Archive

CHAPTER EIGHT

____________________________________________________________________________

Focusing on Vulnerable Populations

Throughout its work, the Commission has been particularly cognizant of the special concerns associated with assuring quality health care for vulnerable populations. Individuals may be vulnerable to health care quality problems for one or a combination of underlying reasons, including those pertaining to their financial circumstances or place of residence; health, age, or functional or developmental status; or ability to communicate effectively. Other personal characteristics, such as race, ethnicity, and sex also have been shown to be associated with differential experiences in obtaining quality health care. Furthermore, characteristics of the evolving health system -- such as an eroding safety net -- can interact with personal characteristics to contribute to vulnerability.

There is a clear need to increase the level of attention paid to vulnerable groups, including both those who, because of their chronic illness or disability, have many interactions with the health system, and those who have difficulty accessing the system and may be most likely to fall through the cracks during this period of rapid health system change. In the Consumer Bill of Rights and Responsibilities, the Commission articulated that consumers have the right to considerate, respectful care, free of discrimination. In this chapter, the Commission extends this work by specifically addressing ways in which the particular needs of vulnerable populations should be accounted for in the design of systems for health care delivery, quality measurement, and payment.

RECOMMENDATIONS

Additional investment should be provided for developing, evaluating, and supporting effective health care delivery models designed to meet the specific needs of vulnerable populations. The lack of evidence of effective approaches to enhance the health and functional status of persons with chronic illness or disabilities suggests the need for significant additional investment in research and innovation for health care and rehabilitation programs serving these patients. In addition, telemedicine and similar innovations, incentives for health care professionals to practice in underserved areas, and enhanced availability of prehospital emergency services should be assessed as approaches for improving the access to care of those facing geographic barriers to appropriate care.

Development and evaluation of health care quality measures and measurement methods for vulnerable populations should be supported. A focus on vulnerable populations in the design of quality measurement initiatives is valuable, in part, because their experiences may provide new insight on systemwide problems. Attention to vulnerable populations should be integrated within general quality measurement strategies, recognizing that quality measures and measurement methods tailored to vulnerable groups will, in some cases, be required.

Payments to health plans and providers should promote quality health care and improved health and functional status for all patients, including vulnerable populations. Adjusting payments for differences in health or functional status is especially important for Medicare, Medicaid, and other payers that have significant enrollment of individuals with chronic illness or disability so that health plans and providers have an incentive for developing innovative models of care that best serve these individuals. Risk-adjusted payments are also critical to sustaining the safety-net mission of certain providers that provide a disproportionate amount of care to vulnerable populations, such as community health centers, rural health clinics, and academic health centers. The Federal government should convene high-level stakeholders to determine how best to implement risk-adjusted payments for Medicare. Payment systems also can be better aligned with quality care for vulnerable populations by being broadly based across groups of providers and related health care services to encourage multidisciplinary, coordinated care; providing coverage for health care delivery approaches that have been demonstrated to improve outcomes, functional status, and satisfaction; and rewarding quality performance through financial incentives.

SOURCES OF VULNERABILITY

Several specific groups within the population have been shown to be especially vulnerable to health care quality problems, differential experiences in the health system, or increased burden of poor health. Among others, these groups include the uninsured, the poor, the elderly, children, those living with mental or physical disabilities, racial and ethnic minorities, and the terminally ill. The vulnerability that these groups experience usually can be attributed to one or some combination of three factors: economic status or geographic location; health, functional, or developmental status; or ability to communicate. Not all sources of vulnerability can be fully explained by these factors, however, raising important questions for further research.

Economic Status and Geographic Location

Individuals may be vulnerable by virtue of their financial circumstances or geographic location. These factors can present obstacles to obtaining needed health care and can result in increased exposure to health risks. Those who disproportionately experience access problems include those whose income and/or health insurance status place them at increased risk for encountering barriers to accessing needed services and those who live in certain rural or inner-city areas that have a shortage of qualified health care professionals. The estimated 41.6 million Americans who have no health insurance most often face the greatest access barriers (U.S. Census Bureau, 1997). These barriers to access can lead to a lack of continuity, delays in obtaining care, and limited choices about where and from whom care may be received (Newacheck et al., 1996; Lambert and Agger, 1995). These patterns of utilization can contribute to adverse health care outcomes, including higher rates of preventable hospitalizations (Billings et al., 1996).

