 78.96 78.93
- 19. Conclusion Questions/Comments?