Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 82
A Shared Destiny: Community Effects of Uninsurance 3 Community Effects on Access to Care The Committee hypothesizes that the burden of financing care for uninsured persons affects the health services available within the community, especially in urban areas where providers treat large numbers of uninsured persons and in rural areas where providers treat relatively high proportions of uninsured. The strategies taken may be different in rural and urban areas, but the Committee finds plausible the idea that providers in rural and urban settings respond to financial pressures related to uninsurance in similar ways. In an effort to avoid this burden or to minimize its impact on a provider’s bottom line, services may be cut back, relocated, or closed. Staffing may be reduced. This may further strain the capacities of already overcrowded hospital emergency departments, and physicians’ offices or even hospitals may be relocated away from areas of town or entire communities that have concentrations of uninsured persons. Such disruptions may reduce people’s access to care and the quality of care they receive, regardless of their insurance status. In this chapter, the Committee explores the consequences of community uninsurance for access to care in the context of the present era of rapid change in public and private health care financing. As discussed in Chapter 1, the Committee has chosen to interpret its task as concerning community effects of uninsurance but not of underinsurance, despite the difficulties of disentangling the likely mutual influences of uninsurance and underinsurance and the limits of the existing literature. After a brief discussion of access to care considered in general, the chapter is organized by type of service, examining primary care, emergency medical services and other specialty care, and hospital-based care in turn. The chapter concludes with a discussion of questions for further research and a summary. Although many of the data and studies are organized by institution or profes-
OCR for page 83
A Shared Destiny: Community Effects of Uninsurance sion, the Committee takes the perspective of community residents as consumers of health care and reports on local health services arrangements as typically encountered by someone seeking care. Focusing on the types of services as they may be affected by uninsured rates, rather than on providers and institutions, allows a more systematic perspective on the adequacy of health services throughout the community (see IOM, 2001b). In addition, the mere survival of particular institutions may not be a good proxy for the availability of high-quality care (particularly primary care) or for judging the impacts of uninsurance in a rapidly changing health care system or market. Nonetheless, the value of a hospital or clinic may include contributions to a community’s social and economic base or serve as a source of community identity. Thus, changes related to uninsurance may have additional effects beyond those related to access and quality of care. These values are discussed in the next chapter, which addresses some potential social and economic effects of uninsurance. ACCESS TO CARE Finding: Persons with low to moderate incomes (less than 250 percent of the federal poverty level), nearly one-third of whom are uninsured, and uninsured persons have worse access to health care services in communities with high uninsured rates than they do in communities with lower rates. The causal influence of the local uninsured rate on measures of access is unclear. While insurance facilitates access to health care by removing or diminishing financial barriers faced by individuals and families, it also secures a revenue stream for health care providers and institutions (IOM, 2001a, 2002a, 2002b). The Committee hypothesizes that when this revenue stream is diminished by higher local uninsured rates, providers may opt to change their patterns of service provision within the community in an effort to preserve revenues. Depending on the local organization of health services, this provider phenomenon would be expected to affect the opportunities of insured residents to obtain health care. Lower-income residents are particularly likely to be affected by loss of access to care because of their limited economic ability, for example, in rural areas to travel long distances to obtain services unavailable locally or to pay higher costs for care. Three recent studies examine the relationship between self-reported access to care and community uninsurance in urban areas (Cunningham and Kemper, 1998; Brown et al., 2000; Andersen et al., 2002). All of the studies rely on survey data about self-reported access to care, using a number of different measures (e.g., difficulty obtaining care, not having seen a physician in the previous 12 months). The first study, based on the nationally representative 1996–1997 Community Tracking Study (CTS) Household Survey, examines the differences among metropolitan statistical areas (MSAs) larger than 200,000 population in the rates
OCR for page 84
A Shared Destiny: Community Effects of Uninsurance that uninsured residents reported forgoing, postponing, or having difficulty obtaining needed medical care as a function of the community uninsured rate (Cunningham and Kemper, 1998). The authors show that the rates at which uninsured residents reported difficulties in obtaining care differed more than two-fold over the range of high-uninsured-rate and low-uninsured-rate communities, from about 41 percent (Lansing, Michigan) to 19 percent (Orange County, California), with a national average of 30 percent.1 Only about 15 percent of the variation among the communities in rates of difficulty in accessing care is related to measured differences in the health needs or other population characteristics of the uninsured respondents, indicating that most of this variation is related to unmeasured factors. Concluding that an uninsured person’s place of residence plays a key role in determining ease of access to care, the authors hypothesize that differences in access may reflect the economic make-up of the uninsured person’s community (e.g., providers in wealthier areas may be better able or inclined to deliver free or reduced-price care) or the availability of health care outside of the community (e.g., proximity of multiple health care markets in the larger MSAs) (Cunningham and Kemper, 1998). Although the first of the three studies discussed focuses only on uninsured residents, it points up how their ability to obtain care depends on their relative numbers within the overall community. Further, the possibility remains that insured residents are affected by local uninsurance in ways not measured in the analysis. The authors report no relationship (correlation) across the 60 communities studied between the proportion of privately insured respondents reporting difficulty obtaining access to care and the proportion of uninsured respondents with access difficulties, as well as different reasons for difficulties, with 90 percent of uninsured respondents citing cost as the key constraint, compared with 48 percent of insured respondents. Yet the often separate experiences of uninsured and privately insured with the organization and delivery of care in urban areas (as described in Chapter 2) may be linked through the effects that uninsurance exerts on the local health care market. The second study considers the aggregate experiences of all low- to moderate-income residents (earning less than 250 percent of the federal poverty level [FPL]), under age 65 in the 85 largest MSAs in 1997, without distinguishing between insured and uninsured respondents (Brown et al., 2000). The authors examine measures of health care access including forgone or delayed care, no physician visit within the year, and having a regular source of care, as a function of uninsured rate (Brown et al., 2000, data for 1997). Residents of the 12 MSAs with 1 This analysis was unadjusted for the many individual and community-level covariates of uninsured rate. Adjusting for health care need (health status, age, gender) and socioeconomic and demographic characteristics (income, education, family size, race and ethnicity, whether Spanish language interview) narrowed the range in reported difficulty in obtaining access to care to between 36 percent for Cleveland, Ohio, and 22 percent for Newark, New Jersey.