Poverty and lack of insurance can result not only in decreased access to health care, but also increased risk of poor health. Any of these factors can magnify exposure to environmental risks (such as secondary tobacco smoke, poor sanitation, or lead exposure), safety risks (such as traffic hazards and family violence), social and psychological stressors (such as fear of crime), and lack of infrastructure supports (such as counseling or educational services) that contribute to an increased burden of poor health.

Health, Age, and Functional or Developmental Status

Some individuals are vulnerable because the nature and/or severity of their illnesses or disabilities create special challenges in obtaining the needed range of services. For example, a recent study of Medicare enrollees in managed care plans found that nonelderly individuals with disabilities, the frail elderly, the functionally impaired, and persons in fair or poor health were much more likely than the general beneficiary population to report problems with access to care (Nelson et al., 1997). Many of these same groups also have been shown to be vulnerable to access problems in fee-for-service Medicare (PPRC, 1995). These vulnerable groups may be especially prone to difficulties in using the health care system, including discontinuity and lack of coordination among multiple providers, inability to obtain care from providers who have expertise in treating their conditions, and difficulty in receiving approvals for treatments that are expensive or not widely used. In addition, individuals with some conditions, such as mental illness or HIV disease, may face or fear undue stigma that makes it difficult to receive appropriate care.

Developmental status and age can also be associated with vulnerability. Children have health and developmental needs that are markedly different from adults, and require age-appropriate care. Developmental changes, dependency on others, and different patterns of illness and injury require that attention be paid to the unique needs of children in the health system (IOM, 1996). The elderly also have unique health care needs due not only to the increased incidence of illness and disability, but also to the multiple and complex interactions of other types of physical and social consequences of aging.

Communication Barriers

Vulnerability can in some cases be attributed to limitations in the ability to communicate with providers and other actors in the health care system. Communication difficulties may be associated with a person's level of education or development, language or cultural differences, health condition, or physical or mental disability. Persons who have difficulty communicating may experience problems in expressing treatment preferences, providing informed consent, obtaining services that are consistent with their cultural norms, finding providers who are sensitive to their particular concerns, getting problems resolved, and understanding or complying with treatment options.

Unexplained Vulnerability Associated with Race, Ethnicity, Sex

While categorizing sources of vulnerability in this manner is helpful in a number of respects, the above conceptual framework does not fully encompass the full range of vulnerable population groups. Other characteristics also have been shown to be associated with differential health care and inferior health status.

Significant differences in treatment have been documented by race, ethnicity, and sex that are not explained by other demographic differences, insurance status, clinical factors, or provider characteristics. For instance, African Americans with colorectal cancer have been found to be treated less aggressively than their white counterparts (Ball and Elixhauser, 1996). A number of studies have consistently demonstrated that African Americans are about half as likely as whites to receive interventional therapy for coronary artery disease (Ford and Cooper, 1995). Differences between Latinos and nonhispanic whites in the use of invasive cardiac procedures also have been documented (Carlisle et al., 1995). Another study showed that women with AIDS receive fewer services than men with AIDS, even after being diagnosed and having accessed the health system (Hellinger, 1993). These studies are just a sample of the large and growing health services research literature documenting the vulnerability of these groups.

Differences across racial and ethnic groups exist not only in treatment patterns, but also in indicators of health status. The persistence of significant racial and ethnic disparities in rates of infant mortality, cardiovascular disease, diabetes, stroke, HIV/AIDS infection, and other areas prompted a recent Presidential initiative to eliminate those gaps by the year 2010 (White House Office of the Press Secretary, 1998).

The underlying reasons for these demonstrated differences are not well understood. Whether these findings can be attributed to discrimination, cultural factors, or other causes is an important area for further study because of the implications for eliminating unwarranted variation in the provision of health services and for improving the health status of all Americans.

IMPLICATIONS FOR THE HEALTH SYSTEM

Understanding the underlying sources of vulnerability is critical, not only because of the need to influence the development of targeted quality improvement efforts, but also because addressing the problems vulnerable groups encounter requires coordinated efforts throughout the health system. In particular, populations vulnerable to health care quality problems need to be accounted for in the design of effective systems for health care delivery, the choice of appropriate health care quality measures, and the adaptation of payment mechanisms.

