Purpose Of The Database And Study Design: MSIS was established by HCFA(now known as CMS) in 1984 as a voluntary system for States and was mandated nationally for FY 1999 to standardize the content and structure of Medicaid eligibility and claims files among State Medicaid Programs. Early Medicaid eligibility and claims files (1984-1998) may vary somewhat but contain most of the variables maintained in the 1999 MSIS files.
Nature Of The Data Collected: Longitudinal administrative records regarding Medicaid eligibility and claims submitted for reimbursement by the Medicaid program
Unit Of Analysis: The unit of analysis in eligibility records is the Medicaid beneficiary. The unit of analysis in claim files is the claim. Episodes of care and provider-focused files may also be constructed.
Data Collection Methods: MSIS files are organized in five major groups: Medicaid eligibility, inpatient claims, long-term care, other claims/encounters, and pharmacy claims/encounters. Each set of files contains the recipient’s unique identifier that allows linkage across all files to construct person-based analysis files. The unique identifier may be a Social Security Number or other unique number that is issued only once to a recipient and not reused.
Eligibility records are maintained in the eligibility portion of the MSIS. Eligibility data are recorded by local departments of public welfare. Eligibility files are usually updated on a weekly basis. The eligibility record contains several types of root fields: fixed variables that do not change during enrollment in the Medicaid account (e.g., sex); non-fixed variables that may be overwritten when a change occurs (e.g., ZIP code), and additive variables where longitudinal changes are recorded (e.g., eligibility periods).
Health care providers submit automated or hardcopy claims to the State Medicaid program for processing and payment. Claims files are commonly updated on a weekly basis. MSIS uses four distinct claims file structures: inpatient claims, long-term care claims, outpatient claims, and prescription drug claims. The claim file tape must contain one record for each claim of the appropriate type paid or encounters processed during the reporting period. Each claim file must include: one record for each line item that is separately adjudicated; all fully adjudicated claims that have completed the State’s processing cycle for which the State has determined that it has liability to reimburse the provider; adjudicated claims that have passed all the State’s automated edits, but which resulted in a zero liability because of payments by responsible third parties; claim records representing capitated payments or fees paid to capitated plans; encounter claims, to the extent that they are routinely received by the State; and Medicare/Medicaid crossover claims.
Effective for FY 1999 data, claims files may contain several types of records: current fee-for-service claims for medical services, capitated payments, and encounter claims. Encounter or “pseudo-claims” simulate claims that would have been generated for members of HMOs and health insuring organizations (HIOs), patients in prepaid health plans (PHP), and recipients in primary care case management (PCCM) if their claims were billed on a fee-for-service basis. Some States also use “service-tracking claims” for special purposes, such as tracking individual services covered in a lump-sum billing. Adjustment claims are identified and categorized by the Adjustment indicator field. Another claim type identifies supplemental payment above the capitation fee or above the negotiated rate, such as in the case of Federally Qualified Health Center (FQHC) additional reimbursement.
The inpatient file includes any service claim billed as an inpatient hospital service. This file also includes records for services billed by religious non-medical institutions. Inpatient psychiatric services provided in a separately administered psychiatric wing or psychiatric hospital are not considered acute and are not included in the inpatient file. Psychiatric hospital claims are included in the long term care claims file.
The long term care file includes Title XIX claims for long-term care services received in an institution such as: nursing facilities (NFs), intermediate care facilities for the mentally retarded (ICF-MRs), psychiatric hospitals, and independent freestanding psychiatric wings of acute care hospitals.
The drug claim file records identify Title XIX claims for prescription drugs, durable medical equipment, and supplies provided by a pharmacist under a prescription. Injectibles and other drugs dispensed as a bundled service are reported for the provider administering the service (e.g., physician-administered inoculations are reported on the outpatient file as a physician service).
All other claims not included in the other specific claims groups are in the outpatient file. The outpatient claims file includes: provider claims for all non-institutional Medicaid services (e.g., office-based physician visits, hospital-based ambulatory claims, etc.); provider claims for services received in hospitals, NFs, and ICF/MRs that are not billed as part of the inpatient or long-term care claims (e.g., physician visits in those facilities, services of private duty nurses, encounters); capitated payments; and claims for medical and non-medical services received under a Title XIX waiver. Outpatient records may contain bills for multiple service units (e.g., several physician visits for the same illness, multiple laboratory tests conducted on the same day).
General Attributes: MSIS data files will vary substantially among State Medicaid programs, reflecting the number of Medicaid enrollees and the number of claims processed.
Major Data Constructs And Key Data Elements: Eligibility fixed field variables include: MSIS identification number (a unique identification number used to identify a Medicaid recipient in MSIS), birth date, death date, sex, race/ethnicity, county of residence, ZIP code, dual eligible flag, and Medicare Health Insurance Claim. Monthly fields completed for each month of eligibility include: days of eligibility, eligibility group, maintenance assistance status, basis of eligibility, health insurance, Temporary Assistance to Needy Families (TANF) cash flag, restricted benefits flag, up to four managed care plan types under which the recipient is covered during the month, and State Children’s Health Insurance Program (S-CHIP) code.
Inpatient variables include: program type, Medicaid amount paid, beginning and end service dates, provider identification number, amount charged, other third party payment, Medicaid-covered inpatient days, Medicare deductible payment, Medicare coinsurance payment, up to nine diagnosis codes (ICD-9-CM), up to six procedure codes (ICD-9-CM procedure or CPT codes), admission and discharge dates, patient status, DRG, and up to 23 UB Revenue Codes
Long term care variables include: Medicaid amount paid, beginning and end service dates, provider identification number, amount charged, other third party payment, Medicare deductible payment, Medicare coinsurance payment, five diagnosis codes, admission and discharge date, patient status, ICF-MR days, leave days, nursing facility days, and patient liability.
