VA Databases and Methods Using Non-VA Data: Medicare

1. VA Databases and MethodsUsing Non-VA Data:Medicare VIReC Cyber Seminar Series 2006 Presented by Kristin A. Koelling, MPH Katherine Mallin, Ph.D.

2. Session Objectives • Why measuring Medicare healthcare use is not always as simple as it seems • Data source review: Data available through the VA-Medicare Data Merge Initiative • Measurement strategies for evaluating Medicare healthcare use • Where to go for help

3. Why measuring Medicare healthcare use is not always as simple as it seems • Medicare data are collected for billing purposes • What data are/are not collected? • Which services are in which files? • How can researchers define events using Medicare data?

4. Session Objectives • Why measuring Medicare healthcare use is not always as simple as it seems • Data source review: Data available through the VA-Medicare Data Merge Initiative • Measurement strategies for evaluating Medicare healthcare use • Where to go for help

5. VA-Medicare Data Merge Initiative • Centers for Medicare and Medicaid Services (CMS) & VHA Interagency Agreement • Medicare data since 1997 are available through VHA • For researchers through VIReC

6. VHA Cohort • Over 12 million vets • VHA users (since 1992) • Enrollees (since 1999) • Compensation & Pension Beneficiaries (since 2000) • Medicare data available only after entering VHA cohort • Example: Veteran enrolled in Medicare in 1997, but never used or enrolled in VHA until 2002 • VIReC has Medicare data for this veteran only from 2002 forward

7. Data Available • Medicare data (Calendar years 1997 – 2004) • Vital Status and Denominator files • Claims • Provider data • Medicare Current Beneficiary Survey (MCBS) • United States Renal Data Systems (USRDS)

8. Vital Status Beneficiaries ever enrolled in Medicare (Cumulative) Demographics More current than Denominator Denominator Beneficiaries enrolled in a given year Demographics Entitlement and enrollment indicators Reasons for entitlement (old age, disability, ESRD) Part A and/or Part B enrollment Managed care enrollment indicators Vital Status & Denominator Files

9. Sources of Claims Data

10. Claims Files • Institutional Standard Analytic Files (SAF) • Outpatient • Home Health Agency (HHA) • Hospice • Inpatient • Skilled Nursing Facility (SNF) • Institutional Stay Level File • MedPAR – Inpatient and SNF stays • Non-institutional Standard Analytic Files (SAF) • Carrier (Physician/Supplier) • Durable Medical Equipment (DME)

11. Which files contain outpatient services? • Because some outpatient services are billed by the physician and some by the facility, use both Carrier and Outpatient files for: • Laboratory • Ambulance • ER visits* • All other outpatient procedures and services *ER visits that result in hospitalization will be combined with the Inpatient claim

12. Which files contain inpatient services? • The Inpatient and MedPAR files contain services provided by the hospital, including physicians on staff • Carrier file will contain services performed by physicians not employed by the hospital • 16.5% of Carrier claims occur in an inpatient hospital

13. Session Learning Objectives • Why measuring Medicare healthcare use is not always as simple as it seems • Data source review: Data available through the VA-Medicare Data Merge Initiative • Measurement strategies for evaluating Medicare healthcare use • Where to go for help

14. Medicare data is used for billing • Files based on type of service and type of bill • No/Limited data on: • Data not needed for billing • Services not billed • Data directly related to billing is most accurate • One claim doesn’t always indicate one event

15. What is not included in Medicare data? • Data not needed for billing • Marital status • Education, income • Smoking status • Vital signs (height, weight, blood pressure, pulse) • Lab results

16. What is not included in Medicare data? • Services not billed • Prospective payment system (PPS) • “Single” payment is made to cover all services • Procedure & diagnosis codes may not be required for reimbursement • Examples: Inpatient, SNF, HHA, Hospice • Managed care (HMOs) • In 2000, 17.5% of veterans were enrolled in managed care at least 1 month • Plans are paid a capitated monthly rate • Claims are not submitted to Medicare

17. What is included in the Medicare data? • Costs • Charges submitted • Payments made by CMS, beneficiaries (deductibles & co-payments), & primary payers • Provider numbers • Dates • Claim or billing dates • Dates of service

18. Dates • Be careful when using dates of claim • Prospective payment system • Hospice – up to 30 days of care on single claim • HHA – up to 60 days of care on single claim • Sometimes can use revenue center dates • Facilities may submit claim before discharge • SNF – half of stays span multiple claims • Use admission & discharge dates

19. What is included in the Medicare data? • ICD-9 Diagnosis codes • ICD-9 Procedure codes • HCPCS - Healthcare Common Procedure Coding System • Procedures, supplies, products and services • CPT procedure codes + others • DRG - Diagnostic Related Group • Used in calculating facility payments for inpatient stays

20. Measuring Medicare Healthcare Use • Stay • Inpatient/SNF claims – not at stay level • MedPAR – stay level file • Claims • Dates • Procedures • Costs

21. Measuring Medicare Healthcare Use: Examples from Three Research Studies • Measuring types of outpatient use • Weeks WB et al. ‘Veterans Health Administration and Medicare Outpatient Health Care Utilization by Older Rural and Urban New England Veterans’ Journal of Rural Health, 167-171, Spring 2005 • Measuring outpatient use using dates • Hynes D, et al. Veterans’ Access to and Use of Medicare and VA Health Care. Medical Care, in press • Measuring procedures • Halanych JH et al. ‘Racial/Ethnic Differences in Diabetes Care for Older Veterans: Accounting for Dual Health System Use Changes Conclusions’ Medical Care, 44(5):439-445, 2006.