OCR for page 85
A Shared Destiny: Community Effects of Uninsurance significantly higher uninsured rates (at least one standard deviation above the mean for all 85 MSAs) reported greater difficulty obtaining health care than did residents of the 17 MSAs with significantly lower uninsured rates, unadjusted for any covariates of unin-surance: 22 percent reported delaying or going without needed health care in MSAs with high uninsured rates, compared to 8 percent in MSAs with substantially lower uninsured rates; the comparable figures for those who reported no physician visit in the past year were 39 and 14 percent in MSAs with high and low uninsured rates, respectively. Since almost one-third of persons under age 65 who earn less than 200 percent of FPL or who are members of families that earn less than 200 percent of FPL are uninsured (Fronstin, 2002), it is likely that a proportion of the effect observed of local uninsured rate reflects the experiences of the uninsured rather than the insured study participants. The third study complicates conclusions that might be drawn from the findings of Cunningham and Kemper about what unmeasured community-level characteristics may influence the access of uninsured persons to care and the findings of Brown et al. about the relative influence of community uninsurance on insured versus uninsured residents. In a multivariate, cross-sectional study of the likelihood (odds) that low-income children and adults under age 65 have seen a physician in the past year in the 25 largest MSAs nationally, the authors show that a number of individual-level and community-level covariates may better explain differences in access to care than does the community uninsured rate (Andersen et al., 2002). While analytically adjusting for measures of community demand for care (percentage in poverty, uninsured rate, and percentage enrolled in Medicaid) did not change the odds that low-income children and adults would have seen a physician in the past year, taking into account differences in a community’s financial and structural capacity to deliver care (e.g., higher per capita income, lower unemployment, greater income inequality, number of public hospital beds per thousand population, number of community health centers per thousand population, lower degree of managed care penetration) did improve the odds of seeing a physician in the past year (Andersen et al., 2002). Echoing the hypothesis laid out by Cunningham and Kemper, the authors suggest that a possible explanation for their findings is that greater community wealth may strengthen the financial viability of safety-net arrangements, either directly or through the support of higher levels of eligibility for public coverage or higher spending on related public programs. Together, these three studies provide evidence that, with limited analytic adjustments for the many covariates of insurance status and uninsured rate, an urban area’s uninsured rate is associated with the ability of uninsured and of moderate- and lower-income residents to obtain needed and regular health care. Such measures of access to health services are basic markers of effective, quality care that improves health outcomes (Millman, 1993; IOM, 2002a, 2002b). Longitudinal studies are needed, with adjustments for covariates of insurance status and uninsured rate, to better distinguish whether and how community uninsurance
OCR for page 86
A Shared Destiny: Community Effects of Uninsurance influences the access experiences of insured and uninsured community members in different ways and to a greater or lesser extent. PRIMARY CARE Except for the studies of access to care described in the preceding section, which include measures that can be interpreted as measures of access to primary care, there is little empirical documentation of the relationship between community uninsurance and access to primary care. In this section, the Committee draws on the literature about primary care providers and sites of care, for the most part, physicians and clinics, to develop background information about potential mechanisms and outcomes of community effects. Uninsurance may compromise access for many community residents to preventive, screening, and primary care services. These services are provided in diverse settings, ranging from general primary care and specialized health department clinics (e.g., immunization, family planning), private physician offices, community health centers, hospital outpatient and emergency departments, and dedicated hospital primary care outpatient clinics (Donaldson et al., 1996). See Box 3.1 for the Committee’s working definition of primary care. Particularly for low-income residents and members of other medically underserved groups, clinics and health centers play a special role in primary health care services delivery due to their close geographical proximity to underserved populations, their cultural competence and history in the community, and their provision of supportive “wrap-around” services that facilitate access to care and quality care (Hawkins and Rosenbaum, 1998). As a result, when uninsurance results in fiscal pressures on community facilities, insured as well as uninsured clients may be affected by reductions in health center services. To make up for the provision of uncompensated care, clinics and health centers may cut back on available services, lengthen waiting times for all patients, and cease operations altogether in some communi BOX 3.1 Definition of Primary Care The Institute of Medicine Committee on the Future of Primary Care defines primary care “as the provision of integrated (that is to say, comprehensive, coordinated, and continuous over time), accessible health care services by clinicians who are accountable for addressing a large majority of personal health care needs, developing a sustained partnership with patients, and practicing in the context of family and community” (Donaldson et al., 1996). Primary care clinicians typically are physicians, physician assistants, or nurse practitioners, sometimes working together in a team. The ongoing personal relationship between provider and patient that is the ideal in primary care is an integral element of quality health care (IOM, 2001b).
