Access to Health Care:

Bridging the Gap Between Policy and Research

Issue Brief No. 08
April 1997
Heidi H. Whitmore

ccess to care has long been studied by researchers, but emphasis has shifted recently from measuring whether access is worse for the poor and uninsured to monitoring how access is changing over time. Only a limited number of measures have been used consistently to monitor changes longitudinally, and these suggest that access by the uninsured is declining. There are potentially more valuable measures to track access, but data that are consistent over time have not been available. This Issue Brief, which is based on a seminar held by the Center for Studying Health System Change, discusses current research efforts to measure access to care.

Defining and Measuring Access to Care

onitoring access to care is not an easy task. Access is a broad and often vaguely defined concept that incorporates various dimensions of health care, including the supply and availability of health care providers, health insurance coverage and problems that individuals encounter in getting care.

There is no gold standard to measure access, and new approaches are constantly being developed to reflect the changes in the delivery of services and an increasing interest in outcome-based measures of access.

This is not an arcane research problem. Policy makers are very interested in research on the health care system -- particularly how policy changes can affect the way health care is delivered. Unfortunately, there is not a lot of research available to give policy makers the longitudinal information they need on access to care. And since much of the research they see has apparently conflicting results, they have not been inclined to make use of it.

While researchers have relied primarily on population-based surveys to study access, a separate area of research has used the level of charity or uncompensated care given by hospitals and other health care providers as a measure of health services provided to the uninsured. This approach has several limitations, however.

Changes in the amount of uncompensated care provided may not be a valid indicator of changes in access because it may reflect the quantity of care supplied rather than the quantity of care demanded by those without insurance. For example, a decline in the amount of uncompensated care could reflect diminished resources available to hospitals to provide it rather than a reduction in needs for free care in the community. Indeed, in some communities, changes in the amount of uncompensated care could reflect changes in supply (all needs cannot be met), while in other communities, changes in the amount of uncompensated care could reflect changes in demand. Only in the latter would uncompensated care trends parallel trends in access to care by the uninsured.

There are other limitations as well. First, using uncompensated care as a measure of access to the uninsured may underestimate the amount of services actually given free to the indigent. This is because government subsidies and private sector grants to hospitals for uncompensated care sometimes are not counted in the total amount of uncompensated care given. Second, uncompensated care may overstate the amount given because uncompensated care traditionally is defined as the sum of charity care and bad debt, and some bad debt may have nothing to do with providing care to the uninsured. For example, a hospital could have a bad debt because an insurance company has not paid for care delivered to one of its insureds. And third, most databases with measures of uncompensated care do not separate inpatient from outpatient care. This means that a shift from inpatient to outpatient care -- which could affect the level of care given -- is not captured.

A more widely used and more direct approach to measuring access -- population-based surveys -- has many advantages over data on uncompensated care. Population-based surveys use a broader set of measures that capture dimensions of access, including aspects of primary care, the process of care seeking (for example, the ease and convenience of getting to a doctor), barriers to care (for example, language and transportation problems) and unmet health needs. With these surveys, researchers can make inferences about who is at greatest risk for lacking access to care by comparing vulnerable populations, such as the uninsured, poor and low income persons and persons in poor health, to the rest of the population.

Population-based surveys, however, are not the best instruments to capture health outcomes as they relate to access. For example, the surveys cannot tell researchers whether individuals are receiving good-quality care in the most appropriate setting and at the least possible cost. To examine these issues, researchers use hospital discharge data to measure of rates of avoidable hospitalizations. Avoidable hospitalizations refer to a set of conditions -- such as asthma and diabetes -- that are considered treatable in ambulatory settings and for which hospitalization indicates a lack of access to high-quality primary care. Interest in this type of outcome-based measure is likely to increase.

Back to Top

Trends in Access

eter Cunningham of the Center for Studying Health System Change highlighted the difficulty of making inferences about access for the uninsured based on changes in uncompensated care. Using physician surveys conducted by the American Medical Association (AMA) and reports of physician visits from the National Health Interview Survey (NHIS), he found that the amount of charity care provided by physicians increased substantially between 1990 and 1994, but the average number of physician visits for uninsured persons did not change during the same period and was consistently well below that of insured persons. These findings call into question whether inferences about changes in access to care by physicians can be made based on changes in uncompensated care.

Changes in the amount of uncompensated care provided may not be a valid indicator of changes in access . . . . A more widely used and more direct approach to measuring access -- population-based surveys -- has many advantages over it.

Inferring changes in access is even more problematic when measures of hospital uncompensated care are used. Although hospital uncompensated care increased somewhat during the 1980s, the number of uninsured persons also increased. Moreover, it is not clear whether increases in hospital use indicate better or worse access overall. For example, the percent of hospital discharges for ambulatory-sensitive conditions (that is, avoidable hospitalizations) increased between 1980 and 1994 for the population-at-large, but the rate of increase was twice as high for uninsured persons. This suggests that access to primary care for uninsured persons deteriorated over this time period. Thus, in an era that increasingly emphasizes primary care and cost-effectiveness, it is unclear whether increases in hospital uncompensated care should be regarded as indicative of an increase or decrease in access for uninsured persons.

Back to Top

Lack of Consistency

n spite of the relatively numerous surveys that have been conducted by both government and private organizations, the ability to observe trends in access to care over the past 10 to 15 years has been limited. This is due in large part to the lack of consistency of measures and to methodological differences among surveys that can affect the results -- even if the measures used are identical.

Comparisons over time generally are valid only when using longitudinal surveys that maintain the same basic design and methodology and employ virtually identical questions.

