Online Performance Appendix: Performance Detail, Medical Expenditure Panel Survey (MEPS)

Budget Estimates for Appropriations Committees, Fiscal Year 2010

This appendix provides more detailed performance information for all HHS measures related to the Agency for Healthcare Research and Quality's budget.

The Medical Expenditure Panel Survey (MEPS), first funded in 1995, is the only national source for annual data on how Americans use and pay for medical care. It supports all of AHRQ's research-related strategic goal areas. The survey collects detailed information from families on access, use, expense, insurance coverage, and quality. Data are disseminated to the public through printed and Web-based tabulations, micro data files, and research reports/journal articles.

The data from the MEPS have become a linchpin for the public and private economic models projecting health care expenditures and utilization. This level of detail enables public and private sector economic models to develop national and regional estimates of the impact of changes in financing, coverage, and reimbursement policy, as well as estimates of who benefits and who bears the cost of a change in policy. No other surveys provide the foundation for estimating the impact of changes on different economic groups or special populations of interest, such as the poor, elderly, veterans, the uninsured, or racial/ethnic groups. Government and non-governmental entities rely upon these data to evaluate health reform policies, the effect of tax code changes on health expenditures and tax revenue, and proposed changes in government health programs such as Medicare. In the private sector (e.g., RAND, Heritage Foundation, Lewin-VHI, and the Urban Institute), these data are used by many private businesses, foundations, and academic institutions to develop economic projections. These data represent a major resource for the health services research community at large. Since 2000, data on premium costs from the MEPS Insurance Component have been used by the Bureau of Economic Analysis to produce estimates of the gross domestic product (GDP) for the Nation. Because of the need for timely data, performance goals for MEPS have focused on providing data in a timely manner.

The MEPS program has met or exceeded all of its data timeliness goals. These performance goals require the release of the MEPS Insurance Component tables within 7 months of data collection; the release of MEPS Use and Demographic Files within 12 months of data collection; and the release of MEPS Full Year Expenditure data within 12 months of data collection. In addition, the program has expanded the depth and breadth of data products available to serve a wide range of users. To date, over 200 statistical briefs have been published. The MEPS data table series has expanded to include eight topic areas on the household component and nine topic areas on the Insurance Component. In addition, specific large State and metro area expenditure and coverage estimates have been produced, further increasing the utility of MEPS within the existing program costs. In 2008, further advances in the timeliness of the MEPS Insurance Component data have been initiated through a change in the questionnaire to obtain employer-sponsored coverage information that is in force at the time of data collection, rather than for a prior year retrospective reference period. Further advances in data accessibility have also been achieved by expanding the number of Data Centers across the Nation to permit access to MEPS restricted data through a collaboration with the Bureau of the Census to utilize nine additional Research Data Centers for approved projects. Since its inception in 1996, MEPS has been used in several hundred scientific publications, and many more unpublished reports to inform health policy decisions and practice.

  • The MEPS data have been used extensively by the Congressional Budget Office, Congressional Research Service, Department of Treasury, Joint Taxation Committee, and Department of Labor to inform Congressional inquires related to health care expenditures, insurance coverage, and sources of payment, and to analyze potential tax and other implications of Federal health insurance policies.
  • MEPS data on health care quality, access, and health insurance coverage have been used extensively in the Department's two annual reports to Congress, the National Healthcare Disparities Report, and the National Healthcare Quality Report.
  • The MEPS was awarded the American Association for Public Opinion Research's 2008 Policy Impact Award in recognition of the extraordinary, long-term group effort in contributing timely data and research that has informed U.S. health care policy decisions.
  • The MEPS has been used in Congressional testimony on the impact of health insurance coverage rate increases on small businesses.
  • The MEPS data have informed studies of the value of health insurance in private markets and the effect of consumer payment on health care.
  • The MEPS data have been extensively used to inform Congressional inquiries tied to SCHIP reauthorization, with particular emphasis on the change in take-up rates among Medicaid- eligible children over the implementation period of SCHIP and the percent of all uninsured children who are eligible for Medicaid or SCHIP.
  • The MEPS-IC has been used by a number of States in evaluating their own private insurance issues, including eligibility and enrollment by the State of Connecticut and by the Maryland Health Care Commission, and for community rating by the State of New York. As part of the Robert Wood Johnson Foundation's State Coverage Initiative, MEPS data was cited in 69 reports, representing 27 States.
  • The MEPS data have been used extensively by the Government Accountability Office to determine trends in employee compensation, with a major focus on the percentage of employees at establishments that offer health insurance, the percentage of eligible employees who enroll in the health insurance plans, the average annual premium for employer-provided health insurance for single workers, and the employees' share of these premiums.
  • MEPS data have been used in HHS Reports to Congress on expenditures by sources of payment for individuals afflicted by conditions that include acute respiratory distress syndrome, arthritis, cancer, chronic obstructive pulmonary disease, depression, diabetes, and heart disease.
  • MEPS data are used to develop estimates provided in the Consumers Checkbook Guide to Health Plans, of expected out-of-pocket costs (premiums, deductibles, and copays) for Federal employees and retirees for their health care. The Checkbook is an annual publication that provides comparative information on the health insurance choices offered to Federal workers and retirees.
  • The MEPS has been used to estimate the impact of the Medicare Modernization Act (MMA) by the Employee Benefit Research Institute (the effect of the MMA on availability of retiree coverage), by the Iowa Rural Policy Institute (effect of the MMA on rural elderly), and by researchers to examine levels of spending and copayments.
  • MEPS data have been used by the CDC and others to evaluate the cost of common conditions, including arthritis, injuries, diabetes, obesity, and cancer.

