An Inventory Of Federally Sponsored HIV And HIV-Relevant Databases. Database: Medicare


Purpose Of The Database And Study Design: Medicare files are used to administer the beneficiary eligibility and claims processing responsibilities of the Medicare program. Medicare databases record information regarding all services billed on behalf of Medicare beneficiaries under its hospital (Part A) and supplemental (Part B) insurance plans. Part A covers acute care hospitalizations and stays in skilled nursing facilities (SNFs) for all Medicare beneficiaries. Part B covers physician, outpatient hospital, home health, and other medical services, such as diagnostic radiology and laboratory testing, for those individuals wishing to purchase supplemental coverage.

Nature Of The Data Collected: Longitudinal administrative records regarding Medicare eligibility and claims submitted for reimbursement by the Medicare program

Unit Of Analysis: The unit of analysis in the eligibility or Denominator File is the Medicare beneficiary. The unit of analysis in the claim files is the claim. Episodes of care and provider-focused files may also be constructed.

Data Collection Methods, Major Data Constructs And Key Data Elements: Eligibility files are maintained in the Denominator File. Demographic data in the Denominator File include beneficiary demographic characteristics, geographic data, date of death, enrollment date, HMO enrollment, Medicaid buy-in, if the beneficiary was enrolled in Medicare due to disability or other enrollment criteria, and other related eligibility data. The Denominator File is a sub-set of the Health Insurance Master File that maintains Medicare benefits data for all individuals eligible to receive Medicare benefits. Due to its size, the Health Insurance Master File is not commonly used for analytic purposes.

Several types of Medicare claims files are maintained. Each claim record includes summary beneficiary demographic characteristics. The Part A Medicare Provider Analysis and Review (MEDPAR) file has summary claims records that condense the services billed for an inpatient stay into a single claim. Data included in MEDPAR claims include: institutional provider number, admission and discharge date, length of stay, admission type, discharge destination, days covered by Medicare, coinsurance amount, inpatient deductible, lifetime reserve days, date that SNF benefits were exhausted, total charges, covered charges, amount reimbursed, service-specific charges (e.g., intensive care unit), up to five ICD-9-CM diagnostic codes, up to three ICD-9-CM procedure codes, dates of procedures, and DRG.

The Part B Medicare Annual Data Beneficiary File IV (BMAD-IV) has separate records for each claim to which selected provider and beneficiary data are attached. Data included in the BMAD-IV claims include: individual provider number, provider type, provider specialty, assignment/participation indicator, submitted charge, allowed charge, reimbursed charge, one CPT-4 procedure code, type of service, place of service, date of service, and units of service.

Outpatient Bill Records include: institutional provider number, dates of first and last service, total charges, covered charges, reimbursed amount, professional charges, up to five ICD-9-CM codes, up to three ICD-9-CM procedure codes, and CPT-4 code.

Hospice and Home Health Agency Bill Records include: provider number, date care began and ended, total visits, total charges, charge per unit, number of visits, reimbursement amount, and ICD-9-CM code.

The Institutional Provider of Services file summarizes the characteristics of inpatient facilities providing care to Medicare beneficiaries. The file includes: provider category, total beds, geographic location, residency programs, affiliations, accreditation, administrative control, service provided, number of salaried physicians, residents, nurses, and total staff.

General Attributes: The size of the files will vary based on the years and populations studied.

Strengths And Weaknesses Of The Study Design And Database: Medicare beneficiary and claims files may be useful in conducting a variety of observational studies of HIV positive individuals. The files have substantial limitations, however. They are large and complex and require substantial experience in their use to accurately construct and use analysis files. There are no direct identifiers for HIV positive Medicare beneficiaries. Indirect methods must be used, such as the application of ICD-9-CM coding nets. Since Medicare does not cover pharmaceuticals, coding nets must rely solely on diagnostic coding. Such coding may undercount HIV positive beneficiaries due to the stigmatization that may be related to being identified as being infected. Coding nets that rely solely on ICD-9-CM codes have relatively poor predictive value, sensitivity, and specificity. Once identified, no direct clinical marker data (e.g., CD4 count or viral load) are available to stage HIV positive beneficiaries. Claims files may not completely identify services received by Medicare beneficiaries, only those services for which a Medicare claim was submitted for payment. Moreover, Medicare beneficiaries may be enrolled in other insurance programs (e.g., Medicaid, Veterans Affairs, commercial insurance) or receive services through clinical trials or other sources. Other design factors associated with Medicare files make their use complex.

Gaps In The Data Collected And Factors Leading To The Gaps: Clinical marker data, date of HIV infection, and other sources of payment and care.

Feasibility Of Linking With Other Databases: To study persons with HIV and AIDS, Medicare eligibility and claims files have been linked to a variety of other databases. Some of these databases include: HARS, Medicaid, other insurance claims files, administrative records, observational databases, clinical trial databases, and the NDI.

Process To Access The Database And Contact Person: HCFA(now known as CMS) Office of Information Services, Enterprise Database Group, Division of Data Liaison and Distribution, (410) 786-3673.

Selected Citations:

Fasciano N, Cherlow A, Turner B, Thornton C. Profile of Medicare beneficiaries with AIDS: application of an AIDS case definition system.Health Care Financing Review, 19(3), 19-38, 1998.

Riley GF, Potosky AL, Lubitz JD, Kessler LG. Medicare payments from diagnosis to death for elderly cancer patients by stage at diagnosis.Medical Care. 33(8): 828-841, 1995.

Mitchell JB, Cromwell J. Impact of Medicare payment reductions on access to surgical services. Health Services Research. 30(5): 635-655, 1995.

Davis MH, Burner ST. Three decades of Medicare: what the numbers tell us. Health Affairs. 14(4): 231-243, 1995.

Gornick M, McMillan A, Lubitz J. A longitudinal perspective on patterns of Medicare payments. Health Affairs. 140-150, 1993.

Gaumer GL, Stavins J. Medicare use in the last ninety days of life. Health Services Research. 26(6): 725-742, 1992.