Systems to Serve Vulnerable Populations

With their focus on patients' acute and urgent problems, traditional models of health care delivery often lack the systematic assessments, preventive interventions, education, coordination and integration of care, and psychosocial support that vulnerable patients, particularly those with chronic conditions, also need (Wagner, 1996). Innovative models of health care delivery have emerged that attempt to better incorporate these elements of care, but continued research and development of more effective and efficient approaches for meeting the health care needs of patients with chronic conditions, disabilities and other sources of vulnerability remain important. While continued innovation in this area is important, it also must be accompanied by evaluation to determine effectiveness.

To date, there is a lack of evidence supporting improved functional status or other clinical outcomes resulting from many health care delivery approaches for individuals with chronic illness or disability. Some elements of health care delivery that are particularly relevant to patients with chronic conditions include the use of multidisciplinary teams, continuity of care, patient and family empowerment, case management, and outreach or home-based care. A review of more than 50 published studies assessing whether innovative health care programs emphasizing these elements improved outcomes for individuals with chronic illness or disability(1) found that improved clinical and functional outcomes were not consistently demonstrated. The review did find evidence that in general these approaches improved patient satisfaction with their care, but other improved outcomes were limited to specific models of health care within particular populations of patients. Specific areas where evidence demonstrated that innovative health programs improved outcomes are summarized below:

The lack of consistent evidence for improved clinical outcomes highlights the need for significant investment in research and innovation of health care delivery models addressing the needs of individuals with chronic conditions and other vulnerable populations. Evidence of enhanced patient satisfaction from models of health care delivery that emphasize continuity of care, multidisciplinary approaches, patient empowerment, and outreach to community settings should guide the development of innovative approaches to care that also aim to improve functional status and clinical outcomes. As evidence is developed to demonstrate effective models of care, health care providers should commit to practicing evidence-based care. Appropriate approaches to caring for individuals with chronic conditions, however, should recognize the unique needs of individual patients. For example, while innovative approaches to caring for individuals with mental illness (e.g., medications, a variety of outpatient treatment settings, and effective modes of rehabilitation and support) have benefitted many patients, the treatment of serious psychiatric conditions often continues to require the availability of inpatient care (NAMI, 1997).

Several health plans have designed innovative programs intended to provide better care for individuals with chronic conditions, and further demonstration and evaluation of such programs should be encouraged. Examples of these unique programs include:

Other populations beyond individuals with chronic conditions also face vulnerabilities that should be addressed in designing health care delivery systems. Delivery systems need to assure access to the specific types of care that are needed by subgroups of the population, including women's health care, geriatric care, and pediatric care. The availability of culturally-sensitive health care professionals and systems is particularly important for patients with communication barriers as well as for racial and ethnic minorities (as further discussed in Chapter 13). Telemedicine technologies, public programs that provide incentives for health care professionals to practice in underserved rural areas, and the availability of prehospital emergency services are delivery system characteristics that have the potential to improve health care delivery for patients whose location poses a barrier to accessing quality care. In addition, certain types of providers, including academic health centers and community health centers, have played an especially important role in delivering comprehensive, high quality care to vulnerable populations who otherwise could be underserved. Changes in the health care system that affect those organizations' ability to continue in this role have the potential to increase the problems of access to appropriate care that vulnerable groups disproportionately face.

Finally, establishing national aims for improvement, while not targeted specifically to vulnerable populations, is intended to stimulate efforts to improve the quality of health care delivery (see Chapter 3). The setting of aims, while national in focus, should be sufficiently robust to reflect the needs of specific vulnerable populations and localities. This can be achieved through establishing specific objectives within the broader aims that reflect the unique needs of vulnerable populations, and ensuring that the tracking of national aims can support efforts to identify particular areas of concern to local areas and vulnerable populations.

