rganizations involved in the delivery of medical care to the poor include teaching hospitals and academic medical centers, community hospitals, community health centers, physician groups, clinics, health plans with Medicaid products, and public health agencies. In some communities, individual practitioners play a significant role as well. Our discussion concentrates on the institutional providers where the bulk of care for the poor.
Many factors contribute to the likelihood that particular health care organizations will become involved in care for the poor. The size of the poor population, proportion of the population that is uninsured, the location of concentrations of poverty, and certain characteristics of the poor population (e.g., language and immigration status) are factors. Public policy, financing and underlying attitudes toward the poor also play important roles.4,5
The poor include both the publicly insured poor, as well as those who are uninsured. The publicly insured poor are covered by Medicaid and include primarily women, children, and disabled persons. Some states operate other public insurance or provider subsidy programs for medically indigent persons. The uninsured poor include the very poor who are not eligible for Medicaid - undocumented persons and poor single adults - and the working poor who make too much money to qualify for Medicaid but do not have employer-sponsored health insurance. Changes in Medicaid programs, and government policies designed to cover uninsured persons or to shift employer coverage incentives, can change the extent of coverage among the poor population. Changes in employment patterns and employer benefit practices also affect the extent of uninsurance. These circumstances fluctuate, and the precise size of the uninsured population at any given moment is, at best, estimated. The extent of uninsurance in a population, however, transltes directly into the burden of uncompensated care borne by some set of health care organizations in each market studied. Any factors that change the size or other characteristics of that burden influence delivery of care to the poor.
Table 1 shows the poverty and uninsurance rates for the communities studied. Miami, Phoenix and Little Rock had the highest uninsurance rates, and Syracuse, Lansing and Seattle had the lowest. Miami, Little Rock, Phoenix and Syracuse had the highest poverty rates, and Indianapolis, Lansing, Newark and Seattle had the lowest.
Poverty is distributed differentially in each of the markets studied. These metropolitan statistical areas comprise many smaller geographic and ethnic communities, and often sub-markets for health care. Poverty is typically concentrated in one or more of these component communities.
In all of the study sites, there is some pattern of geographic concentration of poverty. In general, the poor live in neighborhoods residentially segregated by property values, household income, ethnicity or race. The specific pattern of segregation varies among the study sites, but the net effect is to isolate the poor from the non-poor in areas with comparatively fewer job opportunities, social and economic resources. In half the sites, the poor tend to be concentrated in one or two central cities in the metropolitan area: Miami, Cleveland, Boston, Indianapolis, Newark and Seattle. In the other sites, the poor are concentrated in central cities and dispersed to varying degrees in pockets in surrounding suburban or rural areas.
In the 12 sites studied, organization of the delivery system for the poor tends to fall into three categories reflecting the distribution of poverty, local norms and public policy. Most commonly, care for the poor (both uninsured and those covered by Medicaid) was clearly concentrated in a single hospital - usually a public hospital, but occasionally a locally owned, not-for-profit institution - along with one or more community health centers and, in a few cases, a very small number of locally owned mission-oriented hospitals. Six sites fit this pattern: Indianapolis, Little Rock, Cleveland, Newark, Miami and Greenville. In four sites, care for the poor and uninsured was much more widely shared. Seattle, Syracuse, Lansing and Boston share a history of collaboration among public and not-for-profit hospitals to meet the needs of the poor. They have low rates of uninsurance, low proverty rates and in two cases long-standing state reimbursement pools for the uninsured. In the remaining two sites, we observed hybrid of both variants, differentiated by insurance coverage. In Orange County and Phoenix, the poor resided in pockets throughout the metropolitan areas, and well-established mandatory Medicaid managed care programs had broadly dispersed care for those eligible. However, care for the uninsured remained concentrated in a public teaching hospital and community health centers.
