Statement of Curt D. Mueller, Ph.D., Director,
Project HOPE Walsh Center for Rural Health Analysis, Bethesda, Maryland
Testimony Before the Subcommittee on Health
of the House Committee on Ways and Means
Hearing on Rural Health Care in Medicare
June 12, 2001
Good morning, Madam Chairwoman Johnson, Mr. Stark, and others members of the Subcommittee. I am Curt Mueller, Director of the Project HOPE Walsh Center for Rural Health Analysis. The Walsh Center is one of several research and policy centers funded in part by the federal Office of Rural Health Policy (ORHP). My testimony today reflects my views as an economist and health policy analyst; my views should not be regarded as representing the positions of Project HOPE, the Walsh Center, or ORHP.
I am very pleased to be here to discuss access to care issues pertaining to Medicare beneficiaries. My bottom line is that although there is some good news to report, recent evidence suggests that rural residents face access to care problems. Equity of the Medicare program is compromised. At the same time, there are policy approaches that would help address these problems.
What Does "Access" Mean?
Access to care refers to the potential and actual entry by an individual or population group into the health care system (Aday, Andersen, Fleming 1980). At the outset, it is important to recognize that access is not always realized: not every one, for various reasons, experiences actual access to care during a defined period of time. But access is difficult to measure when defined in this way. Among the most widely used measures of access are measures of actual use, or utilization. These include simple counts, e.g., the average, per capita number of physician visits per year, and the percent of the population with at least one hospital stay during the year. These measures, although frequently used, may not take into account differences in the clinical need for care or health status of the population that is being studied, differences in attitudes toward use of the medical care system, differences in practice style of providers, and other factors.
With the recent availability of large, claims- and patient-level data bases and advances in data handling, more sophisticated measures of utilization that account for medical need can be studied for large populations. Examples include measures of the percent of the elderly, diabetic population receiving an annual eye exam, and the percent of women in various cohorts, e.g., defined by age and medical need, who have received a mammogram during the previous year.
Finally, access has been measured in other ways, some of which were designed in attempts to measure the extent to which persons do not achieve access to services and reasons why. Survey respondents, for example, have been asked whether care was recently needed but not obtained due to cost or other reasons (Schur and Franco 1999; Mueller, Schur, Paramore 1998). Researchers have also studied satisfaction with various components of the health care delivery system, arguing that satisfaction is an important dimension of access. A battery of questions that separately assess satisfaction with cost, quality, and other dimensions of utilization are routinely asked as part of the HCFA-sponsored Medicare Current Beneficiary Survey.
Factors that affect access to care - by affecting both whether any care is actually received, and the level or intensity of services received -- include those that health services researchers refer to as determinants of ability-to-pay, need, and availability. For the Medicare beneficiary, these factors include: family income; whether the beneficiary has insurance that supplements Medicare and the nature of the supplemental coverage (e.g., benefits, cost-sharing); health status; attitudes regarding relationships between personal health and the health care delivery system; the supply of providers in the beneficiary's community; practice style of community providers; and the non-monetary "costs" of care, including travel time and expenses.
It is often argued that rural beneficiaries tend to be poorer, which -- by limiting ability to pay for care -- reduces access. Relatively more persons in non-metropolitan areas have individually purchased supplemental coverage than in metropolitan areas where supplemental coverage is relatively more likely to be subsidized by a previous employer. (1) Individually purchased coverage is often less generous with respect to benefits than employer-subsidized coverage. Beneficiary coverage of prescription drugs, for example, is more common in urban than in rural areas because employer subsidized coverage is more common and more likely to include a drug benefit (Mueller and Schur 2001). Finally, the importance of travel and time costs in rural areas should not be underestimated. The distance traveled to providers of care and associated time and travel costs can be very high in rural areas, which reduces access measures by reducing patient demand and utilization.
Access to care among rural and urban Medicare beneficiaries is comparable in many respects.
The good news is that access among rural and urban beneficiaries is comparable in many respects. Access indicators for rural and urban populations are similar for a number of dimensions of care. Rural beneficiaries are as likely to see a physician during the year as their urban counterparts (Mueller, Schoenman, Dorish 1999). Differences in the average annual number of physician visits between urban and rural beneficiaries exhibit some variation, depending on the definition of "rural," are usually small (NCHS, 2000b; Mueller, Schoenman, Dorosh 1999; Coburn and Bolda 1999; Krauss, Machlin, Kass 1999; Clark and Dellasega 1998).