Quality Measurement

Vulnerable populations require special attention in the design of health care quality measurement strategies for three distinct reasons. First, a focus on the quality of care experienced by vulnerable populations could provide new insight into health system problems or identification of problems that otherwise could go undetected. Second, some vulnerable groups are more likely to fall through the cracks. The same factors that contribute to their vulnerability can also affect their ability to safeguard their own needs and interests adequately. Third, for a variety of technical and other reasons, health care quality problems experienced by vulnerable populations are not well captured by measurement efforts designed for the general population. For the most part, these efforts do not utilize data collected over time or across multiple different settings for health care delivery. Finally, because individuals with chronic illness and disability consume a disproportionate share of health care resources relative to their share of the population, measuring the quality of their experiences with the health care system will lead to a focus on high-cost areas.

Measurement Strategy Needed. Health care quality measurement for vulnerable populations either could be undertaken separately or be included in general strategies for evaluating quality. The combination of general and targeted strategies is most likely to identify efficiently and effectively the highest priority quality problems for vulnerable populations. In some cases, it may be less expensive to oversample one group in the population to explore whether the subgroup is experiencing more or different quality problems. In other cases, the use of specially tailored measures may be more effective.

Including vulnerable populations in general strategies for quality measurement is likely to be more sustainable over the long run and sends the important signal that evaluating quality for vulnerable populations is equally, if not more, important. This strategy allows comparisons between vulnerable and general populations when the same measures are used. Such approaches may be more efficient if special sampling procedures can augment a generally applied measure so that evaluating problems for vulnerable populations can be done simultaneously.

There are a few reasons, however, why targeted approaches might be necessary. For persons vulnerable due to economic circumstances, the key to including them in monitoring systems is the availability of variables to identify such persons and methods for bringing data on their patterns of utilization into integrated data systems (e.g., the experiences of uninsured persons who do not file claims). For persons vulnerable due to health status, the key to inclusion in monitoring is the expansion of appropriate criteria for evaluating their care and the development of methods that can support analyses of quality problems using small samples. For persons who are vulnerable due to communication challenges, the key to inclusion will be the development of measurement tools that can be used with such populations.

Different Quality Measures Needed. In some cases, different quality measures may need to be used to assess the care provided to vulnerable populations. Certain health care problems experienced by vulnerable populations differ from those encountered by others in the health care system (either by their nature or in the frequency with which such problems are encountered), and will require a different set of measures to evaluate the quality of care they receive. For example, evaluating the quality of care delivered to persons with certain rare conditions may require developing clinical measures specific to the particular condition (e.g., sickle cell disease). In other cases, existing measures might be adapted or refined to better accommodate vulnerable populations. Medicare and Medicaid, as public programs that insure a disproportionate share of many vulnerable populations (e.g., the frail elderly, nonelderly persons with disabilities), should continue to exercise leadership in supporting the development of quality measures to assess the health care delivered to these populations, in instances where the need for special measures is identified.

In efforts to develop new health care quality measures, the deficits in quality that pose the greatest risk to vulnerable populations and the types of measures most likely to identify such deficits should be taken into account. Focusing new measure development on technical process quality has a number of advantages. First, scientifically sound measures of technical process quality are the most sensitive dimensions for providing an early warning of potential threats to health outcomes. A good measure of technical quality is one that is linked to health outcomes-- meaning that adhering to the indicator process more often will increase the likelihood of producing better health outcomes. Process monitoring provides an early warning signal. When problems with processes are identified, interventions can be launched to improve processes before many individuals experience poor outcomes. Second, process evaluation allows quality improvement efforts to examine structural and access features that may be contributing to poor quality. Third, technical process is the area that is most difficult for individuals to evaluate for themselves. Because specialized work must be undertaken to develop scientifically sound process measures, this is an area requiring focused support (McGlynn, 1997).

In choosing areas for measure refinement, translating existing measures, where possible, into additional languages and exploring their relevance for different cultural or ethnic groups is likely to produce a high payoff (McGlynn, 1997). Measures that rely on data collected directly from patients' reports include those that address issues related to access, interpersonal style, communication, patient satisfaction and some outcomes (e.g., functional status). Rather than developing new measures in these areas, expanding the populations that could be included in such measures and compared with existing national data would produce more information quickly.