Perceptions of the major forces that shape care for the poor contribute to public policy and the strategies of health care organizations. In the sites we studied, care for the poor rarely ranked at the top of concerns about health care expressed by health care and community leaders, policymaker, and other informants. In the majority of sites, indigent care was simply not a widely shared concern. There are several explanations for this. In Seattle, Syracuse, Indianapolis and Miami, the providers most involved in care for the poor were well-regarded and generally deemed adequate to meet the needs. In Orange County, Phoenix, Little Rock and Indianapolis, vocal anti-tax sentiment underlay the lack of priority given these issues. Care for the poor was most evidently a broad concern in the Newark area, where reductions in state financing combined with high need and clear fiscal distress of inner city hospitals. But three sites with highly-regarded delivery systems for the poor also expressed concern over the fut re of these systems. In Cleveland and Lansing, concern centered on whether commitments to the poor would be continued after long-standing not-for-profit safety net providers merged with other entities. In Boston, concern for indigent care has a long history as a social priority. As discussed later, fears related to federal welfare reform surfaced in most sites, and concerns about effects of immigration reform were prominent in communities with large foreign-born populations.
Medicaid is the major source of coverage for poor populations in all the study sites, and a major revenue source to providers serving the poor. Medicaid programs vary in the populations and services they cover (beyond federal mandatory requirements) and in the way providers are paid. For those reasons, Medicaid is a dominant force shaping care for the poor and the behavior of health care organizations serving the poor. Three major changes in Medicaid were reported to influence market dynamics: managed care programs, changes in reimbursing providers and disproportionate share hospital payments.
States have been implementing Medicaid managed care programs rapidly to help bring their Medicaid expenditures under control. Several states were increasing their enrollments in managed care within the framework of voluntary programs in which beneficiaries may choose a primary care case manager as an alternative to a health maintenance organization (HMO). Mandatory Medicaid managed care programs for Aid to Families with Dependent Children recipients were in place in several sites, and new mandatory programs were being phased in in others.
Experience with these programs was mixed. Rapid implementation of Medicaid managed care programs caused problems in several sites. Respondents described instances of inadequate preparation for both patients and providers in Orange County, Seattle and Indianapolis. Health care providers with little experience with managed care contracting reported that developing the required administrative capacities was a challenge. These concerns about early implementation may have contributed to a perception that managed care was more prevalent and influential in the market than its actual market penetration warranted.
But in numerous markets, the introduction of managed care has clearly offered Medicaid patients new choices, including access to more hospitals and physicians. This was widely acknowledged in Orange County, Phoenix, Seattle and Lansing.
A more overarching concern expressed by a number of traditional providers of care for the poor was that this increased choice and access to other providers triggered deterioration in their patient base. Community health centers in Orange County, Miami, Syracuse, Indianapolis and Newark, and the public hospital in Phoenix particularly noted this effect. Providers whose mission revolves around care for the poor said that the loss of part of their publicly insured population, as Medicaid patient care was redistributed across a broader base of providers, would leave them with few sources of funding to provide care for the uninsured.
With or without managed care, Medicaid payment rates influence the willingness of providers to serve Medicaid beneficiaries. Sites in states with low payment rates reported longstanding difficulty getting providers to participate in their programs, particularly private physicians. This was evident, for example, in Greenville and Newark. When managed care organizations with Medicaid contracts offered providers capitation or higher rates than were previously available from the state, new providers were attracted and access was increased. Others felt capitation schemes contributed to competition for the lower cost Medicaid patients, but left the higher cost patients with the traditional providers of last resort. Providers that had benefited from prior special payment rates did not typically see managed care payment as an improvement. Community health centers, in particular, often reported managed care rates were inadequate compared to the previous cost reimbursement they received as federally qualified health centers.