OCR for page 87
A Shared Destiny: Community Effects of Uninsurance ties. In Los Angeles County in the latter half of the 1990s, for example, in a survey of private providers who contracted with the county to care for medically indigent patients, respondents reported finding themselves at financial risk, with the potential loss of the provider’s services to the entire community (Grazman and Cousineau, 2000). See Box 3.2 for case studies of uninsurance and primary care in two of the country’s largest urban areas. Shortages of physicians in rural and urban areas with relatively high uninsured rates can mean less access to primary care for all residents. Access for rural residents is particularly sensitive to reductions in available care, because residents have limited choices to begin with and are likely to face long travel times and distances to reach providers outside their community. Furthermore, rural residents, who tend to be in poorer health and older than residents of suburban and urban areas, are likely to experience greater difficulty traveling these longer distances (Ormond et al., 2000a). Physicians Physicians’ decisions about the nature and location of their practice may be influenced by the uninsured rate of the community among other factors (e.g., malpractice insurance costs, the availability of specialty support, the option to practice in large tertiary care hospitals, the safety and overall attractiveness of the community as a place to live). Considerations about where to locate a new practice are likely to differ from those regarding an ongoing practice because established practices may be better able to accommodate charity or other uncompensated care. Many rural areas and lower-income and economically depressed urban neighborhoods have a harder time recruiting and retaining primary care practitioners than do more populous and affluent communities. For communities in rural areas, uninsurance may make it more difficult for a private physician to maintain a financially viable practice. Physicians in rural areas are more likely to treat Medicaid and uninsured patients and have higher proportions of Medicare patients due to rural demographics (Mueller et al., 1999). As a result, rural physicians may face lower levels of reimbursement. This can place further financial drains on their practices (Coburn, 2002). At the same time, in such areas, safety net arrangements for primary care tend to be informal and not publicly subsidized, especially in the smallest towns (Taylor et al., 2001). A convenience survey of informal safety net arrangements in eight towns across the country (less than 5,000 population) finds that rural hospitals are particularly important in this regard (Taylor et al., 2001). Many rural physicians work in affiliation with or are supported by their community hospital, through either a rural health clinic or the hospital itself. For example, a family practice in Deer Island, Maine, a fishing community in which 60 percent of the residents were uninsured, alleviated its financial problems by affiliating with a local hospital as a rural health clinic (Finger, 1999). Physicians who serve uninsured patients encounter ethical and legal dilemmas
OCR for page 88
A Shared Destiny: Community Effects of Uninsurance BOX 3.2 Effects on Primary Care: Los Angeles and New York City Los Angeles County, California Los Angeles County is home to 10 million people, roughly one-quarter of whom (2.5 million persons) are uninsured. Financial pressures on the county health department related to the care of medically indigent and uninsured persons are leading to a reduction in primary care and preventive services and the closing of sites of care for all county residents. The county health department serves an estimated 800,000 patients each year, many of whom are uninsured. Broad cuts in services were averted during a county budget crisis in the mid-1990s through an increase in federal Medicaid payments for outpatient services (totaling more than $2 billion) under a five-year program waiver from the federal government. This waiver was renewed once and is scheduled to expire in 2005. Under the waiver, county health services were reconfigured to emphasize primary care and preventive services, which improved access to care but did not reduce hospital inpatient utilization significantly or result in appreciable cost savings. Recent years’ state Medicaid budget troubles are shared by Los Angeles County. Despite the county’s Board of Supervisors’ efforts to balance the department’s $2.9 billion dollar budget and voter support of a property tax increase earmarked to support hospital trauma centers ($168 million expected annually), a $700 million to $800 million deficit is anticipated in 2003. To reduce this deficit, the county has made cuts in primary care rather than in specialty services, hospital inpatient services, or emergency medical services. According to one news account, decision makers justified this choice by explaining that primary care services are more affordable for the clinics’ clients than specialty services, even though one of the anticipated consequences is increased emergency department (ED) use for non-urgent, primary care. Like hospital EDs elsewhere, those operated by public hospitals in Los Angeles are precluded by the Emergency Medical Treatment and Labor Act (EMTALA) from limiting the use of their services to persons with health insurance. In June 2002, the county closed 11 of its 18 public health clinics and 1 of its 6 public hospitals. It also scaled back support for 100 private clinics and cut 18 percent of employees (5,000 jobs). This was done as part of efforts to trim the deficit in preparation for negotiations for additional federal or state subsidies. Further downsizing and closings are anticipated to be announced in the coming months.1 New York City New York’s municipal Health and Hospitals Corporation (HHC) is the largest public hospital system in the country, overseeing 11 acute care public hospitals and more than 100 local clinics among its varied responsibilities to provide health care to medically underserved groups. HHC serves approximately 1.3 million per-
OCR for page 89
A Shared Destiny: Community Effects of Uninsurance sons a year, 43 percent of whom are uninsured. About half of all hospital-based care is delivered to uninsured and underinsured residents, and HHC facilities provide a disproportionate share of services more likely to be used by vulnerable populations (e.g., tuberculosis, HIV/AIDs, psychiatric, trauma, alcohol and substance abuse, emergency medical services). Since 1994, HHC has responded to several financial challenges by cutting back funds and shifting patients from hospitals to primary care clinics. During 2001, HHC experienced its first deficit in five years. News accounts attribute part of the $300 million deficit to a 30 percent increase since 1996 in the number of uninsured persons (totaling more than 564,000 uninsured patients in 2000). Other factors include changes in patient mix and volume (e.g., loss of market share to other hospitals), competition for Medicaid patients to fill inpatient beds (given declining inpatient utilization) from private nonprofit hospitals, and a growing gap between uncompensated care and public subsidies (e.g., indigent care dollars from the state, Medicare disproportionate share hospital payments, and appropriated city tax revenue). Uninsured patients account for 10 percent of hospitalized patients and 30 percent of ambulatory care patients in HHC facilities. As Medicaid payments lag behind health care cost inflation, especially for clinic visits, and private sector facilities compete successfully for Medicaid managed care enrollees, Medicaid revenues to HHC facilities have declined. Because patients have been shifted from hospitals to primary care clinics, there has been a 9 percent increase in clinic visits and a 12.5 percent decrease in inpatient admissions since 1992. Because Medicaid managed care payments to primary care clinics are lower than those for hospital care or for specialty services, HHC facilities have been squeezed financially. Despite the known benefits of primary care, HHC has closed and merged preventive and primary health care sites over the past year, including eight neighborhood health centers. After receiving additional funds from the city, HHC decided to keep open 15 school-based clinics slated for closure, although their operations will result in an anticipated $2.7 million loss in 2001. For the most part, the clinics selected for closing are located in immigrant and low-income neighborhoods where roughly two-thirds of the patients are uninsured. Clinics to be closed include 9 of 42 that specialize in children’s health. Overall, the loss of capacity is estimated as 100,000 children served annually. Other cost-control measures include cutting staff and downsizing capacity—for example, decreasing the capacity of the renovated Queens Hospital Center from 300 to 200 beds. These changes have resulted in overcrowding and longer waits at public hospital emergency departments and waiting areas, longer waiting times to get appointments, higher fees for pharmacy services, and lower quality of care. 2 1 SOURCES: Cornwell et al., 1996; Grazman and Cousineau, 2000; Andersen et al., 2002; Cohn, 2002; Larrubia, 2002; Lutzky and Zuckerman, 2002; Riccardi, 2002a, 2002b; Riccardi et al., 2002; Rundle, 2002. 2 SOURCES: Siegel, 1996; Prinz et al., 2000; Ramirez, 2001; Sengupta, 2001; Steinhauer, 2001a, 2001b; United Hospital Fund, 2002.