Improved Measurement Methods. Measurement methods that are appropriate or sufficient for evaluating health care quality in the general population may need modification to account for vulnerable groups. There is a need to address existing obstacles to quality measurement for vulnerable populations to advance in this area, including:

Small Numbers Problem. In terms of specific populations that could be subject to quality measurement, each individual vulnerable group (e.g., terminally ill persons) generally constitutes only a small proportion of the overall population (health plan enrollees, residents of a geographic area, patients of a specific physician) subject to measurement. Therefore, methods that evaluate quality for random samples of diverse populations are unlikely to include an adequate number of vulnerable persons in standard quality measures. A fundamental issue for most vulnerable populations is their low representation in general populations that are currently the subject of quality monitoring activities. Because the numbers are unlikely to change significantly, finding methods that do not require large sample sizes is a priority. Exploring techniques for aggregating measures across populations, conditions, settings, and providers offers one promising approach to the small numbers problem.

Identification of Vulnerable Populations in Existing Datasets. Existing datasets used in quality measurement often lack the variables that facilitate identification of vulnerable populations for quality assessment purposes. Routine availability of such information would make it feasible to undertake monitoring of vulnerable persons as a part of quality measurement initiatives more frequently. Methods for identifying vulnerable individuals may also be hampered by efforts to protect their privacy or to shield them from discrimination. For example, most health plans do not routinely collect information on members' income, education, and race; all of these factors can contribute to increased risk of poor quality care, but it is difficult to systematically identify these individuals.

Value of Qualitative Evaluation Methods. Qualitative methods may also prove valuable for identifying and exploring health care quality problems associated with vulnerable populations. Focus groups and case studies are among the approaches that may be more helpful than quantitative approaches under certain circumstances because they may be more sensitive to certain types of quality problems than other methods. These approaches tend to be more resource and labor intensive, and so must be used selectively.

Aligning Payment Mechanisms with Quality Care

Accounting for vulnerable populations in the payment of health care providers and health plans is critical to promoting access and high-quality health care for these groups. Current payment systems often are designed in ways that do not reward high quality care for patients with chronic conditions and other sources of vulnerability.

Basing Payments to Plans on Health or Functional Status. Health plans commonly receive a prepaid amount per member to cover needed services, but these "capitated" payments usually are not adjusted to reflect members' health and functional status. By adjusting payments for health status, health plans will be rewarded for developing quality programs that attract high-cost individuals with chronic conditions, whereas these incentives are lacking when paying plans average rates. Adjusting capitated payments to better reflect health and functional status (also known as risk-adjustment) is therefore an essential part of efforts to improve the quality of care and promote access to care for patients vulnerable to inadequate care due to the presence of chronic conditions.

Risk-adjusting payments is especially important for Medicare, Medicaid, and other payers that have many beneficiaries with significant disability or chronic illness. The ongoing health care needs that result from aging, chronic illness, and disability accentuate the importance of quality and the disadvantage of unadjusted rates. With unadjusted average rates, plans have an incentive to avoid improving quality in ways that might attract people with chronic conditions, such as kidney disease, paralysis, cancer, or heart disease, who are likely to need expensive services on an ongoing basis.

In addition to enrolling a large share of individuals with chronic conditions, Medicare and Medicaid's size enables them to widely influence the health care market by paying more for the enrollment of people with greater health care needs and achieve economies of scale in implementing an adjustment system. Some private group purchasers, particularly groups that purchase health coverage collectively, have also recognized the importance of adjusting payments for the health and functional status of enrollees and begun implementing risk-adjusted payments. For example, the Minnesota Business Health Care Action Group (a cooperative including 28 employers covering 250,000 individuals) and the Health Insurance Plan of California (a small group purchasing pool covering 120,000 individuals) have adopted risk- adjusted payments to health plans based on health status.

Various approaches have been developed to adjust capitation payments to reflect health needs. Medicare and other payers currently use easily-available demographic information, including age, sex, and place of residence, to set capitation rates, but these demographic characteristics generally are poor indicators of health and functional status. Other approaches for adjusting capitation rates include those based on data of prior health care expenditures; surveys of self- reported health status or assessments of functional status; and reporting of diagnostic information. While each of these approaches has merits, they also have drawbacks.