Before AHRQ reorganized research portfolios, MEPS was part of the Data Collection and Dissemination portfolio. This portfolio underwent a program assessment in 2002, and was found to be moderately effective. The review cited MEPS as a strong attribute of the program. As a result of the program assessment, the program continues to take actions to reduce the number of months that MEPS data is made available after the date of completion of the survey, increase the number of MEPS data users, and increase the number of topical area tables included in the MEPS Tables Compendia.

The FY 2010 President's Budget Request for the MEPS totals $55,300,000 in Public Health Service (PHS) evaluation funds, maintaining the FY 2009 Omnibus level. The funding allocation in FY 2010 for the three core MEPS Component Surveys follows: MEPS Household Component ($33,300,000); MEPS Medical Provider Component ($12,000,000); and the MEPS Insurance Component ($10,000,000).

The FY 2010 funding for MEPS will be used to support the sample size and content of the MEPS Household and Medical Provider Surveys necessary to satisfy the congressional mandate to submit an annual report on national trends in health care quality and to prepare an annual report on health care disparities. The MEPS Household Component sample size is specified at 14,000 households in 2010 with full calendar year information. These sample size specifications for the MEPS permit detailed analyses of the quality of care received by special populations meeting precision specifications for survey estimates. This design, in concert with the survey enhancements initiated in prior years, significantly enhances AHRQ's capacity to report on the quality of care Americans receive at the national and regional level, in terms of clinical quality, patient satisfaction, access, and health status both in managed care and fee-for-service settings.

The MEPS Household Component

These funds will also permit the continuation of an oversample in MEPS of Asian and Pacific Islanders and an over-sample of African Americans. These enhancements, in concert with the existing MEPS capacity to examine differences in the cost, quality and access to care for minorities, ethnic groups and low income individuals, will provide critical data for the National Healthcare Quality Report and the National Healthcare Disparities Report.

The MEPS Insurance Component

Funds will also be allocated to the MEPS Insurance Component to maintain improvements in the availability of data to the States. In FY 2010, data on employer sponsored health insurance will be collected to support separate estimates for all 50 States and these funds would be used to enhance the tabulations we provide to the States to support their analysis of private, employer sponsored health insurance.

The Medical Provider Component

FY 2010 funds will also support the MEPS Medical Provider Component, a survey of medical providers, facilities and pharmacies that collects detailed data on the expenditures and sources of payment for the medical services provided to individuals sampled for the MEPS. Such data are essential to improve the accuracy of the national medical expenditure estimates derived from the MEPS and to correct for the item non-response on expenditures by household sample participants.

Recent enhancements to the estimation capabilities of the MEPS Household Component have also been realized and permit the generation of health care utilization, expenditure and health insurance coverage estimates for some large metropolitan areas and for the 10 largest States. This has resulted in visible improvements in the analytic capacity of the survey without any additional increments to the sample size.

MEPS—Marginal Cost

The Baseline MEPS sample consists of approximately 14,000 households and 32,000 individuals, and includes over-sampling of African Americans, Hispanics, Asians and low income households. With respect to desired levels of precision for survey estimates, a relative standard error (RSE) specification of less than or equal to 10 percent is recommended for survey estimates that characterize policy relevant population subgroups which include racial and ethnic minorities (RSE (X) = standard error (X) divided by the estimate X.). This precision target is not currently being met for estimates of the health care utilization and expenditure patterns for American Indians/Alaskan Natives, subgroups of individuals of multiple races (e.g. race classifications of both African-American and other race), specific Hispanic subgroups (e.g., Puerto Rican, Cuban, Dominican) and Asian population subgroups (e.g., Chinese, Vietnamese, Asian Indian). The FY 2010 cost estimate for MEPS—at an average cost of $6,971 per household for the household and medical provider components of the MEPS survey—would allow for the following sample yields for these racial and ethnic minority population subgroups in MEPS.

MEPS Oversampling

SubgroupBaseline—FY 2010 President's Budget Request
IndividualsRSE (for mean expenditures)
Hispanic Subgroups  
Puerto Rican70011.5%
American Indians/Alaskan Natives40013.2%
Multiple Races5759.0%

The baseline cost components related to the household and medical provider component of MEPS for a full panel of 7,000 households over 3 years are provided below:

Cost ComponentsBaseline
HouseholdsFull MEPS consists of 14,000 households
(1) Sample Selection$0.6 M
(2) Management$1.1M
(3) Hire/Train Household/Medical Provider Survey Staff$3.4M
(4.a) Conduct Household Interviews$20.7M
(4.b)Data Collection-Medical Providers$10.9M
(5) Data Processing/Production of Analytical Files$12.1M
Total Cost$48.8M
Cost per Household$6,971

Costs associated with (1) the sample frame preparation and sample selections for the MEPS Household and Medical Provider Surveys and (2) the management tasks are fixed, while costs associated with the remaining data collection and data processing components are variable.