Prescription drug variables include: Medicaid amount paid, date prescribed, provider identification number, amount charged, other third party payment, quantity (e.g., number of units of a prescription that were filled), days supply, National Drug Code (NDC), prescription fill date, and prescribing physician identification number.
Outpatient variables include: type of claim, date of payment, Medicaid amount paid, beginning and end dates of service, provider identification, amount charged, other third party payment, quantity of service, Medicare deductible payment, Medicare coinsurance payment, two diagnosis codes, place of service, specialty code, service (e.g., CPT-4, ICD-9-CM, and HCPCS), UB-92 revenue code, and provider identification number.
Strengths And Weaknesses Of The Study Design And Database: MSIS is being restructured as a national database for FY 1999, as described above. The system in place prior to 1999 has different variable names, but the design is similar to FY 1999 files. State participation was optional and included approximately 30 States by FY 1998. MSIS eligibility and claims files are 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 Medicaid beneficiaries. Indirect methods must be used, such as the application of ICD-9-CM, NDC, and State-generated specialty HIV rate coding nets. These nets, however, have varying degrees of 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 Medicaid beneficiaries, only those services for which a Medicaid claim was submitted for payment. Moreover, Medicaid beneficiaries may be enrolled in other insurance programs (e.g., Medicare, VA, commercial insurance) or receive services through clinical trials or other sources. Other design factors associated with MSIS make its use complex.
Gaps In The Data Collected And Factors Leading To The Gaps: Clinical marker data, date of HIV infection, death date (often missing from the MSIS eligibility file or delayed in recording), and other sources of payment and care.
Feasibility Of Linking With Other Databases: To study persons with HIV and AIDS, Medicaid eligibility and claims files have been linked to a variety of other databases. Some of these databases include: HARS, Medicare, 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) Center for Medicaid and State Operations, Data and Systems Group, Division of Information, Analysis, and Technical Assistance, (410) 786-0780.
Turner BJ, Newshaffer CJ, Cocroft J, Fanning T. Improved birth outcomes among HIV positive women with enhanced Medicaid prenatal care.AJPH. 90(1): 85-91, 2000.
Turner, BJ, Cocroft J, Newschaffer CJ, Hauck WW, Fanning TR. Sources of prenatal care data and their association with birth outcomes of HIV positive women. AJPH. 20: 118-121, 2000.
Wutoh A, Hidalgo J, Rhee W, Bareta J, et al. Survival differences associated with treatment of cytomegalovirus retinitis in Maryland patients with AIDS, 1987-1994. American Journal of Health-System Pharmacy. 56: 1314-1318, 1999.
Wutoh A, Hidalgo J, Rhee W, Bareta J. A characterization of older AIDS patients in Maryland. Journal of the National Medical Association. 90(6): 369-373, 1998.
Walkup J, Sambamoorthi U, Crystal S. Characteristics of persons with mental retardation and HIV/AIDS infection in a statewide Medicaid population. American Journal of Mental Retardation. 104(4): 356-363, 1999.
Walkup J, Crystal S, Sambamoorthi U. Schizophrenia and major affective disorder among Medicaid recipients with HIV/AIDS in New Jersey.AJPH. 89(7): 1101-1103, 1999.
Gebo KA, Chaisson RE, Folkemer JG, Bartlett JG, Moore RD. Costs of HIV medical care in the era of highly active antiretroviral therapy.AIDS. 13(8): 963-969, 1999.
Laine C. Markson LE, McKee LH, Hauck WW, et al. The relationship of clinic experience with advanced HIV and survival of women with AIDS. AIDS. 12: 417-424, 1998.
Turner BJ, Hauck WW, Fanning TR, Markson LE. Cigarette smoking and maternal-child HIV transmission. Journal of AIDS and Human Retrovirology. 14:327-337, 1997.
Turner BJ, Markson LE, Taroni F. Estimation of survival after AIDS diagnosis: CD4 T-lymphocyte count versus clinical severity. Journal of Clinical Epidemiology. 49:59-65, 1996.
Turner BJ, McKee L, Silverman NS, Hauck WW, et al. Prenatal care and birth outcomes of a cohort of HIV positive women. Journal of AIDS and Human Retrovirology. 12:259-267, 1996.
Friedman L, Hidalgo J, Bartnyska L, Turner B. Analyzing survival of adults using a severity classification system for AIDS hospitalizations.Medical Care, 34(2): 178-189, 1996.
Turner BJ, Markson LE, Hauck WW, Cocroft J, et al. Prenatal care of HIV positive women: analysis of a large New York State cohort.Journal of AIDS and Human Retrovirology. 9:371-378, 1995.
Fanning TR, Turner BJ, Cosler LE, et al. Quality of Medicaid data for HIV/AIDS research: examination of a statewide database. AIDS and Public Policy. 10: 39-47, 1995.
Bartnyska L, Schactman M, Hidalgo J. Patterns in Maryland Medicaid enrollment among persons with AIDS. Inquiry. 32: 184-195, 1995.
Markson LE, Cosler LE, Turner BJ. Implications of generalists’ slow adoption of zidovudine in clinical practice. Archives of Internal Medicine. 154: 1497-1504, 1994.
Moore R, Hidalgo J, Bareta J, Chaisson RE. Zidovudine therapy and health resources utilization in AIDS. JAIDS, 7(4): 349-354, 1994.
Moore R, Hidalgo J, Sugland B, Chaisson R. Zidovudine and the natural history of the Acquired Immunodeficiency Syndrome, NEJM. 324(20): 1412-1416, 1991.
Agency: Health Care Financing Administration(now known as Centers for Medicare and Medicaid Services(CMS)) (HCFA(now known as CMS))