While the need to develop feasible and robust adjustment methods has stymied the application of risk-adjusted payments in the past, several diagnostic classification systems that payers need to implement risk-adjusted payment systems are currently available (Ellis, 1996; Kronick, 1996; Weiner, 1996). Beginning expanded use of such systems can serve both the purposes of developing improved adjustment methods and data collection as well as better aligning payment systems to quality care for individuals with chronic illness or disability.

Several state Medicaid programs, including Colorado and Maryland, have implemented risk- adjusted payment, and other states are moving forward soon or are actively engaged in planning. By 1999 it is likely that at least a half dozen states will be adjusting payments for substantial numbers of beneficiaries (Dreyfus, 1997).

The recent Balanced Budget Act of 1997 directs the Health Care Financing Administration (HCFA) to begin data collection in 1998 that will allow the implementation of risk-adjusted payment for Medicare in the year 2000. Prompt movement toward full diagnostic reporting is key to making risk-adjusted payment work. By announcing its intention of moving soon to risk- adjusted payment systems for Medicare, the Federal government has stimulated the need for improved inpatient diagnostic reporting. Plans are required to report inpatient data including diagnoses in January, with ambulatory data, at the discretion of the Secretary of Health and Human Services, potentially required in July 1998. The validity of ambulatory data in particular needs attention, but the quality of data systems is likely to improve rapidly once they are used for payments and as group purchasers increasingly request this information (see Chapter 14).

To assist Medicare in its transition to risk-adjusted payments, a group of representatives from HCFA, health plans, and Medicare beneficiaries should be convened to develop a broader consensus on adjusting payments to account for differences in enrollees' health or functional status and to help plan key steps in implementation. In addition, implementation should be phased-in to protect plans, providers, and beneficiaries from sharp changes in payments and to allow for development of efficient data reporting systems. Finally, HCFA should coordinate with state programs seeking to adopt risk-adjusted payment systems for Medicaid to avoid unnecessary duplication and reduce the data-reporting burdens facing health plans. As Medicare and Medicaid develop further experience with developing and implementing risk-adjusted payment systems, more private payers may find it feasible to also adopt payment systems that more accurately reflect the health and functional status of the populations they cover.

Other Payment Approaches. Payments to hospitals, physicians, and other health care providers also should be designed to encourage coverage of cost-effective models of care delivery that have been shown to attain better outcomes (including satisfaction and functional outcomes). Health plans that capitate provider groups should examine whether the groups vary significantly in the health and functional status of their enrollees and whether their payments should be adjusted accordingly. Payments can encourage multidisciplinary, coordinated care by being broadly- based among groups of providers and across the continuum of care, rather than fragmented among individual providers for specific services. Similar effects might be achieved through improved coordination of various funding sources for vulnerable groups covered by multiple sources (e.g., persons dually eligible for Medicare and Medicaid coverage). Aligning payments with effective health care practices and preventive services will be an important impetus for promoting evidence-based health care delivery.

Groups vulnerable due to financial or geographical issues may benefit from payment systems designed to reduce these barriers to access. For instance, Medicare currently offers bonus payments to physicians who provide services in geographic areas that have been designated as health professional shortage areas. Furthermore, as a competitive health care market makes paying for uncompensated care increasingly difficult, Medicare, Medicaid, and private plan payment approaches should support safety-net providers that provide a disproportionate share of uncompensated care. Vulnerability associated with communication problems may be addressed through payment, in part, by providing reimbursement for translation services, ombuds programs, or other assistance needed to overcome barriers.

Further experimentation with performance-based payment mechanisms can reward health plans and providers that demonstrate delivering high quality care. For example, the Pacific Business Group on Health withholds 2 percent of HMO plan payments, with the release of the money dependent on the plan reporting data on selected quality indicators and meeting performance targets in areas such as customer service and rates of various preventive services (e.g., mammography screening and childhood immunization)(Schauffler, 1996). Similarly, some health plans link physician compensation to quality performance. Performance-based payment systems could explicitly incorporate measures of performance for vulnerable populations to encourage attention to these groups in improvement efforts.

REFERENCES

Advisory Commission on Consumer Protection and Quality in the Health Care Industry, Consumer Bill of Rights and Responsibilities, November 1997.

American Association of Health Plans, "Health Plans and the Chronically Ill: Meeting Members' Needs," January 21, 1998.