In 2007, a marginal cost analysis was completed to determine the marginal cost of increasing the degree of oversampling in the MEPS sample among certain minority sub-groups. This oversampling would allow estimates for these subgroups to be more precise, allowing the implications of program and policies to be more accurately estimated for these groups using MEPS data. As indicated, many estimates for these subgroups have relative standard errors that are higher than the recommended maximum threshold of 10 percent. The marginal cost to reach the recommended RSE of 10 percent for these minority subgroups in 2010 and 2011 is $4,300 per additional minority household surveyed, which is lower than the current average cost per household of $6,971.

The table below indicates the percent reduction in relative standard errors in survey estimates that could be achieved by a targeted MEPS sample augmentation of 1,000 additional households.

SubgroupRSE (for mean expenditures) with
MEPS Sample Augmentation
Reduction in RSE (for mean expenditures) with
MEPS Sample Augmentation
Hispanic Subgroups  
Puerto Rican9.6%15%
American Indians/Alaskan Natives10.0%24%
Multiple Races7.6%16%

Long-Term Objective 1: Achieve wider access to effective health care services and reduce health care costs.

1.3.16: Insurance Component tables will be available within months of collection
2010TBDOct 31, 2010
2009Set BaselineOct 31, 2009
(Target Met)
(Target Met)
(Historical Actual)
(Historical Actual)
1.3.17: MEPS Use and Demographic Files will be available months after final data collection
201011Oct 31, 2010
200911Oct 31, 2009
(Target Met)
(Target Met)
(Historical Actual)
1.3.18: Number of months after the date of completion of the MEPS data will be available
201010.8Oct 31, 2010
200911Oct 31, 2009
(Target Met)
(Target Met)
200612 months12 months
(Target Met)
200512 months12 months
(Target Met)
1.3.19: Increase the number of topical areas tables included in the MEPS Tables Compendia (TC)
2010Add additional variables to MEPS NetOct 31, 2010
2009Update State-level tablesOct 31, 2009
2008Add Prescribed Drug TablesPrescribed Drug Tables Added
(Target Met)
2007Add Insurance TablesInsurance Tables Added
(Target Met)
2006Add State TablesState Tables Added
(Target Met)
2005Add Access TablesAccess Tables added
(Target Met)
1.3.20: Increase the number of MEPS data users
Baseline FY 2005: 10 Data Center Projects (DCP), 15,900 TC, 13,101 Household Component/Insurance Component (HC/IC)
2010Exceed baseline standardOct 31, 2010
2009Exceed baseline standardOct 31, 2009
2008Exceed baseline standard41 DCP
(Target Met)
2007Exceed baseline standard

23 DCP

19,989 TCP

14,809 HC/IC
(Target Met)

2006Exceed Baseline standard

14 DCP

16,200 TCP

11,600 HC/IC
(Target Met)

2005Maintain Baseline standard

10 DCP

15,900 TC

13,101 HC/IC

1.3.21: The number of months required to produce MEPS data files (i.e., point-in-time, utilization and expenditure files) for public dissemination following data collection
201010.8 monthsOct 31, 2010
200911 monthsOct 31, 2009
200811 months11 months
(Target Met)
200711 months11 months
(Target Met)
200612 months12 months
(Target Met)
1.3.49: The average number of field staff hours required to collect data per respondent household for the MEPS (at level funding level)
(Annual Efficiency Measure)
201012.8 hoursOct 31, 2010
200913.0 hoursOct 31, 2009
200813.5 hours13.5 hours
2007Baseline14.2 hours


MeasureData SourceData Validation
1.3.16MEPS Web siteData published on Web Site
MEPS Web siteMonthly meetings with contractor, careful monitoring of field progress and instrument design, quality control procedures including benchmarking with other national data sources.
1.3.19MEPS Web siteData published on Web site.
1.3.20MEPS data: List of ongoing projectsPublications.
1.3.49 The number of field staff hours required to collect data per respondent household for the MEPS is logged by field staff in an automated system. Data quality and validation is monitored in several ways: (1) validation interviews are conducted for a sample of respondents, in which questions concerning the interview process are asked; (2) response rates are monitored to ensure that they stay high; and (3) the duration of interviews are tracked to ensure that interviewers are following proper protocol and not skipping questions during the interview.

Return to Contents
Proceed to Next Section

Current as of May 2009
Internet Citation: Online Performance Appendix: Performance Detail, Medical Expenditure Panel Survey (MEPS): Budget Estimates for Appropriations Committees, Fiscal Year 2010. May 2009. Agency for Healthcare Research and Quality, Rockville, MD.