Health Economics: Hot Topics and Research in Progress

1. Health Economics: Hot Topics and Research in Progress Richard E. Nelson, PhD Division of Epidemiology University of Utah School of Medicine Salt Lake City Veterans Affairs Healthcare System

2. Presentation Outline • Brief overview of healthcare costs in the US • Affordable Care Act • Oregon Health Insurance Experiment • Cost of healthcare-acquired infections • Methods • Application of these methods to VA data

3. Economics • “Economics examines economic events and arrangements through the lens of economic theory” • The study of how individuals, governments, firms, and nations allocate scarce resources to satisfy their unlimited wants • The study of choices

4. Health Economic Evaluation • Bang for the buck • Inputs (costs) • Outcomes (benefits) • Cost-effectiveness • Achieving objective at least cost, or • Maximizing benefits from given amount of resources

5. Total Healthcare Expenditures per Capita OECD. 2010

6. Total Health Expenditures as a Share of GDP, 2008 OECD. 2010

7. Commonwealth Fund 2013

8. Why Cross-Country Differences in Healthcare Expenditures • Administrative costs • US = 25% of healthcare expenditures • Other countries = 10-15% of healthcare expenditures • Duke University • 900 beds • 1,300 billing clerks • Typical Canadian hospital • 10 billing clerks David Cutler, Harvard University

9. Why Cross-Country Healthcare Expenditure Differences? IFHP 2012 Comparative Price Report

10. Why Cross-Country Healthcare Expenditure Differences? IFHP 2012 Comparative Price Report

11. Why Cross-Country Healthcare Expenditure Differences? IFHP 2012 Comparative Price Report

12. Why Cross-Country Differences in Healthcare Expenditures • The same patients get more medical care in the US • Ontario, Canada • 11 hospitals that can do open heart surgery • Pennsylvania • 60 hospitals that can do open heart surgery • Life expectancy and one-year mortality following heart attack roughly the same David Cutler, Harvard University

13. What do we get for our healthcare dollars?

14. What do we get for our healthcare dollars?

15. Geographic variation in health care spending Institute of Medicine 2013

16. Lowest and Highest Spending Medicare HRRs Lowest Highest Rochester, NY Stockton, CA Sacramento, CA Buffalo, NY Bronx, NY Santa Cruz, CA Santa Rosa, CA Medford, OR San Francisco, CA Salem, OR Miami, FL McAllen, TX Monroe, LA Houston, TX Alexandria, LA Lafayette, LA Shreveport, LA Baton Rouge, LA Fort Lauderdale, FL Metairie, LA Institute of Medicine 2013

17. Geographic variation in health care spending Baicker and Chandra (2008) Health Affairs

18. Geographic variation in healthcare spending • Potential reasons • Differences in prices paid for similar services • Differences in illness between regions • Differences in volume of health care services received by similar patients

19. Geographic variation in healthcare spending • Why higher volume of care • More effective care? • More preference-sensitive care? • More supply-sensitive care?

20. Geographic variation in healthcare spending • Higher volume of care does not produce better outcomes for patients • Worse adherence to evidence-based guidelines1-3 • Worse mortality after heart attack or hip fracture4 • Worse communication among physicians5 • Worse access to care and greater waiting times4 • Worse patient-reported inpatient experience6 Fisher et al (2003) Ann Intern Med Baicker et al (2004) Health Aff Fisher et al (2004) Health Aff Fisher et al (2003) Ann Intern Med Sivovich et al (2006) Ann Intern Med Wennberg et al (2009) Health Aff

21. Affordable Care Act • Signed into law March 23, 2010 • Major components • Individual mandate • Employers must offer insurance coverage • No denying coverage if preexisting condition • Creating health insurance exchanges • Expand Medicaid

22. Affordable Care Act • Medicaid expansion • Prior to ACA • Pregnant women and children < 6 with family incomes < 133% of FPL • Children age 6-18 with family incomes < 100% of FPL • Parents, caretaker relative meeting certain financial eligibility requirements • Elderly and disabled individuals who qualify for Supplementary Security Income

23. Affordable Care Act • Medicaid expansion • After ACA • All non-Medicare eligible individuals < 65 up to 133% FPL • $14,856 for individual in 2012 • $30,657 for family of 4 in 2012 • Federal government pays for expansion • Supreme Court decision 2012 • Medicaid expansion violates Congress’ spending clause power

24. Affordable Care Act

25. Oregon Health Insurance Experiment • Oregon Medicaid • Did not allow new enrollment from 2004-2008 due to budget constraints • Expanded in 2008 • Excess demand • So created a lottery • Treatment group = 29,834 • Control group = 45,088 • Sneak peak at possible impacts of ACA

26. Oregon Health Insurance ExperimentResults • Increased hospital admissions Finkelstein, et al Quarterly Journal of Economics(2012)

27. Oregon Health Insurance ExperimentResults • Increased Rx, outpatient encounters Finkelstein, et al Quarterly Journal of Economics(2012)