Ball, J.K., and A. Elixhauser, "Treatment Differences between Blacks and Whites with Colorectal Cancer," Medical Care 34(9);970-984, September 1996.

Billings, John, Geoffrey M. Anderson, and Laurie S. Newman, "Recent Findings on Preventable Hospitalizations," Health Affairs 15(3):239-249, Fall 1996.

Carlisle, David M., Barbara D. Leake, and Martin F. Shapiro, "Racial and Ethnic Differences in the Use of Invasive Cardiac Procedures among Cardiac Patients in Los Angeles County, 1986 through 1988," American Journal of Public Health 85(3):352-356, March 1995.

Dreyfus, Tony, and Richard Kronick, "Promoting Quality through Health-Based Payment," paper presented to the Subcommittee on a Quality Improvement Environment of the Advisory Commission on Consumer Protection and Quality in the Health Care Industry, December 1997.

Ellis, Randy P. et al., "Diagnosis-Based Risk Adjustment for Medicare Capitation Payments," Health Care Financing Review, 1996:17(3).

Ford, Earl S. and Richard S. Cooper, "Racial/Ethnic Differences in Health Care Utilization of Cardiovascular Procedures: a Review of the Evidence," Health Services Research 30(1 part 2):237-252, April 1995.

Hellinger, Fred J., "The Use of Health Services by Women with HIV Infection," Health Services Research 28(5):543-561, December 1993.

Institute of Medicine, Paying Attention to Children in a Changing Health Care System, (Washington, DC: National Academy Press, 1996).

Kronick, Richard et al., "Diagnostic Risk Adjustment for Medicaid: the Disability Payment System." Health Care Financing Review, 1996:17(3).

Lambert David and Mark S. Agger, "Access of Rural AFDC Medicaid Beneficiaries to Mental Health Services," Health Care Financing Review 17(1):133-145.

McGlynn, Elizabeth A., "Quality Monitoring for Vulnerable Populations," paper presented to the Subcommittee on Quality Measurement of the Advisory Commission on Consumer Protection and Quality in the Health Care Industry, November 7, 1997.

National Alliance for the Mentally Ill, Stand and Deliver: Action Call to a Failing Industry, 1997.

Nelson, Lyle, Randall Brown, Marsha Gold et al., "Access to Care in Medicare HMOs, 1996," Health Affairs 16(2):148-156, March/April 1997.

Newacheck, Paul W., Dana C. Hughes, and J.J. Stoddard, "Children's Access to Primary Care: Differences by Race, Income, and Insurance Status," Pediatrics 97(1):26-32, January 1996.

Physician Payment Review Commission, "Implementing Risk Adjustment in the Medicare Program," Annual Report to Congress 1997 (Washington, D.C.: 1997).

Physician Payment Review Commission, Monitoring Access of Medicare Beneficiaries, No. 95- 1 (Washington, D.C.: 1995).

Schauffler, Helen Halpin and Tracy Rodriguez, "Exercising Purchasing Power for Preventive Care," Health Affairs, 15(1):73-85, 1996.

Soumerai Stephen B., Jerome Avorn, Dennis Ross-Degnan, Gortmaker S., "Payment Restrictions for Prescription Drugs under Medicaid: Effects on Therapy, Cost and Equity," New England Journal of Medicine 317(9):550-556, 1987.

Wadhwa, Sandeep, "Models for Effective Delivery of Care," paper presented to the Subcommittee on a Quality Improvement Environment of the Advisory Commission on Consumer Protection and Quality in the Health Care Industry, December 8, 1997.

Wagner, E.H., B.T. Austin, and M. Von Korff, "Organizing Care for Patients with Chronic Illness," Milbank Quarterly 74(4):511-544, 1996.

Weiner, Jonathan P. et al., "Risk-Adjusted Medicare Capitation Rates Using Ambulatory and Inpatient Diagnoses," Health Care Financing Review, 1996:17(3).

White House Office of the Press Secretary, transcript of February 21, 1998 press briefing by Secretary of Health and Human Services Donna Shalala and Surgeon General David Satcher.

(1)This review focused on terminally ill individuals, children with special needs, people with mental illness, and the frail elderly.

(2)For some individuals with severe mental illness, community-based care was not appropriate and inpatient care remains necessary.