The Medical Expenditure Panel Survey (MEPS) is designed to provide nationally representative estimates of health care use, expenditures, sources of payment, and insurance coverage for the U.S. civilian noninstitutionalized population. MEPS is cosponsored by the Agency for Healthcare Research and Quality (AHRQ) and the National Center for Health Statistics (NCHS). MEPS comprises three component surveys: the Household Component (HC), the Medical Provider Component (MPC), and the Insurance Component (IC). The HC is the core survey of MEPS.
The MEPS HC, a nationally representative survey of the U.S. civilian noninstitutionalized population, collects medical expenditure data at both the person and household levels. The HC collects detailed data on demographic characteristics, health conditions, health status, use of medical care services, charges and payments, access to care, satisfaction with care, health insurance coverage, income, and employment.
The HC uses an overlapping panel design in which data are collected through a preliminary contact followed by a series of five rounds of interviews over a 2½-year period. Using computer-assisted personal interviewing (CAPI) technology, data on medical expenditures and use for 2 calendar years are collected from each household. This series of data collection rounds is launched each year on a new sample of households to provide overlapping panels of survey data and, when combined with other ongoing panels, will provide continuous and current estimates of health care expenditures.
The sample of households selected for the MEPS HC is drawn from respondents to the NHIS, conducted by NCHS. The NHIS provides a nationally representative sample of the U.S. civilian noninstitutionalized population, with oversampling of Hispanics and Blacks.
National health care estimates from MEPS for the 2006 NHDR were derived from the 2003 MEPS HC survey, including the Self-Administered Questionnaire (SAQ), the Child Health and Preventive Care section, and the Diabetes Care Survey (DCS).
The SAQ was a supplement to the MEPS HC and was completed in late 2003 (Panel 7 Round 4) and early 2004 (Panel 8 Round 2). It included health care quality measures taken from the health plan version of CAHPS®, an AHRQ-sponsored family of survey instruments designed to measure quality of care from the consumer's perspective; general health questions; attitudes about health questions; and health status questions as measured by the SF-12 and the EuroQol 5D.
The Child Health and Preventive Care section was part of the regular MEPS HC interview that, like the SAQ, took place in late 2003 and early 2004. It included health care quality measures taken from the health plan version of CAHPS®; the Children with Special Health Care Needs (CSHCN) Screener questions; children's general health status as measured by several questions from the General Health Subscale of the Child Health Questionnaire; Columbia Impairment Scale questions about possible child behavioral problems; and child preventive care questions. Researchers should note that the CAHPS® and CSHCN questions changed from a self-administered parent questionnaire in 2000 to an interviewer-administered questionnaire starting in 2001.
A third supplement to the MEPS HC, the Diabetes Care Survey (DCS), was a self-administered questionnaire given to persons identified with diabetes. It asked about the care they received in the treatment of their diabetes.
Estimates derived from MEPS are presented at both an aggregate level and for select subpopulations. Characteristics used to define subpopulations included age, gender, race, ethnicity, poverty status, education, insurance coverage, proximity to metropolitan areas, preferred language at home, whether or not the person was born in the United States, and children with special health care needs. A brief description of how each of these population characteristics was defined is provided below.
Age—With the exception of analytic variables associated with round-specific questions noted below, age was defined as a person's age on December 31 of the data year.
For measures using analytic variables associated with round-specific questions (e.g. questions from the SAQ, the Child Health and Preventive Care supplement, as well as access-to-care measures), corresponding round-specific age variables were used to determine age.
Gender—Male and female.
Race—Variable RACEX classifies persons into five single race categories (1-5) and one for multiple races. For all tables, RACEX was used and re-ordered as: 1) White, 2) Black, 3) Asian, 4) Native Hawaiian or Other Pacific Islander, 5) American Indian or Alaska Native, 6) multiple races. The race groups include persons of Hispanic and non-Hispanic origin. From 2002, MEPS has transitioned to the most recent Office of Management and Budget standards issued in 1997 for collecting racial and ethnic data which allows respondents to identify more than one racial/ethnic group. Therefore, estimates for racial categories using 2002 and later years' data are not directly comparable with estimates from previous years' data.
Ethnicity—Ethnicity was determined to be either Hispanic or non-Hispanic. Persons of Hispanic origin may be of any race. Estimates were derived for both Hispanic and non-Hispanic subpopulations. In addition, race was crossed with ethnicity and estimates were reported for persons classified as "non-Hispanic, White," "non-Hispanic, Black," and "non-Hispanic, other." For 2002 and later years, "non-Hispanic, White" and "non-Hispanic, Black" categories excluded multiple race individuals and estimates are not directly comparable to data from previous years.