28. Oregon Health Insurance ExperimentResults • Reduced probability of unpaid medical bill sent to collection agency Finkelstein, et al Quarterly Journal of Economics(2012)

29. Oregon Health Insurance ExperimentResults • Increased self-reported health and probability of not screening positive for depression Finkelstein, et al Quarterly Journal of Economics(2012)

30. Oregon Health Insurance ExperimentResults • Increased ED use Taubman, et al Science (2014)

31. Oregon Health Insurance Experiment and ACA • Summary • Improvements in self-reported health • Decreases in financial hardship • Increases in healthcare utilization

32. HAI and MRSA • Healthcare-acquired infections (HAI) • Infections that result from encounters with healthcare system • About 1 in 20 hospitalized patients in US • Methicillin-resistant Staphylococcus aureus (MRSA) • Bacteria resistant to many antibiotics • One of the leading causes of invasive infections in healthcare settings • Bloodstream, pneumonia, and surgical site infections

34. Accurate cost of HAIs • Nicholas Graves • The purpose of cost-of-illness studies for HAIs is to inform decisions about how to reduce HAIs • If we know how much they cost, we will know how much we will save if they are prevented • 2 measures of cost appropriate for HAIs • Excess length of stay • Opportunity costs associated with lost bed-days • Variable inpatient costs • Variable vs. fixed costs

35. Accurate cost of HAIs • Excess LOS • Variable (and total) inpatient costs • Post-discharge costs

36. Goal of my current research • Estimate the cost per healthcare-acquired MRSA infection in the VA using these 3 components: • Excess LOS • Variable (and total) inpatient costs • Post-discharge costs • And use that estimate to estimate the budget impact of VA MRSA Prevention Initiative

37. Veterans Affairs MRSA Prevention Initiative • Began October 2007 • Consisted of a “bundle” of prevention strategies • Universal nasal surveillance for MRSA • Contact precautions for patients colonized or infected with MRSA • Hand hygiene • Institutional change • HAI prevention is everyone’s responsibility

40. Estimating cost of MRSA HAI in VA • Need way of identifying healthcare costs • VA DSS data • Activity-based accounting system in VA • Extracts information from general ledger and VA payroll system • Specific job categories, supplies or equipment • Costs are allocated to cost centers • Primary care clinics • Intensive care units • Administration • Environmental services • Costs are allocated based on employee activities

41. Estimating cost of MRSA HAI in VA • Need way of identifying MRSA infections • ICD-9 code (V09) is not good for MRSA HAIs • V09 = infection with drug-resistant microorganisms • Microbiology data • Unstructured Schweizer et al ICHE 2011

42. VA Microbiology Data

43. Progress to date • Excess LOS • In progress • Variable (or total) inpatient costs • Preliminary results • Post-discharge costs • Preliminary results

44. 1. Impact of HAI on Excess LOS • Important because each extra bed-day taken up by a patient with HAI represents opportunity cost for hospital • Many studies compare total LOS between patients with HAI and those without • But not all of the days are attributable to the HAI • This leads to “time-dependent bias” HAI Patient 1 Admission Discharge Patient 2 Discharge Admission Barnett et al AJE (2009) Barnett et al Value in Health (2011)

45. 1. Impact of HAI on Excess LOS • Multi-state models (Beyersmann method) HAI (1) Admission (0) Discharge/death(2)

46. 1. Impact of HAI on Excess LOS • Using VA data to estimate this • In progress

47. 2. Impact of HAI on Inpatient Costs • Many studies compare total inpatient costs between patients with HAI and those without • But not all of the costs are attributable to the HAI • This leads to “time-dependent bias” HAI Patient 1 Admission Discharge Patient 2 Discharge Admission

48. 2. Impact of HAI on Inpatient Costs • Can we identify costs before and after HAI with VA data? • Separate observations for each patient-treating specialty-calendar month 2009-11-01 2009-12-01 txsp 52 txsp 63 txsp 52 txsp 63 txsp 22 5 treating specialties txsp 22 6 observations txsp 63 txsp 52 txsp 63 txsp 52

49. 2. Impact of HAI on Inpatient Costs • Options to separate pre-HAI costs from post-HAI costs • Hope that patients with HAI had a new treating specialty • Try to get daily costs for all admitted patients • Exploit the quirk that generates a new observation each month

50. Option 2 • GEE model on patient-day data • Gamma distribution • DSS Daily Cost Resource (DCR) • Daily inpatient costs • DSS Production-Level Data Admitdt Dischdt HAI Patient 1 Day 1 Day 2 Day 5 Day 9 Day 11 Day 3 Day 4 Day 6 Day 7 Day 8 Day 10 Admitdt HAI Dischdt Patient 2 Day 1 Day 2 Day 5 Day 9 Day 11 Day 3 Day 4 Day 6 Day 7 Day 8 Day 10 Day 12 No HAI Admitdt Dischdt Patient 3 Day 5 Day 1 Day 2 Day 5 Day 3 Day 4 Day 6