Medicare Claims Data

Uses for Asthma Surveillance

Several questions can be addressed statewide using Medicaid Data:

What are the annual rates of hospitalizations for asthma among Medicare recipients?
How do rates of asthma hospitalizations vary by age, sex, race/ethnicity, and county among Medicare recipients?
What are the annual rates of ED visits and office visits for asthma among Medicare Part B recipients?
What is the prevalence of “persistent asthma” among Medicare recipients with Part B and supplemental medication coverage?
What is the direct annual cost of asthma medications and asthma-related services per Medicare recipient with supplemental medication coverage?
What percentage of Medicare recipients with asthma and supplemental medication coverage are receiving appropriate long-term control medications?

History of Medicare Data Collection and suggested Asthma Case Definitions

The CMS administers Medicare, the nations’ largest health insurance program, which covers nearly 40 million Americans. Persons aged 65 years and older, some disabled people under age 65, and people with End-Stage Renal Disease are eligible for Medicare. Medicare has two parts, Part A and Part B. Most people get Part A automatically when they turn age 65. They do not have to pay a monthly payment called a premium for Part A because they or a spouse paid Medicare taxes while they were working. Part B is supplemental medical coverage that requires an additional monthly fee. Part B covers doctors’ services, outpatient hospital care, and some other medical services (e.g., the services of physical and occupational therapists, and some home health care) that Part A does not cover. Part B helps pay for these covered services and supplies when they are medically necessary. Some private companies also contract with the Medicare program to offer Medicare health plans. These are called Medicare Plus Choice (MPC) plans. Managed care MPS plans charge low-to-medium out-of-pocket costs, require a person to see a participating physician, but provide additional coverage for things like prescription drugs, eye exams, hearing aids, or routine physical exams.  Private fee-for-service MPS plans charge medium-to-high out-of-pocket costs, allows for a choice of physician, and provides extra coverage for foreign travel or extended stays in a hospital. Additional information and resources pertaining to the Medicare program can be found at http://www.medicare.gov and http://www.cms.gov/Medicare/Medicare.html

Asthma Case Definitions: Descriptive data elements that are typically included in Medicare data include recipient age, sex, race, county/zip code of residence. There are several asthma-related case definitions that have been developed for hospitalization data that also may be applicable to Medicare data:

  1. The Council of State and Territorial Epidemiologists (CSTE) (http://c.ymcdn.com/sites/www.cste.org/resource/resmgr/PS/1998-EHCD-1.pdf) has jointly developed a standardized case classification for asthma in hospital discharge data. This classification also may be applicable to Medicare data for claims for inpatient and outpatient (including ED) visits.
    1. Confirmed Case: There is no confirmed case classification for Medicare data.
    2. Probable Case: Part A or Part B inpatient/outpatient Medicare claims listing asthma (any ICD-9-CM Code 493 and ICD-10-CM Code J45) as the primary discharge diagnosis.
    3. Possible Case: Part A or Part B inpatient/outpatient Medicare claims listing asthma (any ICD-9-CM Code 493 and ICD-10-CM Code J45) as a secondary discharge diagnosis or a primary discharge diagnosis of acute bronchitis and bronchiolitis (ICD-9-CM Code 466; ICD-10-CM Code J20-J21) in children younger than 12 years, or chronic bronchitis (ICD-9-CM Codes: 491.20 and 491.21; ICD-10-CM Codes J40-J41) in children younger than 12 years.
  2. The National Committee for Quality Assurance (NCQA) also has developed standards to define persons with “persistent asthma.” This may be used within the Medicare population that receives Part B and supplementary medication coverage.1 Persistent asthma can be defined as:
    1. Four or more prescription medications used in the treatment of asthma in one year, OR
    2. One or more inpatient hospital visits with a primary diagnosis of asthma in a year, OR
    3. One or more ED visits with a primary diagnosis of asthma in a year, OR
    4. Four or more outpatient visits with asthma listed anywhere as one of the diagnoses and two (2) or more claims for a prescription drug used in the treatment of asthma within 1 year.

    The NCQA has developed a list of Common Procedural Terminology codes and Uniform Billing-92 revenue codes that are designed to more sensitively capture inpatient and outpatient encounters in Medicaid data.  This is because Medicaid data are based on paid claims and not all encounters.  These lists may also be useful for identifying asthma cases in Medicare data. Persons can obtain more information about this at http://www.ncqa.org.

