Medical Expenditure Panel Survey

Studies highlight the value of the Medical Expenditure Panel Survey to inform trends in care costs, coverage, use, and access

The Medical Expenditure Panel Survey (MEPS) is an ongoing national survey of medical care costs, coverage, use, and access, which is sponsored by the Agency for Healthcare Research and Quality. The MEPS consists of a family of three interrelated surveys: the Household Component, the Medical Provider Component, and the Insurance Component. The MEPS provides annual national estimates of health care use, medical expenditures, sources of payment, and insurance coverage for the U.S. civilian noninstitutionalized population. It also provides estimates of health status, demographic characteristics, employment, and access to health care. Estimates can be made available for individuals, families, and various subgroups.

MEPS data can also be used to study factors that determine the use of medical care services and expenditures; changes in the provision of health care in relation to social and demographic factors such as employment or income; the health status and satisfaction with healthcare of individuals and families; and the health needs of specific population groups such as the elderly and children. A special May 2006 issue of Medical Care features nine original articles that demonstrate the utility of MEPS to inform trends in medical care costs, coverage, use, and access. Following are brief summaries of the introduction and nine articles that appear in the issue:

Cohen, S.B., and Buchmueller, T. "Trends in medical care costs, coverage, use, and access: Research findings from the Medical Expenditure Panel Survey," pp. 1-3.

This introduction summarizes the nine articles in the journal issue. The first set of articles illustrates the capacity of the MEPS to address important policy issues related to the availability and take-up of employment-related health insurance coverage. The next set of articles demonstrates the unique features of MEPS for studying the behavior and characteristics of individuals who experience high medical costs. The final article in this special issue evaluates the reliability of self-reports of recovery from disability in the MEPS Household Component. All of the articles illustrate recent research efforts using MEPS data to aid development, implementation, and evaluation of policies and practices addressing health care and health behaviors. They also represent AHRQ's emphasis on research initiatives that yield findings that can be translated into practice to improve the quality, safety, efficiency, and effectiveness of health care.

Cooper, P.F., Simon, K.I., and Vistnes, J. "A closer look at the managed care backlash," pp. 4-11.

Much has been written about the consumer backlash against managed care, but limited empirical evidence is available. These investigators analyzed data from the MEPS Insurance Component (MEPS-IC) to understand trends in enrollment in health maintenance organizations (HMOs) between 1997 and 2003. Consistent with anecdotal evidence and previous studies, the researchers documented a decline in HMO enrollment since 1996. HMO enrollment rates fell from about 32 percent to 26 percent between 1997 and 2003, with most of the decline occurring after 2001. This overall trend, however, masked interesting differences by firm size. MEPS-IC data revealed a decline in the HMO enrollment rate for large employers starting in 1998, which was driven by employees shifting to preferred provider organization coverage. However, this was offset by an increase in the HMO enrollment rate by employees of small firms. Nevertheless, when workers were given a choice between an HMO and other plan types, they increasingly opted for the non-HMO plan.

Bernard, D.M., and Selden, T.M. "Workers who decline employment-related health insurance," pp. 12-18.

Families of workers who decline employment-related health care coverage represent a substantial share of the uninsured and publicly insured population in the United States. The authors of the paper used MEPS data from 2001 to 2002 to focus on these families. They found that a majority of children from low-income families whose parents declined insurance obtained coverage through public programs. However, nearly all adults who declined employer-sponsored coverage were uninsured. The differences in availability of public insurance had important implications for access to care. Decliners who took up public insurance were as likely as individuals with employer-sponsored insurance and significantly more likely than uninsured decliners to report a usual source of care and at least one doctor visit during the year. Families turning down employer-sponsored coverage were more likely to face higher medical expenditure burdens as a percentage of income and to have financial barriers to care. They also tended to rely heavily on the safety net of public coverage and uncompensated care.

Selden, T.M., and Hudson, J.L. "Access to care and utilization among children: Estimating the effects of public and private coverage," pp. 19-26.

Despite concerns that conventional estimates overstate the impact of insurance coverage on care access and use, this study of children suggests that the reverse may be true. The magnitude of the impact of coverage on children's care access and use underscores the importance of reducing uninsurance among children, note the researchers. They pooled MEPS data from 1996 to 2002 to estimate the impact of insurance coverage on children's access to and use of care. They found, as previous studies have found, that public and private coverage were both associated with large increases in care access and use. The large differences between public and private coverage were reduced (and often reversed) when they controlled for other characteristics of children and their families. The effect of coverage on care access and use was substantially greater using instrumental variables estimates compared to conventional estimates across a wide range of access and use measures.

Banthin, J.S., and Miller, G.E. "Trends in prescription drug expenditures by Medicaid enrollees," pp. 27-35.