OCR for page 90
A Shared Destiny: Community Effects of Uninsurance that physicians who do not see the uninsured (either by choice or by practice location) can avoid. Anticipating little or no reimbursement for their services, they may have to decide between reducing the amount or nature of the services they provide to patients from whom they do not expect any compensation or seeing fewer uninsured patients in their practices (Weiner, 2001). The decision to limit service for uninsured persons may preserve the financial basis for individual or group practice, but it presents an ethical or moral conflict for many physicians (Weiner, 2001). Over the 1990s, the proportion of physicians who provided charity care (whether to uninsured or insured persons) declined. The CTS results presented in Chapter 2 show a decline in the proportion of physicians reporting giving charity care from 76.3 to 71.5 percent between 1997 and 2001 (Reed et al., 2001; Cunningham, 2002). The proportion of physicians devoting more than 5 percent of their clinical practice time to charity care also decreased over this period from 33.5 to 29.8 percent, while the proportion devoting less than 5 percent of their practice time to charity care increased from 66.5 to 70.2 percent (Cunningham, 2002). This change is thought to be related to health care market changes, including the shift of physicians from solo and group ownership of practices to becoming employees of larger health services institutions. It may also be related to the financial strain on physician practices and increased time pressures. According to the results of a 2000 CTS survey, even at academic health centers (which serve a safety-net function in many urban areas), 63 percent of affiliated medical school faculty surveyed report that they are discouraged from seeing uninsured patients, by either inadequate payment, their teaching hospital, or their practice group (Weissman et al., 2002). Physicians who provide charity care also have concerns about malpractice liability that may affect their provision of such care. A random sample survey of California generalist physicians (n = 124) finds that these practitioners’ decisions about accepting new uninsured patients were influenced by the concern that they would be more likely to be sued (49 percent) (Komaromy et al., 1995). This study reported that just 43 percent of those surveyed accepted new uninsured patients compared with 77 percent who accepted new privately insured patients. Community Health Centers and Clinics Finding: Serving a high or increasing number of uninsured persons reduces a community health center’s capacity to provide ambulatory care to all of its clients, insured as well as uninsured. By mission, community health centers (CHCs) are located in areas with limited access to primary care (Davis et al., 1999). For uninsured persons, there is evidence that CHCs provide better access to primary health care, as measured by having a regular source of care and having a physician visit in the previous year, and higher-quality care, as measured by patient satisfaction and meeting Healthy
OCR for page 91
A Shared Destiny: Community Effects of Uninsurance People 2000 goals for health promotion, prevention, and screening than other ambulatory care providers (Carlson et al., 2001). Expanding the number of CHCs or their capacity to provide services is likely to improve access to primary care for medically underserved groups (Forrest and Whelan, 2000). One survey of primary care access at four types of ambulatory care sites (public facilities, private nonprofits, federally qualified health centers, and other freestanding health centers) in New York City in 1997 found that public sites have the strongest performance for low-income patients in terms of coverage for uninsured persons, services available after business hours, and the availability of wraparound services.2 This study concluded that serving increasing numbers of uninsured persons, in the context of fiscal pressures from Medicaid managed care and fewer Medicaid patients overall, may weaken ambulatory sites’ capacity to provide primary care to all members of the community (Weiss et al., 2001). The loss of ambulatory care clinical services would be particularly damaging to vulnerable populations because these clinics serve higher proportions of low-income and minority group members under age 65 and have safety net missions (e.g., serving farmworkers, homeless persons, or undocumented immigrants) or patients with specific high-risk diagnoses (e.g., tuberculosis, sexually transmitted diseases, substance abuse). CHCs face financial pressures on a number of fronts: from reduced income from Medicaid patients, related to the growth of state Medicaid managed care contracting; from reduced public subsidies; and from an increased proportion of their patient mix that is uninsured (Lewin and Altman, 2000). As discussed in Chapter 2, since the 1980s, Medicaid revenues have become more central to supporting clinics and supporting, indirectly, primary care for uninsured persons. During the 1990s, however, the rapid growth of state Medicaid managed care contracting destabilized this revenue stream for CHCs. Between 1980 and 2000, inflation-adjusted federal grant support for CHCs actually dropped 30 percent, even as the number of centers increased 22 percent and the number of uninsured patients served by CHCs grew by 54 percent (Markus et al., 2002). As a result, changes in the Medicaid program both influence uninsured rates and affect the capacity of CHCs to serve all of the members of their target community. For example, when a CHC is not included in a local Medicaid managed care network or when substantial discounts are negotiated as a part of its contract with a managed care network, the CHC’s scope of operations may be compromised (Cunningham, 1996; IOM, 2000a). 2 As part of their mission, CHCs offer wraparound or enabling services that facilitate access to care for members of medically underserved groups. These services include transportation, translation, social services and assistance in enrolling in public assistance and other programs, community outreach and health education and promotion, and environmental health; commonly used services that are easier to access at one site (e.g., pharmacy, radiology, and laboratory services); and specialty services targeting the special needs of the client population (e.g., nutrition, prenatal care, dedicated services for HIV/ AIDS, substance abuse treatment, and mental health services) (Davis et al., 1999; USGAO, 2000).