Poverty status—MEPS includes a five-level categorical variable for family income as a percentage of poverty. For construction of this variable, definitions of income, family, and poverty are taken from the poverty statistics developed by the Current Population Survey (CPS). For the purposes of analysis and reporting in the NHDR the near poor and low income categories were combined. This resulted in a four-level categorical variable of poverty status: (1) negative or poor refers to household incomes below the Federal poverty line; (2) near poor/low income, over the poverty line to just below 200 percent of the poverty line; (3) middle income, 200 percent to just below 400 percent of the poverty line; and (4) high income,over 400 percent of the poverty line. Beginning with the 2002 file, there were substantial revisions made to the skip patterns in the Income section. These changes have increased response rates, resulting in a small impact on income estimates for persons under 65, with a somewhat larger impact on person 65 and older.
Education—In MEPS, a person's educational attainment is indicated as the number of completed years of education. For the NHDR, this continuous measure was grouped into three categories: (1) less than high school refers to persons with less than 12 completed years of education;, (2) high school graduate, persons with exactly 12 completed years of education; and (3) at least some college, persons with greater than 12 completed years of education. This variable was constructed only for persons 18 years of age or older and any measure presented for the education subpopulations includes only persons in this age cohort.
Insurance coverage—The insurance coverage variable was constructed in a hierarchical manner and in relation to a person's age. For persons less than 65 years of age, those who were uninsured for the entire year were classified as uninsured; those who had private coverage at any time during the year (including CHAMPUS/VA) were classified as having had private insurance; and persons who had only public coverage (i.e., no private) at any time during the year were classified as public only. Persons over 65 years of age were categorized as having Medicare only, Medicare and private, or Medicare and other public assistance. A small number of persons (n=56) were identified as having no Medicare and no other private or public assistance. These people were not included in insurance related analyses.
Residence location—A variable was constructed by combining the 12 levels of the 2003 Urban Influence Codes to 5 levels: (1) metropolitan-large is defined as in a metro area with at least 1 million residents; (2) metropolitan-small is defined as in a metro area with fewer than 1 million residents; (3) micropolitan is defined as adjacent to a large metro area, small metro area, or not adjacent to a metro area; (4) noncore-adjacent is defined as noncore, adjacent to a large or a small metro area, or adjacent to a micro area and contains a town of 2,500-9,999 residents; and (5) noncore-not adjacent is defined as noncore, adjacent to a micro area and does not contain a town of at least 2,500 residents, or not adjacent to a metro/micro area. For measures with subpopulations, such as diabetes measures, levels 4 and 5 were combined as "noncore" to increase the possibility that reliable estimates could be obtained.
Children with special health care needs—The variable CSHCN42 identifies children with special health care needs based on the CSHCN Screener instrument developed through a national collaborative process as part of the Child and Adolescent Health Measurement Initiative (CAHMI) under the coordination of the Foundation for Accountability. Missing values (CSHCN42=3) were excluded.
Preferred language at home—For each individual family member, the Access to Care section ascertained what language the person prefers to speak at home (LANGHM42). Persons who prefer to speak Spanish or a language other than English (LANGHM42 = 2 or 3) were combined as "Other" group.
Born in U.S.—Family members were also asked whether they were born in the United States (USBORN42).
Standard errors of the estimates were provided to permit an assessment of the sampling variability. All estimates and standard errors were derived using SUDAAN statistical software which accounts for the complex survey design of MEPS.
All estimated proportions and ratios are weighted to reflect the experiences of the U.S. civilian noninstitutionalized population at the aggregate and subpopulation levels. For example, the SAQ person-level weight (SAQWT03F) was used for measures from the SAQ, and the diabetes person-level weight (DIABW03F) was used for measures from the diabetes care survey. For other person-level measures, including those from the Child Health and Preventive Care section, the overall person-level weight (PERWT03F) was used. For the Child Health and Preventive Care measures, child population included children (under 18) with a positive person-level weight who had been asked these questions during the later half of 2003 (where PERWT02F>0 and PSTATS42 was not equal to 31 (deceased)).
To obtain estimates at the family-level, it is necessary to prepare a family-level file containing one record per family and containing family-level summary characteristics, to use the family reference person's values for demographic characteristics, and to use the MEPS family-level weight (FAMWT03F).
Tables containing estimates from MEPS are included in Appendix D: Data Tables. Consistent with the established criteria for data reporting in the NHDR, MEPS estimates are suppressed when they are based on sample sizes of less than 100, or when their relative standard errors are 30% or more. In the tables, these data are replaced with symbols to indicate that data are statistically unreliable. Records in which analytic variables or demographic variables have missing values were excluded for analysis.