  3. For the assessment of appropriate medication use, HEDIS has set a standard for appropriate prescription drug treatment that requires a person with asthma to have received any of the drugs identified by the NCQA as a primary therapy for the long-term treatment and control of asthma.1 The NCQA puts out annual lists of National Drug Codes (NDCs, http://www.ncqa.org/hedis-quality-measurement/hedis-measures/hedis-2016/hedis-2016-ndc-license/hedis-2016-final-ndc-lists, scroll down and click accept to enter the site) that indicate the drugs classified as long-term asthma controllers as well as NDCs of current long-term and short-term asthma medications. These drug codes provide a standard that may be used when classifying an asthma “case” based on medication use, or when calculating the proportion of persons with asthma who are on long-term control medications. However, any asthma case definition will need to identify the proportion of all asthma cases that were defined through diagnosis, medication prescription, or both.

The NCQA has developed a list of Common Procedural Terminology codes and Uniform Billing-92 (UB 92) revenue codes that are designed to more sensitively capture inpatient and ambulatory prepaid managed care encounter claims.  This is because Medicaid data are based on paid claims (fee-for-service claims, per-service managed care claims, primary care case management fees, etc.) and rather than encounters.  Persons can obtain more information about this at http://www.ncqa.org.

Source of Data (How to Access the Data)

The Medicare Provider Analysis and Review file contains records for Medicare beneficiaries who use hospital inpatient services.  The records are stripped of most data elements that will permit identification of beneficiaries.  The six-position Medicare billing number identifies the hospital.  These data can be obtained from the CMS web site at https://www.cms.gov/Research-Statistics-Data-and-Systems/Files-for-Order/IdentifiableDataFiles/index.html.  In addition, data pertaining to Part B claims can be obtained from https://www.cms.gov/Research-Statistics-Data-and-Systems/Downloadable-Public-Use-Files/Part-B-National-Summary-Data-File/Overview.html.

A detailed description of available Medicare data can be found at https://www.cms.gov/Research-Statistics-Data-and-Systems/Statistics-Trends-and-Reports/Medicare-Geographic-Variation/GV_PUF.html, or the CMS statistics and data information page at (https://www.cms.gov/Research-Statistics-Data-and-Systems/Research-Statistics-Data-and-Systems.html).  A list of Part A and Part B Intermediary insurance providers by state can be found at https://www.cms.gov/medicare/medicare-contracting/medicare-administrative-contractors/medicareadministrativecontractors.html. As Medicare data are national data, a subset of data limited to your state will need to be requested.  Additional questions related to Medicare claims data can be directed to the Centers for Medicare & Medicaid Services, 7500 Security Boulevard, Baltimore, MD  21244-1850; 877-267-2323.

Preparation of Data and Standard Methods for Rate Calculations

In order to enhance the comparability of asthma-related morbidity estimates between states, the following steps outline a standard process for preparing Medicare data for asthma surveillance purposes.

  • It is important to be sure that the criteria used to determine inclusion in the numerator is as similar as possible as the criteria used for assignment in the denominator. If Medicare enrollees with additional insurance (including Medicaid) are included/excluded from the numerator of rates, the same exclusion/inclusion criteria will need to be applied to the denominator of rates.
  • Remove claims for persons younger than age 65 with disabilities from the data set. However, even if this is done, Medicare data may not be representative of the broader state population aged 65 and older (see the “Anticipated Questions and Answers” section below).
  • The date of any asthma encounter is identified by the service data of the claim, not the date on which the claim was paid.
  • The denominator for any rate is the number of persons continuously enrolled in the state Medicare program during each year in the time period covered by the rate. We suggest that “continuous enrollment” be defined as enrollment of a full year with no more than a one month gap in enrollment (i.e., 11 or more months of enrollment per year). If calculating the HEDIS asthma measure, a stricter criteria for continuous enrollment is needed (i.e., 11 or months of measurement year and in the measurement year.

Following these steps will increase the compatibility of Medicare-derived asthma morbidity estimates among continuously enrolled Medicare recipients over 65 between states. Denominator data sets persons enrolled in Medicare by State (and master files to annotate Part-B buy-in beneficiaries) can be obtained from https://www.cms.gov/Research-Statistics-Data-and-Systems/Files-for-Order/LimitedDataSets/DenominatorLDS.html.  A detailed description of these denominator data can be found at https://www.cms.gov/research-statistics-data-and-systems/files-for-order/limiteddatasets/denominatorlds.html.

Analysis Standards

The NATC Surveillance module includes a discussion of standard demographic breakdowns that should be used when analyzing asthma surveillance data.  The applicability of these breakdowns to Medicare data are summarized below.