OCR for page 92
A Shared Destiny: Community Effects of Uninsurance The increasing proportion of CHC clients who are uninsured may reflect not only greater numbers of uninsured persons but also the practice, becoming more common, of health care providers’ referring uninsured patients to CHCs rather than accepting such patients for treatment themselves (Hawkins and Rosenbaum, 1998). This phenomenon is akin to “patient dumping,” the transfer of medically indigent patients from private hospitals to public facilities. The result is likely to be the concentration of uninsured patients in safety-net institutions, including CHCs. Over the long term, the increasing reliance of CHCs on Medicaid revenues, often through managed care contracting, has adversely affected CHCs’ capacity to serve all of their patients, insured as well as uninsured (Shi et al., 2001; Rosenbaum et al., 2002). A survey of CHCs between 1996 and 2000 found that participation in Medicaid managed care contracts was associated with financial losses for CHCs. Still, health centers’ participation in managed care, particularly for Medicaid enrollees, increased more than 50 percent over this four-year period (Rosenbaum et al., 2002). As a result, federal operating grants awarded to CHCs to support care for uninsured persons actually may be compensating the centers for losses incurred from treating insured patients enrolled in Medicaid managed care programs. The financial difficulties experienced by community health centers may result in narrowing the scope of services offered at these sites and a decrease in the center’s capacity to deliver primary care, including wraparound services that facilitate access to care, for all members of a center’s target population (Hawkins and Rosenbaum, 1998). Because wraparound services are, for the most part, not reimbursed by Medicaid or other third-party payers, federal grants support these services (McAlearney, 2002). In light of this fact, any increase in the number and proportion of uninsured users at a health center may create financial strains that result in the reduction or elimination of one or more of these enabling services. This is likely to decrease both access to care and quality and continuity of care (USGAO, 2000). The degree to which CHCs have lost capacity to provide primary care is not uniform nationally. A longitudinal study of 588 CHCs between 1996 and 1999 found that reductions in capacity appear to be concentrated among subsets of health centers with a sizable number or share of uninsured patients or those that have experienced a recent increase in uninsured patients (McAlearney, 2002). Nationally, one-fifth of federally funded CHCs (130 of 700) have a high proportion of uninsured clients (62 percent or higher, rather than the national average of 41 percent). More than half are located in the South, West, or other areas with high uninsured rates (Markus et al., 2002). EMERGENCY MEDICAL SERVICES AND TRAUMA CARE Finding: Although hospital emergency department (ED) overcrowding is not primarily a consequence of uninsurance within a community, rising uninsured rates can worsen ED overcrowding and the
OCR for page 109
A Shared Destiny: Community Effects of Uninsurance Rural (County)2 Weighted and unadjusted for concentration Weighted and adjusted/not concentratedd Coeff. Change associated with % uninsured discharges 1 S.D. above the meanb (%) Coeff Change associated with with % uninsured discharges 1 S.D. above the meanb (%) –0.45 –11.1 –0.54 –13.3 –0.64 –15.8 –1.38 –25.6 –0.96 –66.0 Not studied Not studied –1.67 –34.5 –1.99 –41.1 –1.41 –40.0 −0.91 –54.1 –1.04 –61.8 –1.27 –30.2 −0.17 −37.4 –0.22 −48.4 a proportion of hospital beds in the MSA (a measure of availability). For the rural analysis, inpatient capacity is weighted by average county population and services by the percentage of hospitals offering the services. dAdjusting for the relative concentration of uninsured patient discharges, or the relative degree of clustering of uninsured patients at hospitals within the area, as measured by the ratio of the Herfindahl index for uninsured discharges over the Herfindahl index for all discharges. SOURCES: 1Gaskin and Needleman, 2002; 2Needleman and Gaskin, 2002.
OCR for page 110
A Shared Destiny: Community Effects of Uninsurance the case of psychiatric inpatient services, this association translates into a decline in the proportion of hospitals offering this service from 39.6 percent (the mean across all MSA observations) to 33.8 percent. Adjusting the analysis by hospital size, a 5.3 percentage-point higher uninsured rate is associated with between a 9.8 percentage-point and 14.6 percentage-point lower proportion of psychiatric beds in hospitals offering these services and an 18.4 percentage-point lower proportion of alcohol and chemical dependence beds in hospitals that offer that service. Compared with an urban area with the “average” uninsured rate, a 5.3 percentage-point higher uninsured rate is associated with a decline from 37.8 percent to 30.9 percent of alcohol and chemical dependence beds in hospitals offering this service. For high-technology services, a 5.3 percentage-point higher uninsured rate is associated with a 14.7 percentage-point lower proportion of hospitals offering trauma services, a 17.2 percentage-point lower proportion offering burn services, and an 8.4 percentage-point lower proportion offering tomography (SPECT). These changes translate into a decline from 22.3 to 19.0 percent of hospitals offering trauma services, from 5.2 to 4.3 percent of hospitals offering burn services, and from 41.2 to 37.7 percent of hospitals offering SPECT. Adjusting the analysis for hospital size, a 5.3 percentage-point higher uninsured rate is associated with an 8.7 percentage-point lower proportion of trauma beds in hospitals offering the service, a 7.5 percentage-point higher proportion of neonatal intensive care, or NICU, beds and a 4 percentage-point higher proportion of angioplasty beds in hospitals offering these services. The case of greater NICU availability may reflect the greater relative need for neonatal intensive care in communities with higher uninsured rates, the fact that this need is usually accompanied by eligibility for public coverage, and the more extensive coverage of pregnant women and newborns by public insurance programs, compared with public coverage of the population overall (Howell, 2001; IOM, 2002b). The finding of greater availability of angioplasty services in higher uninsured rate urban areas may reflect regional differences in practice patterns (e.g., the procedure is more common in the South and West, which also have higher uninsured rates) that are unmeasured and unadjusted for in the analyses (Pilote et al., 1995; Krumholz et al., forthcoming 2003). Rural Areas The associations between uninsurance in rural counties (measured as the percent of all hospital discharges in the county that are uninsured) and measures of beds, services, and financial margins follow a pattern similar to that seen in the urban area analysis. In addition, the rural analysis includes an adjustment for the degree to which uninsured patients were clustered at particular hospitals within a given county. The concentration of patients may reflect competitive strategic interactions among hospitals. For example, a hospital may eliminate a service it cannot support with its patient and payer mix. This could allow a competing
OCR for page 111
A Shared Destiny: Community Effects of Uninsurance hospital to add to its caseload while lowering per-patient costs and making a profit on the service. Table 3.2 includes two sets of coefficients for the rural area analysis, one in which a general effect is found related to the uninsured rate and a second set in which a significant association is observed only when uninsured patients are not clustered at particular hospitals within the county; see Appendix D for details. For the rural counties studied, higher uninsured rates are associated with lower inpatient capacity for ICU services and, where uninsured patients are not concentrated at specific hospitals, psychiatric inpatient services. A 4.4 percentage-point higher uninsured rate (one standard deviation above the mean percent uninsured discharges of 4.4 percent for the 426 nonmetropolitan county observations) is associated with a 11.1 percentage-point decrease in the number of ICU beds per 100,000 population. When patients are not concentrated at hospitals (adjusted model), the size of the effect increases. A 4.4 percentage-point higher uninsured rate is associated with a 13.3 percentage-point decrease in the number of ICU beds per 100,000 population and a 15.8 percentage-point decrease in the number of psychiatric beds. Viewing the adjusted percentage changes in terms of the number of beds in the “average” rural county, the findings indicate a decrease from 423.3 ICU beds (the mean number) to 367 ICU beds and a decrease from 10 to 8.4 psychiatric beds per 100,000 population. Both the urban and the rural analyses show a consistent association between higher uninsured rates and lower levels of provision of high-technology services and community services. In rural areas, a 4.4 percentage-point higher uninsured rate is associated with a 34.5 percentage-point lower proportion of hospitals offering magnetic resonance imaging (MRI) and a 54.1 percentage-point lower proportion of hospitals offering lithotripsy (ESWL). Compared with the “average” uninsured rate rural county, this difference results in a decrease from 20.1 to 13.9 percent of hospitals offering MRI and a decrease from 6.6 to 3.0 percent of hospitals offering ESWL. When patients are not concentrated at specific hospitals, the size of the effect increases and additional findings become statistically significant, including decreased availability of transplant services, MRI, radiation therapy, ESWL, and Meals on Wheels. In contrast to urban areas, for rural counties there is a strong association between higher uninsured discharges and lower financial margins, whether or not the analysis is adjusted for the concentration of uninsured patients. Unadjusted, a 4.4 percentage-point higher uninsured rate is associated with a 37.4 percentage-point lower margin, which translates into a decrease from a 2.0 percent to a 1.3 percent aggregate margin. With the adjusted model, a 4.4 percentage-point higher uninsured rate is associated with a 48.4 percentage-point lower financial margin, representing an even larger decline in aggregate margin, from 2.0 percent to 1.0 percent. The lower margins may account in part for the lower likelihood that hospitals offer specific costly high- technology services. They may also affect quality and staffing in ways not observable in the analysis.