22. Measuring types of outpatient use Example from Weeks et al, Journal of Rural Health, 2005 • Goal • Compare outpatient medical health care utilization for urban and rural Medicare enrolled veterans • Cohort • New England VHA users (1995-1999) 65 years and over, enrolled in Medicare Fee for Service

23. Measuring types of outpatient use Example from Weeks et al, Journal of Rural Health, 2005 • Methods • Claims files – Outpatient and Carrier files • Categorized outpatient use using • Medicare “Type of Service” Codes • VHA “Stop Codes” • Number of outpatient visits assigned to primary care, mental health, other specialty, emergency room care, and other outpatient care

24. Measuring types of outpatient use Example from Weeks et al, Journal of Rural Health, 2005 • Selected Results • All veterans obtained 2-3 times more primary care in Medicare as VHA • Urban veterans had more specialty care visits in VHA than in Medicare • Urban and rural veterans had more mental health visits in VHA than in Medicare

25. Measuring outpatient use • Benefits of measuring use with claims • Allows categorization of types of care • Caveats • One event may be recorded in multiple claims • i.e. radiology • One claim may contain more than one service • Occurs less often in outpatient files • Some inpatient claims billed prospectively

26. Measuring outpatient use using datesExample from Hynes et al, Medical Care, in press • Goals • To measure VA and Medicare reliance among dually eligible veterans • Cohort • 2.6 million veterans in VHA cohort enrolled in Medicare Fee-for-Service in 1999

27. Measuring outpatient use using dates Example from Hynes et al, Medical Care, in press • Methods – Measuring Outpatient Reliance • Claims files – Outpatient, HHA, Carrier • Divided Carrier file into inpatient and outpatient services • Based on dates in Inpatient file, or • Place of Service Codes in Carrier file • Counted unique dates of service in VA and Medicare

28. Measuring outpatient use using dates • Benefits of using Dates • Easy to calculate • Disadvantages of Using Dates • Different meanings in VA vs. Medicare • VA – often multiple “events” in 1 day • Medicare – often different day for each “event”

29. Measuring procedures Example from Halanych et al, Medical Care, 2006 • Goal • Compare racial/ethnic disparities in quality of care for VHA patients with diabetes • Cohort • National sample of non-institutionalized VHA patients, 65 years and over, with diabetes • Diabetes determined from VHA prescription and ICD-9 codes and from Medicare ICD-9 diagnostic codes for FY 1997-1998

30. Measuring proceduresExample from Halanych et al, Medical Care, 2006 • Methods • Race/ethnicity obtained from both VHA and Medicare data • Claims files – Inpatient, Outpatient, Carrier • Care measurements • VHA data-- Laboratory result files and CPT codes used • Medicare data – CPT/HCPCS codes used (outpatient, carrier) ICD-9 procedure codes (inpatient) • Three quality of care measures used • Annual hemoglobin A1c (HbA1c) test • Annual Low Density Lipoprotein (LDL) test • Annual Dilated eye examination

31. Measuring procedures Example from Halanych et al, Medical Care, 2006 • Results • VHA only data-Rates of receiving three procedures were equal to or higher for Blacks and Hispanics than for white patients (except LDL testing in Black patients) • VHA + Medicare data– Results were reversed. Rates of receiving procedures were lower in Blacks and Hispanics compared to whites • More care captured in Medicare for Whites as compared to Blacks and Hispanics

32. Measuring procedures • Benefits of Procedure codes • Claims types where procedure codes are required for billing are likely to be complete and accurate • One example: HCPCS in Carrier File • Limitations of Procedure codes • May be incomplete for some types of care • One example: Only up to 6 procedures allowed on inpatient claims

33. Session Learning Objectives • Why measuring Medicare healthcare use is not always as simple as it seems • Data source review: Data available through the VA-Medicare Data Merge Initiative • Measurement strategies for evaluating Medicare healthcare use • Where to go for help

34. VIReC’s VA-Medicare Web Page • Links to Medicare Files Data dictionaries • SAS Proc Contents • Frequencies for selected Medicare variables • List of source files for VHA cohort • Instructions for requesting Medicare data from VIReC • www.virec.research.va.gov/DataSourcesName/VA-MedicareData/Background.htm • Links to Medicare Files Data dictionaries • SAS Proc Contents • Frequencies for selected Medicare variables • List of source files for VHA cohort • Instructions for requesting Medicare data from VIReC • www.virec.research.va.gov/DataSourcesName/VA-MedicareData/Background.htm • Links to Medicare Files Data dictionaries • SAS Proc Contents • Frequencies for selected Medicare variables • List of source files for VHA cohort • Instructions for requesting Medicare data from VIReC • www.virec.research.va.gov/DataSourcesName/VA-MedicareData/Background.htm

35. Research Data Assistance Center (ResDAC) • CMS contractor based at the University of Minnesota • Provides free assistance to • Researchers • Government agencies • Not-for-profit organizations • CMS 101: Introduction to the use of Medicare data for research • www.resdac.umn.edu

36. CMS and Medicare Websites • CMS Home Page • www.cms.hhs.gov • CMS Research, Statistics, and Data Systems • www.cms.hhs.gov/home/rsds.asp • For Medicare beneficiaries • www.medicare.gov • CMS Home Page • www.cms.hhs.gov • CMS Research, Statistics, and Data Systems • www.cms.hhs.gov/home/rsds.asp • For Medicare beneficiaries • www.medicare.gov