  • Age Categories: After removing persons less than 65 from the data set, rates can be calculated by age for 5-year age intervals for groups aged 65 and older (e.g., 65-69, 70-74 etc.).  Note that there is the potential for misclassification with other respiratory diseases.
  • Sex Categories: Rates can be calculated separately for “male” and “female” categories.
  • Race Categories: “White,” “Asian,” “Hispanic,” “and “Other” are typically mutually exclusive categories in Medicare data.  If possible, rates can be calculated separately for each of these categories.
  • Time Trends: Rates may be calculated by year.  Rates also can be calculated by month, day of week, and time of day.  However, when calculating trends in Medicare data, care should be taken to determine that no changes in individual eligibility criteria have occurred during the time period in question.  Such changes could influence the number of asthma cases “captured” and the type of population that is represented by the Medicare data.
  • Geographic Categories: Rates can be calculated at the state, county, city, and zip code levels if the cell sizes permit.  Age standardized rates are used to compare geographic units.

 NOTE:  Small sample size can result in release of inferred release of confidential or sensitive information and can also affect reliability of rates.  Please consider collapsing years or demographic groups, presenting confidence intervals, and/or suppressing rates and counts if sample size of the numerator or denominator is inadequate.  In some cases, an aggregation of categories of data may be necessary to achieve the relative standard error of 30% that has been suggested to produce reliable rates (https://www.cdc.gov/nchs/data/statnt/statnt24.pdf).  For example, if the event count is < 20 in any particular year, we recommend combining years to achieve numerators > 20, and/or estimating trends based on 3-year rolling averages (e.g., calculate a single rate for 1978-1980, then 1979-1981, etc.).  To protect patient confidentiality, CMS requires that “data must not be beneficiary-specific and must be aggregated to a level where no data cells have 10 or fewer beneficiaries.”2

Anticipated Questions and Answers

Q   Do Medicare asthma claims data accurately reflect the overall burden of asthma among persons > 65?
A    Nationally, 96% of persons aged 65 and older are enrolled in Medicare.3   However, this varies considerably by age and race.  Among whites, about 90% of the population aged 65-69 are enrolled in Medicare.  Only 79% of Black s aged 65 to 70 are enrolled.  Both of these percentages increase with age.

Q   Will Medicare Part A data sensitively capture all asthma hospitalizations occurring to enrolled members > age 65?
A    Ideally, Medicare Part A data would capture all hospitalizations, but this is not always the case.  One study found that 9.1% of all hip fracture hospitalizations that occurred to Medicare patients nationally could be found in Part B data but not Part A data.  This percentage ranged from 0% to 32.8% depending on the state.3

Q   What surveillance questions can be answered with only Medicare Part A data?
A    Probable and possible hospitalizations because of asthma can be identified for most paid Medicare claims (Part A data).  Analyses will need to be limited to those with Part B or supplemental insurance coverage when estimating asthma-related office / outpatient visits, specialist visits, ED visits, or medication dispensation

State and Federal Contacts and Resources for Medicaid Data

Centers for Disease Control and Prevention (CDC)
National Center for Environmental Health
Asthma and Community Health Branch
General Number: (770) 488-3700

Current contact information for all state asthma contacts: http://www.cdc.gov/asthma/contacts/default.htm

Centers for Medicare and Medicaid Services
Research Data Assistance Center (RESDAC)

www.resdac.umn.edu
General Number: 888-973-7322
resdac@tc.umn.edu

CMS Public Use Files https://www.cms.gov/Research-Statistics-Data-and-Systems/Downloadable-Public-Use-files/SynPUFs/index.html

References

  1. National Committee for Quality Assurance. Use of appropriate medications for people with asthma. HEDIS 2014 technical specifications. Vol 2.
  2. Medicare Fee-For Service Provider Utilization & Payment Data Physician and Other Supplier Public Use File: A Methodological Overview,” Prepared by:
    The Centers for Medicare and Medicaid Services, Office of Information Products and Data Analytics, April 7, 2014. Available at https://www.cms.gov/research-statistics-data-and-systems/statistics-trends-and-reports/medicare-provider-charge-data/downloads/medicare-physician-and-other-supplier-puf-methodology.pdf
  3. Fisher ES, Baron JA, Malenka DJ, Barrett J, Bubolz TA. Overcoming potential pitfalls in the use of Medicare data for epidemiologic research.  Am J Public Health. 1990;80:1487-1490.