Fears of future Medicaid rate reductions were pervasive, often rooted in past experience. In Indiana, for example, the state decreased payments to hospitals and physicians in 1994, leading to reports in the Indianapolis market of the unwillingness of certain providers to serve Medicaid patients. Similarly, after initially positive provider reactions to Washington State’s Medicaid managed care program, subsequent rate reductions were beginning to discourage provider participation in Seattle. Concerns about possible Medicaid rate reductions were frequently cited in Little Rock, Miami and Syracuse. Medicaid programs also may change the determination of which providers are eligible for reimbursement. In Orange County, for example, a community health center that had historically provided family planning services under direct contract with the state found itself losing volume when the State Office of Family Planning permitted any licensed Medicaid provider to receive reimbursement for such services.
The net effect of these changes - recent and anticipated - was general insecurity among Medicaid providers about both the level and form of Medicaid payments, but continued efforts to consolidate and expand their market share for this population.
In many sites Medicaid disproportionate share hospital (DSH) payments, and related state bad debt and charity care pools were critical supports, not only to hospital Medicaid revenues but also as cross-subsidies or direct reimbursement for the uninsured. These are essentially redistributive financing mechanisms that rely on provider contributions, provider or payer taxes, intergovernmental funds transfers, usually matched by additional federal Medicaid dollars, and theoretically targeted to providers delivering high levels of care for the poor and/or uninsured. Threats to these subsidies were cited in nearly all sites: federal legislation limiting these financing mechanisms; state-level legislative and judicial challenges to some variants of them; and more competitive environments.
In both Orange County and Phoenix, longstanding disproportionate share providers reported they were receiving less of such funding than in the past. In Syracuse, hospitals feared reduced funding under the restructuring of the state’s traditional bad debt and charity care pool, though initial experiences following the state change did not bear this out. In Newark, one hospital declared bankruptcy; and the financial condition of inner city hospitals worsened after judicial rejection of the state’s uncompensated care program. In Boston, challenges were mounting to the state’s free care pool, which has long supported indigent care costs in hospitals; concerns included allocation formulas and the offsetting effect of health plan discounts on pool payments to hospitals.
Several states and localities had implemented financing programs for the uninsured in the markets studied. Such programs provide new resources to providers carrying the burden of uncompensated care and may attract other providers to care for the poor in a manner analogous to Medicaid. These programs involve both insurance mechanisms and direct subsidies to providers.
State programs in six sites expanded services to the uninsured. As noted, state indigent care pools were significant sources of financing to hospitals in Syracuse, Boston and Newark. New Jersey’s program was overturned by the courts as a violation of federal Employee Retirement Income Security Act laws, resulting in a dramatic decline in funding from $700 million (1992) to $300 million (1997). New Jersey subsequently redirected state unemployment insurance fund dollars to keep the program going, and devoted special funding to hospitals serving large numbers of patients with AIDS, tuberculosis, mental illness, substance abuse and complex births. The state also explored a managed care approach to charity care. In Seattle, the state’s Basic Health Plan subsidizes insurance for the working poor. One of the few survivors of Washington State’s repeal of its comprehensive health reform act, this highly regarded program was experiencing some adverse selection of the sicker uninsured. Arizona’s state-only progra for the indigent, administered by its Arizona Health Cost Containment System (AHCCCS) Medicaid managed care program, covers part of the population just above Medicaid eligibility in Phoenix. In Orange County, the state-county Medical Services for the Indigent (MSI) program has provided limited funding to some hospitals and physicians. Moving administration of the MSI program to the county’s Medicaid contracting entity, CalOPTIMA, has been slowed by concern funding is inadequate to serve even half of the uninsured.
Two other approaches deserve mention. Dedicated county tax assessments support the Public Health Trust in Miami (Jackson Memorial Hospital and affiliated programs), and Indianapolis Health and Hospitals is a taxing district. While these are institutional subsidies rather than explicit resources for the uninsured, they are cited as supporting these institutions’ dominant roles as indigent care providers. County tax dollars also provide essential supports for Maricopa County Hospital in Phoenix. Finally, we encountered contractual provisions in two sites associated with the sale or partnership of a traditional safety net hospital. Provisions in the partnership agreements of three Cleveland hospitals with national for-profit companies, and of a Lansing public hospital merged into a community not-for-profit hospital, required continuation of traditional distribution of responsibility for indigent care.