Evidence from a recent analysis of Medicare claims submitted on behalf of more than a quarter million program beneficiaries indicates that rural residents are just as likely to receive much of the necessary care received by urban beneficiaries (Hogan 2001). The author of this analysis defines necessary care as care for which benefits are expected to be substantial and outweigh associated risks, and care, which if not provided, was viewed as inappropriate by a panel of physician experts. For example, one necessary care indicator is whether a mammogram is performed annually for patients with a history of breast cancer. The author examines to what extent 46 necessary care components are met in five different types of rural and in urban areas. With respect to mammography, the exam was received by 69 percent of the female population with a history of breast cancer in urban counties, and by between 68 and 69 percent of the populations in rural counties, depending on the extent of "rurality." No important rural-versus-urban differences were found for 27 of the 46 indicators studied, and "for the typical indicator, most geographic areas differ by just a few percentage points" (Hogan 2001). (2) Finally, rural access deficiencies tend to be most severe in the least populated, most-rural of rural areas. (3)
Data also suggest that Medicare beneficiaries in rural and urban areas are satisfied with certain dimensions of access to care. In 1996, most beneficiaries (96 percent) indicated satisfaction with the overall quality of care, and little variation existed within and across residents of rural and urban counties (Mueller, Schoenman, Dorosh 1999). Contrary to expectations, rural beneficiaries are not less satisfied with the availability of care on nights and weekends. While fewer elderly in the most remote rural counties are satisfied with the ease of commuting (92 percent in remote counties, versus 94 percent in the largest metropolitan counties), the difference is small and not statistically significant (Mueller, Schoenman, Dorosh 1999).
But rural beneficiaries face access problems.
The importance of rural deficiencies in access to necessary care identified by Hogan (2001) should not be understated. For example, rural populations are significantly less likely to receive timely EKGs for congestive heart failure and transient ischemic attack, follow-up care after hospitalizations for diabetes and gastrointestinal bleeding, timely gall bladder removal for symptomatic gallstones, and screening mammography (Hogan 2001). Once again, these deficiencies are for care deemed medically necessary.
It is also important to observe that these access deficiencies are most severe in the most rural of rural areas. Although this is not surprising, given the lack of providers in these areas, this point is very important to emphasize because much of the research on access to date does not thoroughly address the extent to which access differences vary within rural areas. Part of the reason is that national databases that have traditionally been used to research access to care issues do not have enough data on rural populations to generate reliable estimates for the variety of rural settings that often get lumped together under the label, "non-metropolitan."
Evidence from a number of other studies also indicates that rural beneficiaries have less access to care than their urban counterparts. Rural Medicare beneficiaries are less likely than urban beneficiaries to: receive pap smears (Stearns, Slifkin, Edin 2000); have an emergency department visit during the year (NCHS 2000a; Lishner, et al. 2000; Miller, Holahan, Welch 1995); and use auxiliary home health services, including physical therapy and medical social services (Sutton 2000; Kenney, 1993). Rural residents are less likely to have access to specialty physician services, and the average physician visit for rural residents is less resource-intensive than for metropolitan area beneficiaries (Sutton 2000; Miller, Holahan, Welch 1995).
How access to care that requires an inpatient stay varies between rural and urban beneficiaries is unclear. Recent findings indicate that the rural elderly are slightly more likely to be hospitalized during the year and that their lengths of stay tend to be shorter than in urban areas (NCHS 2000a; Schur and Franco 1999; Krauss, Machlin, Kass 1999). In thinking about what these findings say about access, it is important to remember that use of hospitals is not a "complete" picture of the illness episode. Payment system incentives often encourage shorter lengths of stay and discharge to a "lower level" of care. Complete episodes of illness for the elderly often involve skilled nursing care (either in a skilled nursing facility bed or at home), home health services, or both after the patient's discharge for an acute care hospital. We know that rural residents receive more skilled nursing facility care (Coward, Netzer, Mullens 1996; Coward, Horne, Peek 1995), that receipt of this care appears to vary indirectly with the availability of home health services (Dubay, 1993) and directly with the supply of skilled nursing facilities in rural areas (Coburn and Bolda, 1999). While these findings are consistent with the substitutability of certain kinds of services, currently available access measures do not address the mix of care received during the entire episode. Additional study on the use of post-acute care services in rural versus urban areas is needed before we can say more about access to care that requires an inpatient hospital stay.
Do these differences matter?
The evidence cited above suggests that rural residents have less access to certain types of care than urban residents. These differences are important for at least two reasons. First, these differences are of at least some clinical significance. Although there is often disagreement among medical experts in defining medical need and exact prescriptions of what constitutes optimal levels of services for rural versus urban populations, some of the observed differences are large enough to be of concern to clinicians, e.g., the large differences in access to types of services that clinicians have deemed necessary.