Comparisons over time generally are valid only when using longitudinal surveys that maintain the same basic design and methodology and employ virtually identical questions. Until recently, the only annual health care survey was the NHIS, conducted by the National Center for Health Statistics (NCHS). Other surveys have been done infrequently, such as the National Medical Expenditure Survey (NMES) conducted by the Agency for Health Care Policy and Research (AHCPR) in 1977, 1987 and 1996, and the Access to Care Survey conducted by The Robert Wood Johnson Foundation (RWJF) in 1978, 1982, 1986 and 1994.

But even with these surveys, tracking some measures of access is still limited. The NHIS, for example, did not include measures of access other than service use until the late 1980s, and in any case, many researchers consider service use to be of limited importance as a measure of access. Also, many of the measures of access change from survey to survey.

Constructing measures that capture the concept of access is challenging, according to P. Ellen Parsons of NCHS, and the task becomes even more difficult given the rapid nature of health system change. For example, some researchers question whether tracking the usual source of care -- a traditional and widely used measure of access -- is even relevant for people in managed care plans. In addition, traditional survey methods do not collect information on severity and stage of disease at diagnosis, which is critical for the development of outcome-based measures.

Back to Top

A Case in Point

ne of the most important measures of access included in surveys is unmet health needs. It is virtually impossible, however, to make accurate inferences about changes in unmet health needs because there is no standardization among the surveys.

A recent study of unmet health needs, conducted by Karen Donelan of the Harvard School of Public Health and others, looked extensively at access problems and their consequences among the uninsured and insured populations. Some of the measures used in this study to ascertain access problems include the level of difficulty in obtaining needed medical care, difficulties in paying bills for medical care and the severity of health problems that individuals had difficulty getting care for.

Variations in Percent of the Uninsured Reporting Problems Getting Needed Care



6 NMES (1987)
13 NHIS (1993)
20 RWJF Access to Care (1986)
34 RWJF Access to Care (1994)
45 Donelan study (1996)

Unmet health needs were determined by asking respondents whether there was a time when they needed care but did not get it. In contrast to previous measures of unmet need, Donelan also asked whether respondents had difficulty getting care, even if they eventually received it. She found that a considerably higher percentage of the population had unmet health needs than was reported in previous surveys -- 45 percent of uninsured respondents in Donelan’s survey said they had unmet needs compared with a low of 6 percent in the 1987 NMES (see table). Several things could account for the wide discrepancies in estimates of unmet need from other surveys, according to Marc Berk of Project HOPE:

Lack of standardization and methodological differences make it virtually impossible to draw any conclusions about trends in unmet health needs from the various surveys. Furthermore, researchers can come to sharply different conclusions about changes in access to care, depending on which studies are used to compare the measures over time. For example, comparing the 1986 RWJF Access to Care survey with the 1993 NHIS shows that the rate of unmet health needs decreased by about 50 percent, but comparing the 1987 NMES with the 1996 Donelan study shows an increase of about 700 percent.

Back to Top

Addressing the Problem of Consistency

hese problems underscore the need for greater consistency in measuring access to care and for continuous monitoring. A number of activities undertaken recently by the federal government and private organizations suggest that these needs are beginning to be recognized.

Federal Government Activities. Until recently, the annual NHIS incorporated questions on access occasionally, and questions on health insurance coverage only every few years. The NHIS underwent a major restructuring starting in the early 1990s and now addresses barriers to care, unmet needs and other issues relating to access. Periodic supplements to the NHIS also examine the duration of respondents’ relationship with the usual source of care, satisfaction with care and reasons for delays in receiving care. Other measures include efforts to obtain needed care, whether care was eventually obtained and the impact of not getting needed care on health.

AHCPR recently created the Medical Expenditure Panel Survey (MEPS) to replace the NMES. While the NMES was conducted infrequently, the MEPS will be conducted annually and will permit continuous monitoring of access, along with health care utilization, health insurance coverage and expenditures. The MEPS will approach access issues in a manner similar to that of the NHIS and will be able to track the relative burden of out-of-pocket costs for health care.

The Medicare Current Beneficiary Survey (MCBS) conducted by the Health Care Financing Administration (HCFA) is another example of a recent government effort to improve monitoring of access to care. Although the MCBS’s central goal is to track health care expenditures for the Medicare population, the study includes special supplements on access that are administered on a continuous basis. The access measures used are similar to those in the NHIS and MEPS.

Researchers need to get their findings about trends in access to policy makers so they can develop insights into this critical dimension of how well the health care system is performing.

Center Activities. The Center for Studying Health System Change has placed a priority on tracking access through its household and physician surveys, which are a part of the Community Tracking Study. The Center will track changes in access through such measures as difficulty obtaining needed care, ease and convenience of getting care, usual source of care, charity care and physicians’ ability to obtain referrals to specialists. The surveys will be conducted every two years, which will allow the Center to track longitudinal changes in access. The surveys currently underway will provide baseline data.

In addition to these surveys, the Center will track other aspects of access -- for example, the organization and viability of the health care safety net -- as a part of the Community Tracking Study’s site visits to 12 communities across the country.

With so many changes in the health care system that affect access to care by vulnerable populations -- decreased ability of providers to shift costs, cutting back on public funding for the safety net, decline in employment-based insurance coverage -- researchers need to get their findings about trends in access to policy makers so they can develop insights into this critical dimension of how well the health care system is performing.

This Issue Brief is based on a health policy research seminar sponsored by the Center for Studying Health System Change, which took place in Washington, D.C., on March 7, 1997.
Peter Cunningham -- Center for Studying Health
    System Change

Karen Donelan -- Harvard School of Public

P. Ellen Parsons -- National Center for Health
Chip Kahn -- Ways and Means, Subcommittee on
    Health, U.S. House of Representatives
Paul B. Ginsburg -- Center for Studying Health
    System Change

Back to Top