In contrast to these efforts to expand coverage for the uninsured, concern was widespread that welfare reform would swell the uninsured population and increase burdens on traditional providers of care for the poor. In Orange County, Phoenix, Miami and Newark, it was expected that immigrants would be denied health benefits previously available to them or the institutions serving them. A further concern about growth of the uninsured was a decline in employer coverage of dependent health benefits, noted in Orange County and Newark. A number of respondents indicated increases in the uninsured would confront them with the choice of whether to turn away uninsured patients, and, in fact, providers in several sites described reductions in access they and other providers had implemented.
In nearly every market studied, the phenomenon of hospital and health system consolidation confronted the organizations involved in care for the poor, forcing them to revisit their historic referral relationships and affiliations. This impulse to organizational change is driven by changing payment mechanisms, desire to expand or consolidate market share, and fear of being "left out" as other organizations around them consolidate.
Most markets exhibited increasing diversity in payment arrangements, with the introduction of risk-bearing arrangements by both public and commercial payors becoming more prevalent. Traditional providers of care for the poor expressed concern about their ability to negotiate and administer the new organizational and financial arrangements required, and to cope with the attendant drain on scarce administrative resources. Creation of networks for managed care contracting by health plans and health systems in nearly every market presented traditional providers of care for the poor with difficult decisions about potential alliances with these networks or formation of their own networks.
Many of the major hospital providers caring for the poor were reorganizing to bolster their chances of survival in a competitive market, including mergers with other healthcare organizations. Mergers involving important urban hospitals serving the poor occurred in Boston and Newark. The dominant safety net hospitals in Orange County and Miami engaged in exploratory discussions with national for-profit hospital companies, but later abandoned them in the wake of investigations into Columbia/HCA billing practices. Three traditional hospitals for the poor in Cleveland entered partnerships with for-profit companies, Columbia HCA and Primary Health Systems. Two prominent attempts to convert important safety net hospitals failed: Maricopa County, Ariz., rejected a proposed privatization plan that would have leased a county hospital to a not-for-profit organization, and the attempt of not-for-profit Michigan Capital Health Care in Lansing to partner with Columbia/CA also failed. In Indianapolis, the county hosp tal was left out of the merger between Indiana University and Methodist Hospitals.
Like other providers, safety net providers were establishing a wide variety of alliances, networks and contractual relationships with each other and with other organizations in the local market. In Cleveland and Miami, strong public hospitals assumed responsibility for operation of local health department clinics. In Seattle, Boston and Miami, community health centers (CHCs) formed networks and joined new health plans controlled locally or regionally by the centers. In Cleveland, the public hospital entered an extensive managed care contracting network with the Cleveland Clinic and other organizations, and assumed management of emergency medicine for two other systems. Yet all these fledgling alliances notwithstanding, competition among health care organizations involved with care for the poor, including partners in several of these alliances, was often aggressive. Several respondents cited declines in historic collaborations in Syracuse, Indianapolis, Seattle, Lansing and Boston.
Respondents worried that constant change in networks, organizational arrangements and alliances would increasingly disrupt established referral patterns and take time and resources away from a focus on delivery of coordinated care. However, we saw little hard evidence that organizational change was as yet affecting access or delivery of care. The exceptions to this were the creation of primary care links and referral patterns to implement mandatory managed care for Medicaid beneficiaries, where some benefit was widely perceived.
4 Solomon, L. S. "The impact of public policy on health system change." Unpublished manuscript. 1998. (Citation record may require revision upon publication.)
5 Steinberg, C. R., and R. J. Baxter. "Local community norms and values as a force in health system change." Unpublished manuscript. 1998. (Citation record may require revision upon publication.)