Second, these differences are important from an equity perspective. A basic objective of the Medicare program is that beneficiaries - regardless of geographic location - should have equitable access to adequate medical care. Rural residents should have adequate access to basic preventive, primary care, and emergency medical services. But as access differentials do exist, especially for residents in the most rural of rural areas, medical care expenditures by urban beneficiaries are considerably greater than for rural beneficiaries. In 1996, per capita personal health care expenditures for non-institutionalized Medicare beneficiaries residing in metropolitan areas was $6,901 and $5,901 for non-metropolitan beneficiaries (Liu et al. 2000). Differences are even larger after controlling for differences in supplemental coverage across beneficiaries. Per capita expenditures for the populations of beneficiaries without any supplemental coverage - that is, with only Medicare coverage - were $5,248 in metropolitan and $3,860 in non-metropolitan areas (Liu et al. 2000). This difference, 37 percent, is an amount larger than could possibly be accounted for by geographic differences in the costs of providing care.
It is important to emphasize that equity with respect to rural versus urban residence does not necessarily require that per capita program expenditures be equal. A variety of factors determine per capita expenditures on medical care, just as a variety of factors affect access. For example, expenditure differences reflect differences in practice style and in the prices of care. By definition, expenditures will be larger in areas with higher Medicare payment rates and in areas where service mix is relatively resource-intensive. Thus, because payments under the Medicare Fee Schedule have tended to be lower in rural areas and service mix tends to be less resource-intensive, per capita expenditures will be greater in urban areas. On the other hand, removing access to care barriers faced by rural residents will certainly increase rural expenditures relative to urban expenditures, and improve equity of the Medicare program.
What policies might improve access among rural beneficiaries?
Several policy responses might be considered as we think about Medicare reform and restructuring. First, policies that increase the supply of health care resources in rural areas should help improve access. At present, relative supplies of physicians and hospital resources are larger in urban than in rural areas because markets - in terms of numbers of people - are larger. The number of specialist physicians per capita is larger in urban areas, so access to specialty care is easier to obtain. A corollary is that certain "fixed" costs will be lower in more-populated markets, so hospitals will be relatively scarce and more costly to operate in rural areas. Personal preferences of providers also constrain supply in some areas. Some rural markets have difficulties attracting and retaining physicians and other providers. Access indicators for these areas will differ from measures for areas where effective supplies of providers are larger.
Supply in rural areas would be enhanced by workforce policies that provide additional incentives for physicians to locate and maintain practices in rural areas, and by payment policies that assist rural providers and help ensure their financial viability in rural areas. An example is the Rural Hospital Flexibility Program, which has granted cost-based reimbursement to many of the smallest rural hospitals. This program appears to be helping these facilities overcome financial problems associated with low volume and allowing them to continue help meet community needs.
Second, policies that improve rural beneficiaries' access to expanded benefits should help improve access in rural areas. For example, rural access will improve with a Medicare drug benefit so long as rural residents can obtain coverage from a plan that provide drug coverage. Currently, rural residents are less likely than their urban counterparts to have this coverage because they have less access to employer-sponsored private supplemental coverage, Medicare HMOs, and Medicare+Choice plans that offer a drug benefit. Furthermore, rural residents tend to be poorer and less able to afford an individually purchased supplemental plan with drug coverage. For these reasons, federal subsidization of a drug benefit based on income will likely improve access in rural relative to urban areas.
To the extent that these policy approaches improve access in rural areas, equity of the program from the rural perspective would improve because real utilization in rural relative to urban areas would increase. In the short run, other policy options that would reduce rural inequity by reducing rural beneficiaries' share of program expenses might be studied. One approach is to tie cost sharing to program expenditures per capita. For example, the Medicare premium could be reduced in lower cost (mainly rural) areas. Beneficiaries would be expected to pay a premium equal to 25 percent of program expenditures in rural areas of the state or region, for example. Another possibility is to increase program payment rates to providers under traditional Medicare in historically low-cost, under-served areas. This might be accomplished, for example, either through the current bonus payment mechanism that targets the bonus to physicians in shortage-designated areas, or through the work component of the Medicare Fee Schedule. The premium adjustment would benefit rural beneficiaries directly, whereas adjustments in payment rates might affect improvements in access over time by helping to direct physicians to areas with the greatest need for physicians.
One final note. Although the primary focus of this statement is on access to care among rural Medicare beneficiaries, it is important to note that the Medicare program has been very successful at reducing monetary barriers of access to care. Data indicate that access differentials, like those noted above, also distinguish the rural non-elderly from the urban non-elderly, but that monetary barriers are more severe because of the lack of insurance among the non-elderly and that this problem is relatively more severe in rural than in urban areas. Because the Medicare population is relatively larger in rural than in urban areas, policies designed to improve access to care among the elderly are likely to strengthen the rural health infrastructure as a whole, which in turn should improve access for the entire rural population.
Aday L.A. Andersen R. Fleming G.V. Health Care in the U.S. Beverly Hills: Sage Publications. 1980.
Clark D. Dellasega C. Unmet Health Care Needs: Comparison of Rural and Urban Senior Center Attendees. Journal of Gerontological Nursing; 1998; 24(12):24-33.
Coburn A. Bolda E. The Rural Elderly and Long-Term Care. In TC Ricketts, III, (ed.), Rural Health in the United States. New York: Oxford University Press. 1999.
Coward R. Horne C. Peek C. Predicting Nursing Home Admissions Among Incontinent Older Adults: A Comparison of Residential Differences Across Six Years. The Gerontologist; 1995; 35(6):732-743.
Coward R. Netzer J. Mullens R. Residential Differences in the Incidence of Nursing Home Admissions Across a Six-Year Period. Journal of Gerontology: Social Sciences; 1996; 51(5):S258-S267.
Dubay L. Explaining Urban-Rural Differences in the Use of Skilled Nursing Facility Benefit. Medical Care; 1993; 31(2):111-129.
Hogan C. Urban-Rural Differences in the Use of Needed Services: Analysis of the ACE-PRO Indicators Using 1998/1999 Data. Report prepared and presented to the Medicare Payment Advisory Commission, January 3, 2001.
Kenney G. Is Access to Home Health Care a Problem in Rural Areas? American Journal of Public Health;1993; 83(3):412-414.
Krauss N. Machlin S. Kass B. Use of Health Care Services, 1996. Agency for Health Care Policy and Research; 1999; MEPS Research Findings No. 7. AHCPR Pub. No. 99-0018.
Lishner D. Rosenblatt R. Baldwin L. Hart L. Emergency Department Use by the Rural Elderly. Journal of Emergency Medicine; 2000; 18(3):289-297.
Liu H. Ginsberg C. Olin G.L. Merriman B. Health & Health Care of the Medicare Population: Data from the 1996 Medicare Current Beneficiary Survey. Rockville, MD: Westat, 2000.
Miller M. Holahan J. Welch W. Geographic Variations in Physician Service Utilization. Medical Care Research and Review; 1995; 52(2):252-278.
Mueller C.D. Schoenman J.A. Dorosh E. The Medicare Program in Rural Areas. In TC Ricketts, III, (ed.), Rural Health in the United States. New York: Oxford University Press. 1999.
Mueller C.D. Schur C. Drug Coverage of the Rural Elderly and Implications for a Medicare Benefit. Paper presented at the annual meeting, National Rural Health Association, Dallas, May 2001.
Mueller, C. Schur, C. Paramore, C. Unmet Dental Health Needs: Estimates from the Robert Wood Johnson Foundation Access to Care Survey, Journal of the American Dental Association 129 1998, 429-437.
National Center for Health Statistics (NCHS). Health, United States, 2000. Hyattsville, MD. 2000a.
National Center for Health Statistics (NCHS). NCHS Series 10, No. 200. From NCHS publications CD, Vol. 1, No. 6; Hyattsville, MD. July 2000b.
Schur C. Franco S. Access to Health Care. In TC Ricketts, III, (ed.), Rural Health in the United States. New York: Oxford University Press. 1999.
Stearns S. Slifkin R. Edin H. Access to Care for Rural Medicare Beneficiaries. The Journal of Rural Health; 2000; 16(1):31-42.
Sutton J. Characteristics of Rural Home Health Users and Implications for Payment Reform. Project HOPE Walsh Center for Rural Health Analysis Working Paper, 2000.
1. Supplemental coverage comes from a variety of sources. Private supplemental, or Medigap coverage, is either purchased directly by the beneficiary or is obtained through an employer. The term, "supplemental coverage" also includes Medicaid, for those Medicare beneficiaries who are eligible.
2. Other examples of necessary care indicators are: gastrointestinal work-up for patients with iron deficiency anemia; follow-up visits for hospitalization within one week of initial diagnosis of unstable angina; chest x-ray within three months of initial diagnosis of congestive heart failure; and visit within two weeks following discharge of a patient hospitalized for depression.
3. Rural deficiencies are statistically significant and the divergence from the corresponding urban value is 10 percent or greater in the most-rural areas for 15 of the 19 indicators for which rural deficiencies are observed.