OCR for page 112
A Shared Destiny: Community Effects of Uninsurance Hospital-Level Analysis One would expect that the findings in the area-level analyses reflect the strategic behavior of hospitals within urban and rural markets, in response to financial incentives and disincentives to offer services and maintain staffed beds as well as to each hospital’s mission and history in its community. For example, inpatient and emergency psychiatric and transplant services have the potential to bring in uninsured patients whose care is likely to be very expensive. In the case of NICUs, for example, state Medicaid programs almost always cover the costs of care for the newborn of previously uninsured women, so that the burden of uncompensated care is likely to be less. Hospitals may choose not to offer psychiatric services for fear that they will not be reimbursed for a significant portion of the expensive care they provide (Gaskin, 1999). Alternatively, some hospitals may offer these services because there are few other sources of such care in the community and they are willing to absorb the losses in lower margins. While analyses by hospital rather than area ideally should capture information about these relationships, data limitations precluded the development of appropriate models that would have done so. The hospital-level analysis for urban areas (described in Appendix D but not presented here) makes fewer untested assumptions about the interactions among hospitals than do the area analyses, yet the Committee considers the hospital-level findings to be less reliable than the MSA-level findings. The hospital-level model indicates how the average individual hospital may respond to increased demand for hospital care from uninsured persons as measured by the MSA-wide uninsured rate. However, it is unlikely that the overall market response to uninsurance is merely the weighted average of individual hospital responses in the market. This would assume that hospitals neither behave strategically nor respond to the actions of their competitors. For example, a safety-net hospital may eliminate a high-tech service because they cannot support it with their payer mix. This may pave the way for a competing hospital to offer the service. Alternatively, this may reduce the availability of the service to the uninsured and insured populations who use that hospital. Neither of these potential outcomes is observable in the hospital-level model, while the MSA-level models allow for the observation of the overall association of uninsurance and the availability of hospital services, net of the strategic behavior of individual hospitals. The hospital-level findings for urban hospitals indicate that all the measures of inpatient bed capacity are negatively and significantly associated with the area uninsured rate, except for ICU beds, which are positively and significantly associated. Of the other variables, only alcohol and chemical dependency services (services disproportionately utilized by vulnerable populations) are negatively and significantly related in the hospital-level analysis. See Appendix D for a more detailed discussion of findings and the limitations of the analysis.
OCR for page 113
A Shared Destiny: Community Effects of Uninsurance Hospital Financial Margins As with health departments and ambulatory care clinics, many hospitals operate with financial constraints that leave little room for cross-subsidizing the costs of care for uninsured patients, who on average pay a small proportion (7 percent) of their total inpatient hospital bills (IOM, 2002b; MedPAC, 2002a). During the early and mid-1990s, the average total operating margins for public academic health centers, which disproportionately serve uninsured patients, were approximately zero (Commonwealth Fund, 2001). Despite a significantly increasing average total margin for all hospitals from 1.8 to 3.7 percent between 1991 and 1996, among those that provided the equivalent of at least 6 percent of their gross patient revenues in charity care, margins declined by 1.2 percentage points (Commonwealth Fund, 2001). In urban areas, the adverse financial effects on hospitals are more likely to have an impact on low-income residents who have few options for care other than their local public or private safety net hospitals. For example, urban hospitals in Newark, New Jersey, serve a much higher proportion of uninsured patients, compared with hospitals in the surrounding suburbs. This means that these inner city hospitals face great financial pressure to cut their costs by cutting services, but neighboring hospitals that do not serve such a large number or proportion of uninsured patients may continue operations as usual (Draper et al., 2001a). In rural areas, in contrast, there is less opportunity to segment the market and serve only a portion of the community, given the limited supply of health care practitioners and the smaller economic base to support hospitals. As a result, the adverse effects on hospital operating margins are more likely to affect the care of all community residents in rural areas than in suburban or urban communities (Sutton et al., 2001). Within rural communities, however, whether or not uninsured patients are more concentrated in one or a few hospitals than is the general population will determine how their presence affects the hospitals’ financial positions. However, in rural counties where uninsured patients are relatively more concentrated in one or a few hospitals than is the county population overall, hospital margins—both in the aggregate and by hospital—are less affected by rural or county uninsured rates than those in counties where uninsured patients are more widely dispersed among local hospitals (Needleman and Gaskin, 2002).7 7 This finding may be weakened by the fact that many rural counties have only one hospital, especially in western states. Alternatively, if a county has more than one hospital, the concentration of uninsured patients in a single facility may be associated with that hospital having an external source of funding to care for uninsured persons. As a result, the hospital may have a positive total margin, while the other hospitals in the county might be suffering financially and in danger of further financial decline if the burden of caring for the uninsured were to be dispersed among all hospitals in the county.