The double-digit growth in Federal spending on prescription medications has consumed an increasingly large portion of Medicaid budgets. This study used MEPS data to identify trends between 1996 and 1997 and 2001 and 2002 in Medicaid drug use and expenditures. The researchers examined specific therapeutic drug classes and subclasses to identify the fastest growing categories of drugs. They found evidence of the rapid take-up of new drugs as well as rapid growth in expenditures for antidepressants, antipsychotics, antihyperlipidemics, antidiabetic agents, antihistamines, COX-2 inhibitors, and proton pump inhibitors. In some cases, these increases were the result of higher expenditures per user and in others cases, the result of an increase in the number of people using the medications. Medicaid programs may want to reassess their cost-containment policies in light of the rapid take-up of new drugs, suggest the researchers.

Miller, G.E., and Hudson, J. "Children and antibiotics: Analysis of reduced use," pp. 36-44.

This study used MEPS data to examine trends in antibiotic use among children from 1996 to 2001, a period that followed the launch of national campaigns to promote the appropriate use of antibiotics. It also examined how changes in ambulatory visits and prescribing contributed to these trends. From 1996 to 2002, children's use of antibiotics sharply declined by 8.5 percent overall and 5.1 percent for respiratory tract infections. The apparent response to campaigns to reduce inappropriate antibiotic use was widespread, as reductions in use were found in all subgroups of children examined. However, the decline in overall antibiotic use for white children was more than double the decline for black or Hispanic children.

Cohen, S.B., Ezzati-Rice, T., and Yu, W. "The utility of extended longitudinal profiles in predicting future health care expenditures," pp. 45-53.

A small proportion of the U.S. population accounted for a large share of the $810.7 billion in estimated U.S. health care expenses in 2002. This study used MEPS longitudinal data to examine the capacity of alternative models to predict the likelihood of an individual incurring high levels of medical expenditures in a subsequent year. A predictive model that only included medical expenditures from the prior year performed quite well. This model correctly classified half of the 2000-2001 MEPS panel in the top decile of health care expenditures in 2001. Another model that covered health care expenditures for 2 years prior to the target year revealed only marginal gains at best in predictive capacity.

Fleishman, J.A., Cohen, J.W., Manning, W.G., and Kosinski, M. "Using the SF-12 health status measure to improve predictions of medical expenditures," pp. 54-63.

Self-reported health status is useful in predicting future medical expenditures, conclude the authors of this study. The researchers used data from the 2000-2001 MEPS panel, which included the SF-12 Health Survey of physical and emotional functioning. Over 5,000 people completed the SF-12 health status questionnaire and were interviewed about their demographic characteristics and selected chronic conditions. The researchers also examined data on medical expenses incurred subsequent to the interview. Adding data on self-reported health from the SF-12 improved the prediction of medical expenditures. In a model including demographic characteristics, chronic medical conditions, and previous expenditures, adding the SF-12 considerably increased the ability of the model to predict future medical expenditures.

Kirby, J.B., Taliaferro, G., and Zuvekas, S.H. "Explaining racial and ethnic disparities in health care," pp. 64-72.

This article demonstrates the capacity of the MEPS to help explain racial and ethnic disparities in health care. The researchers linked data from the 2000 and 2001 MEPS to detailed neighborhood characteristics from the Census Bureau and local provider supply data from the Health Resources and Services Administration. They found that insurance status and socioeconomic differences explained a significant portion of racial and ethnic disparities in health care. Also, neighborhood racial and ethnic composition accounted for a large portion of disparities in care access (use of ambulatory care during the year). The study also found substantial variation in the level of disparities among different groups of Hispanics. What's more, language differences helped explain observed disparities in the access measures used in the study.

Hill, S.C., and Pylypchuk, Y. "Reports of fewer activity limitations: Recovery, survey fatigue, or switching respondent?", pp. 73-81.

Researchers can be confident in reports of recovery from activity limitations in the MEPS, especially when disability status is self-reported, conclude the authors of this study. They assessed the reliability of reported recovery from activity limitations elicited from two types of questions using the second panel of MEPS. The questions asked about limitations in activities of daily living (ADLs) such as dressing and feeding oneself, and instrumental activities of daily living (IADLs), such as shopping and doing housework. Within an interview, they found substantial reliability for both ADLs and IADLs. Individuals with more severe disabilities were less likely to report functional recovery, which is consistent with accurate reporting. Controlling for disability severity, type of respondent (self- and proxy-response) affected reported recovery.

Editor's Note: A limited number of single copies of Medical Care 44(5 Suppl.) are available from AHRQ Publications Clearinghouse (AHRQ Publication No. OM-06-0074).

Return to Contents
Proceed to Next Article