OCR for page 114
A Shared Destiny: Community Effects of Uninsurance Hospital Closures and Conversions A community’s high uninsured rate or an increasing demand for care by uninsured persons may influence decisions by a local hospital’s governing body to merge; to convert from public to private ownership status, or from private non-profit to for-profit status; or to close. These major changes are usually prompted by a variety of factors. A hospital’s decision to convert its ownership status may be motivated by the need for funds to continue operating, either independently or as one of a group of hospitals, or by the need to raise capital (Meyer et al., 1999; Needleman, 1999). Local uninsured rates and the burden of uncompensated care costs to local and state government contributed to the conversions of three large urban public hospitals—in Milwaukee, Wisconsin, Boston, Massachusetts, and Hillsborough County, Florida—to private ownership during the 1990s (Bovbjerg et al., 2000b). The conversions were spurred by changing market conditions and political decision making, particularly localities’ interest in cutting expenses (Meyer et al., 1999; Bovbjerg et al., 2000b). The conversion, merger, or closing of a hospital may lead to improved and less costly health services locally. It may also promote greater access to higher-quality care by making ambulatory primary care more available with the resources freed up from support of an unviable and inefficient inpatient facility (Needleman, 2000). Chapter 4 discusses such transitions in the case of rural hospitals. However, some changes bring with them the risk that the whole community will end up with less access to services or lower-quality care. For example, the loss of a physical plant and with it the relatively automatic support that a bricks-and-mortar institution commands may result in a net loss of public dollars for indigent health care (Bovbjerg et al., 2000b). A study of three public hospital closings in the 1990s finds that the basic lack of capacity to finance needed services for local underserved populations was not addressed, although the delivery of health services was better integrated through managed care arrangements that followed the closures (Bovbjerg et al., 2000b). For the most part, local public funders did not reward providers who inherited the patient base of the closed public hospitals for the cost savings and improved quality of care that came out of the conversion process, which led to a decline in local dollars for services to uninsured persons within these communities following the closings. In addition, a change in ownership status or operations can result in the relocation of a hospital from a neighborhood with a high uninsured rate to an area with a lower uninsured rate. This was the case in one Florida hospital, where, as part of its reorganization, a public hospital affiliated with a medical school shifted its base of operations to a new campus away from its inner city location and opened new outpatient facilities in more prosperous neighborhoods (Meyer et al., 1999; Bovbjerg et al, 2000b). The proposed closing of a public hospital may be perceived by other providers in the area as threatening the financial stability of their own institutions by potentially spreading the effects of uninsurance more broadly throughout the
OCR for page 115
A Shared Destiny: Community Effects of Uninsurance community. As a result, some closings may be successfully opposed. In Los Angeles County in the mid-1990s, for example, uncompensated care costs attributed in part to a sizable uninsured population led the county’s Board of Supervisors to consider closing a public hospital affiliated with the University of Southern California (one of many public hospitals in the county and the largest). Neighboring private hospitals threatened to close if faced with the public hospital’s patients, and this political resistance was sufficient that the county decided to keep its public hospital open (Cornwell et al., 1996). The heterogeneity from community to community in the degree of concentration or dispersion of safety net arrangements is described briefly in Chapter 2. The relative influence of greater or lesser dispersion of uninsured patients on uninsured rate is hinted at in the Committee’s commissioned papers and is likely to depend on the local configuration of health services organization and financing. When public hospitals convert to private ownership, however, the availability of services for all community residents is more likely to be adversely affected, as documented in two studies. The first is the previously cited study of hospital conversions in Florida between 1981 and 1996 (Needleman et al., 1999). The authors conclude that public hospitals provided less care after conversion. Although converting public hospitals provided lower levels of uncompensated care on average (9.1 percent) before their change in status, compared to public hospitals that did not convert (14.7 percent), they provided even less after conversion (6.8 percent). For the two converting hospitals in counties with other hospitals, their market share of uncompensated care declined significantly, from 65 percent to 39 percent (Needleman et al., 1999). The second study followed 52 privatizations of public hospitals to nonprofit or for-profit status in California, Florida, and Texas between 1980 and 1997 (Desai et al., 2000). The authors find a statistically significant 23 percent decrease in the level of uncompensated care given by public hospitals that converted to for-profit status, compared with care provided by comparable hospitals that did not convert, and greater decreases in uncompensated care among converting hospitals in lower-income areas compared with those in higher-income areas (Desai et al., 2000). There are regional variations in the level of uncompensated care in converted hospitals after such changes. In California, for example, private hospitals have lower uncompensated care levels, on average, than Florida, reflecting both a greater availability of public hospital services and more extensive Medi-Cal eligibility and enrollment (Needleman et al., 1999). Accounting for differences in size, teaching status, and urban or rural location, levels of uncompensated care at 15 Florida private nonprofits that converted to for-profit status between 1981 and 1996 remained relatively stable, at a lower-than-average rate compared with all private nonprofit hospitals in the state. The hospitals that converted, however, provided lower levels of unprofitable services such as emergency and trauma services after the conversion (Needleman, 1999). In contrast to public hospital conversions, for nonteaching hospitals, the shift from private nonprofit to for-profit status resulted in little change in the amount of
OCR for page 116
A Shared Destiny: Community Effects of Uninsurance uncompensated care delivered, the level of community benefits, and the price of care for all community residents (Mark and Cheng, 1997; Young et al., 1997; Needleman et al., 1999; Young and Desai, 1999; Hadley et al., 2001). Conversions of 137 hospitals from private nonprofit to for-profit status between 1985 and 1996, compared with similar hospitals that did not convert, resulted in few changes in the numbers of hospital births or emergency department visits, but the numbers of total outpatient visits and Medicaid inpatient discharges increased more slowly (Hadley et al., 2001). Neighboring hospitals (within 5 miles) of conversions reported larger (but not statistically significantly larger) increases in the volume of publicly insured patients (Medicare and Medicaid), no significant change in the number of emergency or outpatient visits, a significant increase in the number of births (particularly in public and large teaching hospitals), and a shift in the pattern of emergency department visits to large teaching hospitals, with a decline at neighboring public hospitals (Hadley et al., 2001). RESEARCH QUESTIONS 3.1 Access to Care Across the Community Does the local uninsured rate, independent of other factors, affect residents’ access to care throughout the community? Existing studies of the relationship between community uninsurance (state or MSA uninsured rate) and access to care offer preliminary evidence that, particularly for low- to moderate-income and uninsured populations, higher local uninsured rates are associated with worse access to care. These provocative findings should be corroborated with additional studies that include more refined measures of access as outcome variables. Well-controlled, longitudinal study designs could allow researchers to tease out the difference between the effects of the uninsured population as exerting an aggregate influence on local health services (e.g., because the uninsured constitute a large portion of the population being studied) as compared with having an ecological impact (e.g., the independent effect of uninsured rate on access to care for insured persons). Is the supply of services in more affluent urban neighborhoods and suburbs affected by proximity to large or small populations of uninsured persons, and how? 3.2 Access to Primary and Preventive Services How does the local uninsured rate affect the availability of primary and preventive services for the community’s insured as well as uninsured residents? Because primary and preventive services are often considered elective by patients (unlike emergency medical services and hospital inpatient care), use of these services may fall off more quickly when patients lack the financial means to pay for care. When the number or proportion of uninsured patients increases within a primary care practice, the combination of decreasing patient utilization
OCR for page 117
A Shared Destiny: Community Effects of Uninsurance and an increasing proportion of uninsured visits may also adversely affect the financial position of the practice. As a result, primary care practices may become financially unviable. A conservative estimate of the amount of charity care that physicians provide annually to uninsured patients is roughly $5 billion (Hadley and Holahan, 2003). Better information about the distribution and impact of the burden of providing uncompensated care among physician practices and its implications for the availability of high-quality, stable primary care services is needed in order to understand the dimensions of the problems that uninsurance poses for communities. 3.3 Access to Specialty Care, Including Emergency Medical Services How does the local uninsured rate affect the availability of specialty services, including emergency medical services and trauma care for the community’s insured as well as uninsured residents? The legal duty of hospital EDs and trauma units to screen and medically stabilize all patients regardless of ability to pay is one source of financial stress on hospitals. What is less clear is how important the financial demands of providing care to uninsured patients are compared with other reasons for emergency department overcrowding and causes of financial strain (Tsai et al., forthcoming 2003). Further study is needed to understand the degree to which ED overcrowding and financial instability could be ameliorated by reductions in the number of uninsured persons, the restructuring of financial subsidies for their care, or the strategic diversion of emergency department patients to alternative sites for primary care. In addition, more research is needed to understand the degree to which administrative and staffing decisions made within hospitals (e.g., the specialized services such as psychiatric inpatient care) reduce community access and what alternatives exist to address the problem without jeopardizing the financial health of hospitals. While emergency medical services and trauma care provide some of the strongest evidence about access to specialty care, the Committee’s findings about the difficulty community health center physicians have obtaining referrals for their uninsured and other patients, and about hospital-based specialty services suggest that further investigation of the availability throughout a community to a wide range of specialty services is merited. How does the institutional setting (e.g., AHC) influence or moderate the effect of uninsurance on access to other specialty services? And what impacts does uninsurance have on the other related missions of these institutions, for example, medical education in academic health centers? 3.4 Access to Hospital-Based Services How does the local uninsured rate, in conjunction with public institutional support such as DSH payments, affect hospital service offerings, financial stability, and decisions to close? The limitations and preliminary findings of the Committee’s commissioned analyses of hospital inpatient capacity, services, and financial margins suggests a number of ways that community effects on access to care might be explored.
OCR for page 118
A Shared Destiny: Community Effects of Uninsurance Perhaps most important is the examination of strategic or competitive responses of hospital administrators and their boards to local market conditions, including the decisions of competing hospitals to expand or shrink inpatient capacity and boost or trim the provision of services, and the relationship between these responses and individual hospital financial margins. More refined studies, based on market areas smaller than MSAs, are needed to understand better the impacts of uninsured populations on hospital services and operations. For example, findings for rural areas that the concentration of uninsured patients may lessen the size of the effect of local uninsured rate on inpatient capacity, services, and margins of all hospitals in the county, on average, suggests that such concentration of uncompensated care caseload among a few providers may be beneficial to the health care system as a whole. On the other hand, concentrating all care for uninsured persons in one facility, such as a county hospital, may both limit access to care, compared with more dispersed safety net arrangements, and lead to poor quality of care. Further research is needed to assess the overall effect of concentration and dispersion on access to hospital-based care. Although there were hundreds of hospital closures during the 1980s and 1990s, there has been little documentation of the role that a community’s uninsured rate may have played in these closures. Most studies of hospital closings do not directly address the influence of uninsurance because of the limited information about local uninsured rates and about hospital payer mix. Closures of public hospitals are of particular concern because these hospitals often serve as providers of last resort. Studies that examine reasons for these closings and the impacts of these closings or conversions are needed. 3.5 Quality of Care How does uninsurance within communities affect the quality of care available to and provided to all residents, insured and uninsured alike? An important and underexamined potential impact of uninsurance within a community is on the quality of care provided to all residents. The Committee’s earlier reports document problems of quality (based on measures of process and outcome) in the care of uninsured patients. This report focuses more holistically on the effects on a provider’s performance overall when uninsured patients or uncompensated care overwhelm or impair the provider’s capacity to provide quality services. No studies have directly assessed the correlation of community uninsured rates with hospital quality of care. There are at least two areas in which high uninsured rates could adversely affect quality of care for all patients. First, if high rates of uninsurance adversely affect hospital margins, this could lead to cutbacks in nursing staff, which in turn threaten quality of care (Needleman et al., 2002). Second, if high uninsured rates contribute significantly to emergency department overcrowding at certain facilities, those hospitals may deliver poorer quality of care during times of patient overload. Detailed studies at the level of the patient are needed to understand the relationship between overcrowding, patient safety, and quality of care.
OCR for page 119
A Shared Destiny: Community Effects of Uninsurance SUMMARY The Committee’s hypothesized effects of uninsurance on access to health care services are borne out by evidence that, although preliminary, offers reason for concern. For primary and specialty care delivered by physicians in private, clinic, and hospital settings and in hospital emergency, outpatient, and inpatient settings, the burden of providing unreimbursed care to uninsured patients threatens and impairs the capacity of a community’s health care institutions and providers. Based on the evidence developed in this chapter, the Committee concludes that moderate- to lower-income residents of communities with higher uninsured rates are more likely than all residents to experience adverse spillover effects of uninsurance. The findings of the Committee’s commissioned analyses, while preliminary, indicate that residents with moderate or higher incomes may also be at risk for diminished access to care. Health care professionals and institutions exist in social and economic relationships with the rest of the community that go beyond their roles as providers of health services. Community effects related to these relationships are the subject of Chapter 4.
Representative terms from entire chapter: