The Expansion of Coverage

The original framers of Medicare were well aware that the inclusion of nursing home services could destroy the fiscal viability of the system, particularly if it covered the kind of long-term custodial care most nursing homes were then providing. Despite this, substantial inflation in hospital rates during the 1950s and early 1960s made it desirable to have less costly alternatives than hospitals for patient convalescence (interview with Wilbur Cohen, personal communication, 1980; U.S. Senate Committee on Finance, 1970; Vladeck, 1980). Thus, as originally conceived, Medicare covered only hospital care, but was amended to add coverage for convalescing patients in "extended care facilities" (ECFs). With anticipated daily costs about half those of convalescent care in a hospital, ECFs were seen as a cost-efficient form of institutional care.

Although the extended care provision of Medicare was soon to haunt its framers, its impact pales beside that of another amendment to the Social Security Act, Title 19 or Medicaid. Medicaid attracted little attention from policymakers, and coverage of skilled nursing home care was included almost by default—again through vendor payments and without much definition or a budgetary limit. This coverage (and the eventual extension to intermediate care as well) provided the financial fuel for the further growth of the nursing home industry.

With the passage of Medicare and Medicaid, sufficient funds were available to many of the elderly for much pent-up need to be translated into demand for and utilization of nursing home care. It was not, however, only the infirm aged residing in the community who came to demand nursing home care. Nursing home care also emerged as a substitute for housing (and some care or supervision) previously provided to many of the elderly in other settings, particularly mental institutions (Dunlop, 1979; Manard et al., 1975; Vladeck, 1980; Scanlon and Feder, 1984; Harrington and Grant, 1985). One observer estimated that 25 percent of the increase in nursing home utilization between 1960 and 1970 can be attributed to the deinstitutionalization or diversion of individuals from mental institutions into nursing homes (Morton Research Company, 1982).

Consumers of care and their families had a multitude of reasons for choosing nursing homes over other forms of institutional care. But it was also in the interest of the states and localities. In these other settings, most of the costs were borne by state and local governments, while transfer to a nursing home usually meant the state could collect the federal matching funds for the care of these individuals in nursing homes under the Medicaid program. As a result of these diverse factors, including reductions in hospital lengths of stay, the percentage of elderly persons in nursing homes rose 58 percent in the decade between 1950 and 1960 and 107 percent between 1960 and 1970.

A combination of factors, therefore, fueled the demand for and utilization of nursing home beds. Increased but unevenly distributed expenditures under Kerr-Mills led to significant growth, but it was the passage of Medicare and Medicaid in the mid-1970s, in conjunction with demographic trends, that led to substantially extended demand by the elderly for nursing home care. From the viewpoint of potential providers, expansion was a result not only of increased demand, but also the increased availability of public reimbursement and the attractiveness of the rates set up under these new programs, a result of the policymakers' desire to assure substantial provider participation. While they expanded eligibility, Medicare and Medicaid could accomplish little if too few individuals and organizations were willing to provide services to program beneficiaries. To attract providers to the program—and increase the supply of nursing home beds—the federal government pursued two basic policies, one related to reimbursement and another to regulation.

Increased Reimbursement Rates. First, Medicare adopted a cost-based reimbursement policy very, similar to that it employed for hospitals. Providers were basically reimbursed for their reported costs. The formula placed no ceiling on reimbursement rates, and variations in reported costs between providers were ignored. In addition, the program provided proprietary nursing homes a "profit" based on their net invested equity in the facility. Finally, the American Hospital Association (AHA) lobbyed effectively for Medicare to reimburse hospitals for mortgage interest as well as depreciation of capital equipment, including the facility. This was extended to nursing home reimbursement. As a result, investors were virtually assured of covering their mortgage payments—and having a positive cash flow (from depreciation) as well in the first years of the mortgage—through the Medicare program. They were also virtually assured of having their costs covered as well as receiving in effect a guaranteed profit (Shulman and Galanter, 1976; Vladeck, 1980; Ohio Nursing Home Commission, 1979; Washington Senate, 1978).

Lenient Regulatory Posture. Medicare also adopted a policy of easing nursing homes' entrance into the program. Although enacting mandatory minimum health and safety regulations for facilities participating in Medicare, the federal government estimated that few facilities could actually meet even the minimum standards. Consequently, granted nursing homes were given participatory status in Medicare and Medicaid if they were in "substantial" (rather than full) compliance with the new federal minimum health and safety standards. The homes simply had to present to the survey and certification agency a plan for correcting their violations in order to be certified and receive payments.

This lenient regulatory posture, combined with open-ended, full-cost reimbursement, a guaranteed profit factor, and interest plus depreciation for capital reimbursement in a program financially supported by the federal government, made investment in nursing homes very attractive. In addition, government funding and the aging of the population ensured the industry of a ready supply of customers. Thus, government policies eliminated much of the risk generally associated with most investments and nearly all new ventures—while promising substantial profits to those with the knowledge and ability to manipulate the system. As a result, in less than a decade the industry expanded its total supply of beds from approximately 460,000 (in 1965) to more than 1.1 million beds (by 1973) an increase of 139 percent in less than a decade.⁶

Expenditures grew even more rapidly. In 1950, total expenditures for nursing home care were less than $190 million, with government paying only 10 percent of that total. By 1960, government paid 22 percent of a total of $526 million. By 1965, just before the onset of Medicare and Medicaid, expenditures totaled $1.328 billion, an increase in just 5 years of 153 percent. Between 1960 and 1974, expenditures on nursing home care grew approximately 1400 percent (U.S. Senate Special Committee on Aging, 1974). By 1978, the nation spent $15.8 billion on nursing homes, with government paying 53 percent of the total (Gibson, 1979). And by 1982, expenditures for nursing home care exceeded $27 billion annually with government paying nearly 55 percent of the total (GAO, 1983b). In 1984, expenditures on nursing home care exceeded $32 billion with government paying 49 percent (Levit et al., 1985).

Impressionistic Evidence

Impressionistic findings suggest that nonprofit facilities tend to provide better quality of care and quality of life, but these findings are hardly conclusive. Brooks and Hoffman (1978) visited many intermediate care facilities (ICFs) in the Cleveland metropolitan area and concluded,

We . . . are impressed with the noticeable differences between the two types of ownership. In general, homes for the aged (nonprofits) appear to have a more "homey" atmosphere, to be more adequately staffed, to be kept in better repair, to be more closely linked to community organizations, and to have more patient activities than do proprietary intermediate care nursing homes.

Vladeck (1980) deals with this issue in his book, Unloving Care. On the basis of his observations and interviews, Vladeck concludes,

. . . on the average, voluntary, facilities are somewhat better than proprietary ones. The worst nursing homes are almost exclusively proprietary. But in the middle ranges, there is substantial overlap. The best way to visualize the difference might be to conceive of the range of quality in each of the two types of nursing homes as a quasi-normal distribution .... The two distributions overlap markedly, with the mean for the voluntaries slightly higher than the mean for proprietaries, and with the voluntaries having a shorter low-quality tail.

Winn and McCaffree (1976) conducted a study of 282 nursing homes identified by state officials (in 16 states) as being high-quality homes and developed a different impression· (Their purpose was to study how the patterns of services received by patients in these excellent facilities differed from other facilities.) They observed that the proportion of nonprofit facilities identified as "high quality" did not differ from the proportion of nonprofit facilities in the population. However, without information about the distribution of the "adequate" and "seriously substandard" facilities, one cannot conclude that no difference exists.

The performance of nursing home chains is even more difficult to evaluate, even impressionistically, since it seems to vary. significantly from chain to chain. In a study for a congressional committee, Richard Levinson (1984) and other staff interviewed state Medicaid and health department officials regarding the impact of chain ownership on quality. Most state officials lacked systematic information about quality of care, much less how it varies by type of ownership. Also, state officials often do not know which facilities are part of multifacility. chains. However, while officials are concerned about the growing concentration of nursing home chains and the future implications of this trend, many currently view the chains in a fairly favorable light. Thus, Levinson notes, that though responses "are anecdotal and impressionistic, at best,"

. . . respondents from the state Medicaid administrations commonly report that the chains have been responsible for bringing a number of local privately owned facilities ("room and pop" homes) which were substandard operations up to the minimum acceptable level . . . . The chains have the capital to upgrade institutions they purchase and are sometimes unwilling to jeopardize their name for deficiencies that can be remedied. Indeed, they indicated that some chains would not allow a facility they purchased to run under their name until it was brought up to—or close to—the minimum acceptable standard. It must be emphasized that these facilities are not necessarily viewed by state administrators as offering "good quality" care, but are just good enough to meet state standards .... Some credit the chains with helping them to remove the worst abuses in nursing home facilities.

In interviewing officials from several states, we heard much the same about the nursing home chains, but the view differed markedly depending on the chain under discussion. In four states where one chain operated, state officials cited it as the worst performer among all the chains—with quality of care a major problem in many of its facilities. Two other nursing home chains, however, were cited as making efforts to upgrade the facilities they purchased.²²

Some state officials, like Governor John Carlin of Kansas, have expressed concern about the growing concentration of chain ownership, which they fear may lead to decreased competition and increased costs. Also corporations with facilities in one state and headquarters in another state are seen as more difficult to regulate. Several state officials expressed concern that large chains could exercise too much influence in state rate setting and regulatory. issues with threats to withdraw from the Medicaid program. In addition, some state officials felt "outclassed" when regulatory actions involved legal proceedings because of the chains' high-priced and highly qualified legal representation, while state agencies typically must rely on relatively inexperienced lawyers.

Other state officials, however, felt that the chains were often more responsive to state Medicaid and survey agencies than were individual owners. Some cited the managerial ability of some chains in this regard. Others, like Dr. Janice Caldwell of the Texas Department of Human Resources, felt that a chain is more likely to take corrective action promptly when one of its facilities is cited for deficiencies because the parent corporation would be unwilling to jeopardize its relationship with state agencies and place all its facilities at risk of increased regulatory activity.²³

Our impression, based on visits to more than 200 facilities and interviews with state officials, ombudsmen, and families of nursing home patients, is that nonprofit, church-affiliated homes tend to provide better quality of care and quality of life than proprietary facilities. Excellent facilities are found in each category, but truly wretched ones are almost always proprietary. However, one's evaluation of the performance of nursing homes may differ according to the dimension of quality under consideration. For example, Ms. Iris Freeman, head of the nursing home advocates program in Minnesota, observes that her office receives complaints about all types of facilities; complaints about quality of care tend to be about proprietary facilities, while complaints about violations of patients' rights (particularly about paternalism toward patients) tend to be lodged against non-profits (Freeman, personal communication, 1985). Clearly, definitive conclusions about differences in quality are difficult to reach· However, most observers' general impressions seem to be that nonprofits, on the average, provide better quality. as it is generally understood.

Empirical Research: Variations in Input Measures of Quality

The findings of empirical studies on the relationship of ownership to resource input and structural measures of quality. are occasionally inconsistent, although most have found significant differences between for-profit and nonprofit facilities in the amount of resources allocated to direct patient care. What these differences indicate, however, is a matter of substantial debate. The validity of resource inputs, such as nursing hours per patient or expenditures on raw food, as surrogates for quality. has been questioned by critics who argue that such measures may hear little relationship to either quality. of care or quality of life in a facility, (O'Brien et al., 1983; Levey et al., 1973; Kane et al., 1983). Several studies have failed to find any statistically significant relationship between structural factors or resource inputs and process and outcome measures of quality (Lee, 1984; Kurowski and Breed, 1981).

Despite such criticisms of input measures of quality, their use can be supported on several grounds. First, the availability of resource inputs is clearly a precondition to their use in patient care (Kosberg and Tobin, 1972; Kosberg, 1973). Second, the lack of correlation between resource input or structural measures of quality and process or outcome measures may be due to the lack of variation in these independent variables in the homes studied (Linn et al., 1977; Kurowski and Breed, 1981). Third, some studies have found some resource inputs to be related to both process and outcome measures of quality of patient care (e.g., Kurowski and Breed, 1981; Shaughnessy et al., 1980). Linn et al. (1977) found that the number of registered nurse (RN) hours per patient day was positively associated with patient survival, improvement, and discharge from the nursing home. Raw expenditures on food and the availability of individual dietary, planning were also associated significantly with two of the outcome measures of quality of care. In a 1977 study of homes rated as superior, the American Health Care Association found that these homes outscored other facilities in the number of RN hours per patient day but were lower in licensed practical nurses (LPNs).²⁴

Two state studies also examined the value of various resource inputs as indicators of quality. The Virginia Joint Legislative Audit and Review Commission (Virginia JLARC, 1978) identified facilities that state inspectors had found deficient in "meal menus, adequate food portions, or the quality of meals," and facilities that had "complaints about food." Raw food expenditure data were available for 9 of the 17 such nursing homes. JLARC found that

In all nine cases, the per day expenditure was below the state average. This supports the belief that low raw food costs and poor meal quality are related, and that food costs could serve as a measure of dietary. adequacy.

The Ohio Nursing Home Commission (1979) studied differences between facilities rated as "high quality" and "low quality." On the basis of agreement from three sources (a reputational survey of nursing home ombudsmen, hospital discharge planners, and state inspection staff; licensure and federal certification survey reports; and inspections of facilities by commission staff), 60 nursing homes were identified as providing either high or low quality. Of these, 27 low-quality homes and 28 high-quality homes had current Medicaid cost reports on file with the Ohio Department of Public Welfare. A comparison of facilities' allocation of funds (Table 6) shows that "high-quality'' homes have higher expenditures on direct patient care items, such as food, RNs, aides, medical and rehabilitative care, housekeeping, and laundry and linen.²⁵

TABLE 6

Spending Differences per Patient Day Between a Sample of High-Quality and Low-Quality Nursing Homes Participating in the Ohio Medicaid Program.

While the validity of resource inputs as indicators of quality is not definitively established, such studies support the utilization of resource input as at least partial indicators of quality of care. Several studies of for-profit and nonprofit facilities use such measures, and they yield fairly consistent findings about ownership. Two studies of resource inputs found no significant differences between for-profit and nonprofit facilities. Holmberg and Anderson (1968) interviewed administrators in 118 of the 392 nursing homes in Minnesota in 1967 and found several structural differences between for-profit and nonprofit homes.²⁶ Although average staffing patterns (RN hours per patient day, average number of patient care staff) were very similar, nonprofits had significantly more physician hours per patient per week than the proprietaries, and the for-profits had more administrative hours per week.

However, Holmberg and Anderson (1968) concluded that most quality. indicators were not significantly related to ownership type and that the differences between ownership types were less significant than the differences among facilities within each category. Levey et al. (1973) reached a similar conclusion. They studied 129 of 690 nursing homes in Massachusetts, using an index of nursing services, patient activities, and physical plant characteristics from survey and patient care reports. They found no significant differences in their quality index on the basis of facility. ownership.

Some observers argue that the findings of Holmberg and Anderson (1968) and Levey et al. (1973) are not typical because Minnesota and Massachusetts are thought to have more-stringent regulatory. policy and higher-quality nursing homes than is average and the quality proxies used were based on resource inputs regulated by the state. A study by Gottesman (1974) supports these findings. Gottesman gathered data on 40 nursing homes and on the social, physical, and mental characteristics of 1,144 residents in those facilities. He also observed the activities of nursing home residents and staff during two 12-hour periods. Gottesman found better psychosocial activities and staff involvement in nonprofit homes and that "church-related facilities seem to have a strong sense of 'family' in their sponsorship." However, he argues that other factors, such as facility size, patient characteristics, and the mix of public and private-pay patients, may have more significance for these differences in quality than does ownership. For instance, of the proprietary facilities with high levels of public-pay patients, Gottesman argues,

These homes have a high proportion of socially marginal residents, fewer financial resources, and a resident group with many social disabilities and little community involvement. It should not be surprising that they come off most poorly with regard to basic medical and psychosocial activities as well as staff involvement.²⁷

Other studies of resource inputs consistently find differences between for-profit and nonprofit nursing homes. For example, the concern about profit-taking at the expense of patient care has been heightened by a study (Fottler et al., 1981) that examined the relationship between profits and four measures of quality of patient care.²⁸ This study found that there is a consistently negative relationship between a nursing home's profitability and resource input measures of patient care quality. That is, the study found that quality decreases as profits increase. While reliance on structural/input measures of quality, like those employed in the Fottler study, may be subject to debate (O'Brien et al., 1983), few observers would argue that staffing is unrelated to quality. Most would agree that sufficient staffing is, at the least, a precondition to high quality of care. Moreover, the study's findings confirm the view of many long-time observers of the nursing home industry. As Vladeck (1980) observes,

The motivations, or actual behavior, of profit-seeking firms may run counter to the well-being of nursing home residents. Proprietary. facilities tend to be physically smaller than voluntary ones, especially in the provision of public space. They have incentives to discriminate against Medicaid recipients and against admission of sicker patients. They are more likely to cut corners on supplies and staffing. And there can be no denying that most of the really scandalous conditions found in nursing home investigations have been in proprietary facilities.

Other studies find substantial differences between for-profit and nonprofit homes on resource inputs. The 1973 National Nursing Home Survey demonstrated that nonprofit homes had significantly more staff than for-profit facilities, with twice as many clerical, food service, housekeeping, and maintenance staff. Proprietary facilities had more administrative hours, as well as medical and therapeutic (e. g., physical therapy) staff hours. In both the 1973 and 1977 surveys, nonprofits had more RN hours (U.S. Department of Health, Education, and Welfare, 1975b, 1979).

Winn (1974) found similar differences in a study of a matched sample of 24 proprietary and 24 nonprofit nursing homes in Washington state. Nursing hours per patient day were slightly higher in nonprofit homes, and total employee hours per patient day were significantly higher in nonprofit facilities. The New York State Moreland Act Commission (1975) found that nonprofit facilities spent an average of $16.02 per patient per day on nursing, while for-profits spent only $12.78 per patient day. When the analysts controlled for facility size and location, however, much of the cost difference between for-profit and nonprofits was eliminated. In general, "ownership variables added only minor explanatory power," but it did account for some of the variation in spending on nursing staff.

The Minnesota Senate/House Select Committee on Nursing Homes also examined the spending patterns of nursing homes and found significant ownership differences in the distribution of expenditures (Minnesota Senate and House, 1976). Proprietaries spent nearly twice as much as nonprofits on cost of capital (property and related expenses), while non-profits outspent the proprietaries on dietary, housekeeping, and plant operation (maintenance). In addition, the Minnesota study found

Proprietary SNFs incur 37. 7% of their costs for care-related items and 20.8% for property-related items. For nonprofit SNFs, the percentages were 41% for care-related costs and 11.1% for property-related items. The differentials for ICFs were about the same.

The Virginia study (Virginia JLARC, 1978) also found substantial differences in the spending patterns of for-profit and nonprofit facilities, including that "profit status was found to have an effect on the distribution of nursing home expenditures but not on overall cost." Using data from facilities with the same average number of patient days, the Virginia study found striking differences, depicted in Table 7.

TABLE 7

Spending Patterns in Virginia Nursing Homes, 1977.

Caswell and Cleverley (1978) also found significant differences in both the costs and distribution of expenditures by different nursing home ownership types. They found that while for-profit facilities had lower overall costs, the savings were primarily in areas that may have direct or indirect effects on the quality of patient care: housekeeping, maintenance, nursing, and dietary. The for-profits, however, reported higher costs of property (interest, depreciation, and rent or lease payments). Caswell and Cleverley concluded that many proprietary facilities were probably staffed at lower levels than minimum health and safety, standards required.

The Ohio Nursing Home Commission's analysis of 1977 costs reported by nursing homes to the state Medicaid agency revealed a similar pattern. On average, nonprofit facilities out-spent proprietaries on raw food and dietary, supplies ($4.67 per patient day versus $3.01), as well as on a composite of direct patient care items that includes nursing salaries, medical supplies and expenses, social services, and medical and rehabilitative salaries (Ohio Nursing Home Commission, 1979).²⁹

Data from Texas (Moden, 1982) reveal similar differences in the distribution of spending. For instance, in 1978 the average for-profit facility had total net revenues 16.5 percent less than the average nonprofit. But the for-profits, on the average, spent 37 percent less on patient care (e.g., nursing staff) and 31 percent less on dietary than nonprofits. In 1980, the pattern was similar. Proprietary facilities had total net revenues (on the average) that were only 7.5 percent less than the average for nonprofit facilities, and spent 29 percent less on patient care items (e.g. nursing, medical, and rehabilitative care) and 27 percent less on food than the average nonprofit nursing home. Table 8 displays data on per diem expenditures broken out by cost category for 1981.

TABLE 8

Data on Revenues, Expenditures, and Spending Patterns in Texas Nursing Homes, 1981.

Elwell (1984) studied 424 skilled nursing homes in New York State. His quality measures were "multiple scales of staffing and resource distribution" among the facilities. The distribution of resources was measured in terms of per diem expenditures on administration, medical, nursing, rehabilitation, social services, activities, nutrition, housekeeping, and other professional services. Staffing measures included RN, LPN, and aide hours per patient day, physician hours per patient week, and other professional staff hours per patient day. In addition, as a measure of quality, of life (privacy and personal space), Elwell included the proportion of patients in multiple-bed rooms.

The proportion of Medicaid patients per facility, the proportion functionally impaired, and the size and location of facilities were included to help isolate the effect of ownership on the measures of quality. Even when controlling for these, in each patient service area Elwell measured, nonprofits spent more than the proprietaries.³⁰ He concludes that "the differences among ownership types were consistent with the hypothesis that... voluntary institutions are superior medical and personal care facilities" (Elwell, 1984). When controlling for location, Elwell found that

In each of the six regions, proprietary facilities spent significantly less than the regional mean for administration, nursing, nutrition, and housekeeping services. Proprietary facilities spent significantly less for medical services in five of the regions, for social and rehabilitation services in four of the regions .... In no region did proprietary facilities spend significantly more than the regional mean for any of the service areas measured.

He concluded,

The results of this analysis provide considerable confirmation for the hypothesis that government and voluntary facilities offer superior medical and personal care to their patients. The allocation of financial resources as well as the number of staff hours per patient day were all consistent with the hypothesis.

Data on the impact of for-profit nursing home chain ownership on quality, of patient care are even more scant than that focusing on differences between for-profit and nonprofit facilities. Greene and Monahan (1981) analyzed data from 24 SNFs in the greater metropolitan area of Phoenix, Arizona. They felt that in the relatively unregulated market of Arizona, which at the time had no state Medicaid program (and thus no federal health and safety standards or inspections), might make the orientation of nursing home owners more apparent. As a measure of quality, Greene and Monahan used a composite of facility spending on such items as RN hours, expenditures on RNs, and dietary expenditures. Greene and Monahan examined the behavior of nonprofit homes, individually owned or locally headquartered proprietaries, and for-profit nursing home chains. They found that "for-profit institutions give significantly lower levels of care, even with charges and other factors controlled," and that "distantly headquartered chain operations tend to provide lower levels of care than locally-owned facilities." However, since the number of homes in the Greene and Monahan study is small and the unregulated market of Arizona may not be typical, their findings cannot safely be generalized.

Data on Texas nursing home expenditures in fiscal year 1983, displayed in Table 9, further illuminate the relationship between ownership and allocation of funds to resources thought to affect quality of patient care. The nonprofit facilities outspend the for-profit facilities, on the average, in every, category, including facility (property) and administrative costs. The differences in spending, however, are most notable in the patient care and dietary categories. Data on variations among the four largest nursing home chains operating in Texas are shown in Table 10. In general, the differences are consistent with earlier findings. The statewide mean for spending on the patient care category is 44 percent of total expenditures. For proprietaries it averages 43 percent, while for chain-owned facilities, spending on patient care is 42 percent of total expenditures. The two largest chains in Texas both spend less than 39 percent of the total on patient care cost items. Nonprofit facilities spend an average of 47 percent of their total expenditures in the patient care category.

TABLE 9

Revenues, Expenditures, and Spending Patterns in Texas Nursing Homes, 1983.

TABLE 10

Revenues, Expenditures, and Spending Patterns in Nursing Home Chains, Texas, 1983.

Variations in Licensure and Certification Deficiencies and Complaints

Another source of data that can be used to evaluate the performance of different nursing homes is information on state licensure and federal certification reports regarding violations of minimum health and safety standards. These reports have often been justifiably criticized in terms of failing to reveal an accurate, multidimensional picture of the quality of care provided to patients in nursing homes, much less quality of life (e.g., Hawes, 1983; Ohio Nursing Home Commission, 1979; New Jersey State Nursing Home Commission, 1978; Missouri State Senate, 1978). Because inspections are usually announced or regularly scheduled, owners and administrators can anticipate the approximate date of inspections. In addition, the survey forms focus almost entirely on structural and input characteristics—on the capacity of the facility to provide care rather than on the condition and satisfaction of patients. For these reasons, critics argue that the inspection system—and the violation citations that emerge from this process—yields insufficient information about quality. On the other hand, there is no reason to expect that the deficiencies/violations do not accurately reflect failures by the facility to meet existing standards, nor to expect that the citation system is biased in favor of any particular ownership type. Thus, while they do not fully measure quality, licensure and certification deficiencies may capture important aspects of quality of care.

Complaints reported by patients, family members, or others familiar with the facilities are another potential source of data about the performance of nursing homes. Many argue that complaints are especially valuable since they may reveal patients' and families' perceptions of quality. and may reveal more than regulatory reports about quality of life, augmenting inspection reports (Weisbrod and Schlesinger, 1983). There is some concern that complaint data may be biased against higher quality facilities, since in these homes patients and families may feel less fear of reprisal and less reluctance to complain. In addition, individuals may be more willing to lodge complaints when they feel most confident that corrective action will occur.

Existing studies and available data on regulatory violations and complaints indicate that nonprofit facilities offer higher quality of care and quality of life than proprietary ones. There are too few systematic data on the performance of nursing home chains—and too much apparent variation among different chains—to permit definitive conclusions.

The Minnesota Joint House/Senate Select Committee on Nursing Homes analyzed differences in complaints about poor quality of care (Minnesota Senate and House, 1976). The data from 1974 revealed that while only 39.1 percent of the licensed nursing homes in Minnesota were proprietary, 78.2 percent of the complaints filed with the state health department were lodged against for-profit homes. Of the 41 nursing homes that were the subject of three or more complaints during 1974, 39 (95 percent) were proprietary facilities.

The Ohio Nursing Home Commission (1979) analyzed characteristics of all (45) nursing homes that had such serious violations of health and safety standards that the Ohio Department of Health began proceedings to revoke their licenses in 1976. Only one (2 percent of the total) was a nonprofit facility, although nonprofit nursing homes constitute more than 20 percent of Ohio facilities.³¹

Koetting's study (1980) is perhaps the best-designed empirical study to date. Koetting selected a sample of 136 Illinois facilities divided into five ownership categories: (1) private non-profits; (2) government-owned nonprofits; (3) proprietaries with working owners; (4) chain-owned proprietaries; and (5) other proprietaries. Quality-of-care measures were derived from licensure surveys and inspection-of-care reports. Using regression analysis to isolate the impact of ownership on these measures, and controlling for facility size, patient mix, and occupancy rate, Koetting concluded that nonprofit homes had superior quality.³²

Riportella-Muller and Slesinger (1982) studied quality of care in 462 of 533 Wisconsin nursing homes. Measures of quality included numbers of code violations and complaints. On the basis of the complaint data, the nonprofit homes provided better quality, but the data on violations yielded more complex findings. Small nonprofit homes had fewer violations than small proprietary homes, but large non-profits (150 + beds) had more violations than similar proprietaries. Riportella-Muller and Slesinger concluded that although nonprofits performed somewhat better than proprietary nursing homes, the size and certification status of the facility were important mitigating factors and that simply eliminating for-profit entities would not necessarily significantly improve quality.

Weisbrod and Schlesinger (1983) also studied code violations and complaints filed against nursing homes in Wisconsin; however, their findings in favor of the nonprofits were much stronger than those of Riportella-Muller and Slesinger. Based on analysis of 431 nursing homes and data for 1976, Weisbrod and Schlesinger found that nonprofit homes had significantly fewer complaints than did proprietary firms. The picture for differences in ownership performance as measured by code violations was more complex. Nonprofit homes that were not church owned had significantly more regulatory violations than their for-profit counterparts of similar size. Church-owned nonprofits performed better on each measure (complaints and regulatory violations) than did for-profits.

States generally do not systematically report data on violations of state licensure and federal certification standards in a form that is useful to researchers. California, however, does assemble such data. The California Health Facilities Commission (1982) examined licensure citations to nursing homes that failed to meet minimum standard for the period 1977-1979. Proprietary firms generally performed worse than nonprofits. The average number of licensure citations per facility was 0.46 among church-related homes, 0.49 for other non-profits, and 1.8 for proprietary homes³³ (see also AFL-CIO, 1983a, b).

Texas also systematically records serious violations on which the Department of Human Resources has taken action. We analyzed three types of increasingly severe punitive actions imposed by the department on long-term care facilities in Texas between January 1 and April 29, 1983. During that time, 181 compliance letters were issued, 35 vendor holds on Medicaid payments were imposed, and 42 facilities were terminated from the Medicaid program. Punitive actions were almost exclusively taken against proprietary facilities. For-profit homes (840 or 88 percent of facilities in the Medicaid program) received 169 (93.4 percent) of the compliance letters, 100 percent of the vendor hold actions, and between 90 and 100 percent of the 42 contract terminations.³⁴ Nonprofit nursing homes (115, or 12 percent of facilities in the Medicaid program) received 12 (6.6 percent) of the compliance letters, none of the vendor hold actions, and between 0 and 10 percent of the contract terminations.

In sum, in terms of code violations, nonprofit nursing homes perform much better than proprietaries. The relative performance of nursing home chains is more difficult to evaluate, although there are indications that chains may differ significantly from one another and that some chains do well in terms of violations of structural health and safety, standards while others do quite poorly. For example, when we analyzed the 1983 Texas data on punitive actions, we found that

1.

Beverly Enterprises owns/operates 121 or 12.7 percent of the facilities participating in the Texas Medicaid program; however, Beverly received 18.2 percent of all compliance letters, 17 percent of the vendor holds, and 14.3 percent of the contract terminations (decertification).

2.

ARA owns 114 or 11.9 percent of the facilities and received 7.2 percent of the compliance letters, 14.3 percent of the vendor holds, and 2 percent of the contract terminations (de-certification).

3.

Hillhaven owns/operates 16 facilities (1.6 percent) and received 1.6 percent of the compliance letters and none of the vendor hold or decertification actions.

Thus, the chains varied in terms of violations cited and serious punitive actions instituted by the state. This suggests that nursing home chains are not a "monolith"—that is, each chain may have a slightly different orientation or organizational structure that affects its performance more than the mere structural characteristic of being a multifacility operation.

Variations in Process and Outcome Measures of Quality

Only a few studies have attempted to measure variations in quality using process and outcome measures (Kane, 1983; Schlenker et al., 1983; Schlenker, 1984). One that directly addresses the impact of ownership on quality focused on nursing homes in Iowa where the state licensure agency developed an "outcome oriented" licensure survey for use in the annual inspection of nursing homes (Lee, 1984). The survey focuses not only on resource inputs but also on resident satisfaction (e.g., feelings of safety and security, enjoyment of food); professional health care (e.g., health care planning for individual residents); implementing the medication, treatment, and diet plans as prescribed by physicians; the quality of the living environment (e.g., cleanliness of rooms); room conditions; and food service (e.g., whether the food provides basic nutritional requirements and whether residents like the food). In analyzing the differences in Iowa facilities, Lee (1984) found not only that proprietary facilities had significantly fewer staff than similar non-profits but also that nonprofits performed better on every outcome category: care planning, quarterly review of patients, room conditions, and quality of living environment.³⁵ The introduction of a "size" variable did not compromise the findings of the superiority of nonprofit facilities. Lee concluded that their findings are consistent with the Greene and Monahan study (1981) that ''nonprofit(s) appear to provide consistently better quality of care when compared with the for-profit facilities" (Lee, 1984).

Summary and Conclusions Regarding Quality

The preponderance of evidence from the relatively few studies that systematically address quality of care and ownership differences suggests the superiority of nonprofits—particularly of the church-related nonprofits. The data from the state surveys strongly support this finding. Studies using a variety of quality measures—resource inputs, licensure violations, complaints, and outcome-oriented measures of quality—are fairly uniform in finding nonprofit facilities superior in quality to for-profit nursing homes. The findings on the relative performance of the chains are less conclusive.

Because few studies have adequately controlled for the possible effects of differences in patient case mix, it is difficult to interpret data on resource inputs. Higher levels of staffing, for instance, may be a sign of higher quality of patient care, of a more intense case mix, or of inefficiency. Most studies have flaws, particularly in terms of the measures of quality (e.g., O'Brien et al., 1983). The nature and strength of the relationship between the proxies for quality and a patient-focused, multidimensional concept of quality have not been conclusively established. At best, such proxies as resource inputs are preconditions to, or partial indicators of, high quality of care. Until better measures are developed, tested, and used in studies, "knowledge" about the impact of ownership and other variables such as size is provisional.

Finally, researchers must use richer models of ownership. Lumping government-owned facilities with church-related homes and private nonprofits (that may be nonprofit in name only) into the same ownership category, may obscure significant differences in performance. The same is true of the for-profits. A multibillion dollar multistate chain of 800 nursing homes, whose stock is publicly traded, is hardly the same organization as a 40-bed facility owned and operated by a practical nurse and her family. Yet, both are for-profits. Finally, there may be significant differences among various multifacility nursing home systems, and understanding the performance of chains may require analyses that distinguish among them.

The Impact of Ownership on Cost

In general, empirical studies indicate that nonprofits report higher per patient per day costs than for-profit facilities. However, this finding requires qualification. First, most studies have employed inadequate measures of the severity of patient case mix and of quality of care. Thus, variations in cost may be due to differences in the facilities' mix of patients served or in the quality of patient care. Second, one of the most significant differences in costs is in the distribution of expenditures by different facilities, that is, in the way they allocate funds to various cost centers or items. This may have significant implications for quality and efficiency. Third, there is some indication that although certain types of facilities have lower costs, the actual price charged to patients varies relatively little.

Cost Function Studies and Behavioral and Market Models

There have been a variety of studies that focus on nursing home costs. Many of the most useful studies utilize secondary data from one or more states and relate differences in costs to such factors as type of ownership, location, bed size, occupancy rate, mix by payee, and measures intended to capture differences in patient case mix and quality of care. Most studies have found the costs reported by nonprofit facilities to be higher than those of similar proprietary homes; however, proxies for such key factors as patient case mix and quality have varied and generally been inadequate. Thus, the results must be interpreted carefully.

Two studies found no significant variations in costs associated with ownership. The New York State Moreland Act Commission (1975) found that voluntary facilities reported significantly higher costs per patient day ($16.02) on the average in 1974 than proprietary facilities ($14.12). However, when the analysts controlled for the size of the facility and regional wage differences, "very little difference in cost was explained by the ownership factor alone."

The Virginia Joint Legislative Audit and Review Commission's (1978) analysis of reported nursing home costs also found little difference between nonprofit and proprietary facilities, although, as discussed earlier, the study found significant differences in the allocation of expenditures, among various cost centers, with "nonprofit facilities . . . [having] lower facility costs (interest, rent, and depreciation) and administrative expenses but . . . [spending] more for nursing care and dietary services."

A few other studies found only slight differences in costs between proprietary and nonprofit nursing homes. Levey et al. (1973) found that higher costs were associated with nonprofit ownership, but the difference was not statistically significant. Winn (1974), on the basis of data gathered from nursing home administrators, found per diem costs to be slightly higher among nonprofits—but not significantly so.

Two other studies reveal significant cost differences. Both Holmberg and Anderson (1968) and Koetting (1980) found that the costs reported by nonprofit facilities were, on average, significantly higher than those of for-profits. Koetting found that for-profit facilities could achieve the same level of quality (as measured by licensure code violations and deficiencies in individual patient medical reviews) at a lower cost than nonprofit nursing homes. Both studies, however, found little difference between the two types of facilities in actual charges to patients. Holmberg and Anderson found, in fact, that once the analysis of the proprietaries included their profit factor, then charges by nonprofit facilities were lower.

Other studies have consistently found non-profits to have higher reported costs. Ruchlin and Levey (1972) studied 4 years of cost data (1965-1969) for 175 Massachusetts nursing homes and found nonprofit ownership generally had higher costs. Caswell and Cleverley (1978) in analyzing costs reported by Ohio nursing homes found that "proprietary homes do appear to be less costly than (nonprofit) . . . homes." Both noted that the allocation of expenditures within total spending differs, with the for-profits spending significantly more on property costs.

Birnbaum et al. (1981) analyzed data from several different data sets, including the National Nursing Home Survey and data on facilities participating in the Medicaid programs in New York, Massachusetts, and Indiana. They found nonprofit facilities (private and government-owned were both included) have higher costs than for-profit nursing homes. The analysis of each data set had some controls for variations in patient mix among the facilities, and in analyzing the New York data, Birnbaum et al. used inspection rating data as a proxy, for quality. Even when these types of case mix and quality measures were taken into account, the cost differentials between for-profit and nonprofit facilities persisted.

In her review of 12 major cost studies, Bishop (1980) concludes that for-profit facilities have lower per diem costs than nonprofit facilities, even when other factors associated with higher costs (such as location and the provision of skilled nursing care services) are held constant. Meiners (1982) also found that proprietary facilities had average costs that were significantly lower than nonprofit nursing homes, even when controlling for some measures of case mix, the range of therapeutic services available, and the type of staff coverage on daily shifts. Meiners also argued that whether a facility was part of a nursing home chain was not statistically significant in explaining variations in cost. Finally, Lee and Birnbaum (1983), Schlenker and Shaughnessy (1984), and Schlenker (1984) all report that nonprofit ownership is consistently related to higher costs.

Few available studies (e.g., Koetting, 1980; Lee et al., 1983; Meiners, 1982) directly address the impact of multistate chains on costs, and their data were from facility costs reported in the early 1970s. Further, we desired a more refined analysis of various ownership types. Thus, we undertook a new analysis of nursing home costs in Ohio. It examined variations among private nonprofit facilities (government-owned were excluded) and three types of proprietary facilities: (1) for-profit facilities owned by individuals who own and/or operate three or fewer facilities in Ohio; (2) intrastate chains, usually consisting of between four and ten facilities (all in Ohio) and owned by an individual, family, or partnership; and (3) multistate publicly held nursing home chains.³⁶

Consistent with much previous research, our analysis found substantial cost differences between nonprofit facilities and those operating under the three proprietary forms. The average total costs per patient day in nonprofit homes were approximately $6.00 to $8.00 higher than average total costs in various proprietary facilities. The average total costs for individually owned homes and for those that are part of intrastate chains were almost identical, while the total costs for homes operated by multistate chains were approximately $1.50 higher than the other two proprietary forms. The same pattern holds for average operating costs per patient day. However, in our analysis we also examined differences in the allocation of expenditures and, using multiple regression techniques, were able to gain a clearer picture of cost differences and the factors that affect them.

Our new Ohio analysis shows expenditures on patient care to be much higher in nonprofit homes than proprietary ones. Nonprofits spent approximately 27 percent more per patient day on raw food than for-profit homes, from 30 to 100 percent more on medical and rehabilitative services, from 250 to 1300 percent more on social services, and from 25 to 32 percent more than the proprietaries on total nursing (which includes RNs, LPNs, and nurse's aides and orderlies). Individually owned and intrastate facilities spent more than the nonprofits on LPNs, in effect substituting one kind of professional nurse for another.

If, as we argue, these measures of resource inputs and service availability are valid proxies for quality of care, Ohio nonprofit facilities offer higher quality of care on the average than do the proprietary facilities. The findings, some of which are displayed in Table 11, also suggest that in Ohio, the multistate chains perform somewhat better than the other two proprietary categories. The intrastate chains commit the fewest resources to direct patient care and thus may provide the lowest quality of care.³⁷

TABLE 11

Comparison of Average Expenditures in Major Cost Centers, Ohio, 1977.

As noted, we found that the difference in costs between the nonprofit and the for-profit categories was between $6.00 and $8.00 per patient day. The regression analysis made it clear, however, that much of that difference derives from differential spending on patient care (quality) and from differences in facility characteristics other than ownership (i.e., occupancy, admissions, patient mix by payer type, facility size, certification status, and location). The analysis also shows, however, that not all of the variation in costs between for-profit and nonprofit owners is related to these factors. The results indicate that controlling for different levels of patient care expenditures (our quality proxies) and different facility characteristics significantly reduced the cost difference; nevertheless, all three forms of proprietary facilities have total and operating costs that are still significantly lower than those found in the nonprofit facilities. For-profit facilities report costs per patient day that are approximately $2.50 lower than the nonprofits.

The analysis displayed in this section leads to several conclusions. First, nonprofit (non-government) facilities, on the average, report higher costs than the average for-profit facility. When we take into account factors other than ownership that might account for cost variations, the difference in costs between nonprofit and for-profit facilities is reduced, but even with these factors taken into account, there is still a significant cost difference between for-profit and nonprofit nursing homes in Ohio. Second, the differences between the various types of for-profits are quite interesting and may be missed by most analyses. When we used total or operating costs as the dependent variable, the three types of for-profit homes displayed little difference. Their behavior appeared quite similar. However, when we examine the for-profits' individual cost functions, their behavior differs; that is, the factors that are related to cost variations differ among the different ownership types. This is especially significant since it may indicate different responses among the various ownership types to regulatory and reimbursement policies. This issue deserves further study to determine whether it is more broadly generalizable. Finally, the variations between the different ownership types in relation to the quality, proxies are provocative. Like most other studies, our analysis shows significant differences between for-profit and nonprofit nursing homes in their expenditures on direct patient care and dietary items.

The Impact of Chains and Chain Acquisition of Facilities

The best way to study the impact of multifacility chains may be to analyze facilities' performance before and after being acquired by a chain, and no empirical studies have systematically attempted this. Some evidence, however, suggests that chain acquisition may lead to higher costs, even when no improvement in the quality, or level of services can be identified.

The facet of chain acquisition that has attracted the most attention is the impact on capital or property, costs. The real estate advantages of nursing home investment have been discussed by several observers (e.g., Shulman and Galanter, 1976) and earlier in this paper. The Ohio Nursing Home Commission (1979) found that there were ownership changes in 20 to 25 percent of for-profit facilities annually—usually to take advantage of various tax and property, cost reimbursement benefits. The result was that the private-pay patients and Medicaid program paid higher and higher costs for the same facility, to provide essentially the same services it had before the ownership change. Vladeck (1980) argues that such turnover causes "a disproportionately high share of industry, revenues . . . to [go to] depreciation, . . . interest premiums and personal profits" and may harm patient care. He argues

The high degree of leveraging also creates a kind of perpetual instability in the nursing home industry as a whole and within the life of any given institution. Nursing homes are always changing hands or changing financial structure; skimping and comer-cutting fluctuate with financing cycles . . . . The one economic certainty in the typical nursing home is that the monthly mortgage payment must be met. Expenditures on staff and supplies can wait, if necessary, but capital financing costs must be covered.

GAO (1983a, b) found that the acquisition of a hospital chain led to significantly higher property, costs being charged to the Medicare program. This also occurs in nursing homes. The Kansas Division of Post Audits found that acquisitions and mergers led to higher property costs. The audit division estimated the takeover of 58 ICFs by a large, multistate chain after March 1982 added $650,000 to the Medicaid program's costs in 1983 alone (Levinson, 1984). Another example is Beverly Enterprises' acquisition of Provincial House nursing homes in Michigan. According to one analysis of Medicaid cost reports, the pre-sale interest expenses reported by Provincial House nursing homes in 1981 totaled $1,814,296. After the acquisition by Beverly, these same facilities reported an interest expense of $4,280,212 as a result of the debt financing of the purchase, a 1-year increase in interest expenses alone of 136 percent (AFL-CIO, 1983a).

Also troublesome is the possibility—suggested by the distribution of spending by the chains—that such increased costs may be covered by chains' reducing expenditures on patient care. The analysis of 1983 nursing home spending patterns in Texas demonstrates that the chains' total expenditures were almost identical to the statewide mean. However, they spend 9 percent more than the statewide mean on property costs and 8.2 percent more on administrative expenses, while spending 4.4 percent less on patient care and 7.4 percent less on dietary services (Texas Department of Human Resources, 1984; see also Table 10). Other data showing similar patterns (e.g., in Virginia) were reviewed earlier.

The large, multifacility chains have argued that their higher administrative costs contribute to improved administration and that their lower patient care costs and dietary expenditures are due to management and purchasing efficiencies. This is not supported by a variety of data discussed earlier, including the records of two chains operating in Texas that spend more than the statewide mean on administration, substantially less on patient care and dietary costs (Texas Department of Human Resources, 1984; see Table 10), and have received a high proportion of punitive actions for violations of minimum health and safety standards.

Two other studies also suggest that higher administrative costs among multifacility chains do not lead to greater operating efficiencies. A Virginia study (Virginia JLARC, 1978) found that homes that are subsidiaries of chains generally paid the home office management fees, which averaged $50,266 per facility in 1976 and added about $400 per year to the cost of each bed. However, these fees did not reduce overall administrative costs or produce savings in other cost categories. The study concluded, "It appears that chain ownership offers no substantial cost savings to either the patient or the Medicaid program." Similarly, the Ohio Nursing Home Commission (1979) found that facilities with management fees had higher "other" administrative costs.³⁸

Although there may be circumstances in which acquisition activity does not lead to higher costs—for example, if payers will not pay for property costs following any sale—available evidence of such cost increases is persuasive. Whether the increased facility costs that result from the chains' acquisition and merger activities are justified by improved quality or more equitable access is difficult to determine since empirical studies have not adequately addressed this issue. Further, since it appears that chains vary among themselves in their total costs, the distribution of their expenditures, their performance in terms of regulatory standards, and their admission policies, more sophisticated analyses are required to address these issues.

Summary and Conclusions Regarding Cost

Most studies find nonprofit homes to have higher reported costs than for-profit homes. Although controls for other factors that lead to higher costs reduce the magnitude of this difference in costs, they do not eliminate them. Even when proxies for quality are introduced, as well as measures of patient case mix intensity, the differences remain. Thus, some analysts, like Koetting (1980), argue that proprietaries are more efficient—that is, they can attain a given level of quality at lower cost than the nonprofits. Indeed, our analysis of the Ohio nursing homes could be seen as supporting this proposition.

This conclusion may be premature. Many studies use inadequate measures of variations in patient case mix intensity. This is critical, since case mix differences between facilities have been found to be the best predictor of variations in costs (Schlenker, 1984) and since some evidence suggests nonprofit facilities typically have a more intense case mix. Thus, case mix may account for some of the cost differences between nonprofit and for-profit facilities (Stassen and Bishop, 1983b; Schlenker et al., 1983; Schlenker and Shaughnessy, 1984). Also nonprofit facilities tend to have a higher proportion of private-pay patients, who tend to be more functionally dependent than Medicaid patients (Liu and Mossey, 1980; Schlenker and Shaughnessy, 1984). Thus, to the extent that prior studies have not captured all relevant dimensions of case mix variation, they may overestimate the cost differential associated with ownership per se.

In addition, the proxies for quality used in cost studies are at best only partial measures of quality and, at worst, are unrelated to either quality of care or quality of life. Host studies use resource inputs as proxies for quality, although studies have found a weak relationship between inputs and process and outcome measures of quality. Even code violations, the majority of which are reflections of deficiencies in structural and input standards, are inadequate measures of quality. Thus, some variation in costs may be associated with unmeasured differences in quality.

Finally, the few analyses that focused on "charges" by facilities—rather than only their reported costs—found little or no differences between for-profit and nonprofit nursing homes. Once the analysis included the proprietary firms' profit factor and total charges, the differences between ownership types diminished. While these results may be an artifact of the states studied or the time period, they raise interesting issues that should be addressed in future research before firm conclusions are reached about the economic efficiency of different ownership forms.³⁹

Discriminatory Practices: Patient "Creaming"

Discrimination takes many forms. Many nursing homes maintain separate waiting lists for individuals seeking admission—one for private-paying patients and another for Medicaid recipients.⁴⁰ Other homes require as a condition of admission that the patient and/or his or her family sign a contract promising to pay the facility, privately (at whatever rate the facility determines) for periods ranging from 6 months to 3 years before the facility will accept Medicaid as payment.⁴¹ For instance, a New Jersey task force estimated that 80 percent of the state nursing homes require fixed periods of private pay for up to 3 years (U.S. Senate Special Committee on Aging, 1984). Other facilities "discharge" or evict individuals when they exhaust their private resources and become eligible for Medicaid. Nursing homes have also evicted Medicaid recipients when a private-paying patient has sought admission (Ohio Nursing Home Commission, 1979; U.S. Senate Special Committee on Aging, 1984). Finally, ombudsmen charge that nursing homes also discriminate against Medicaid recipients in room assignments and services (U.S. Senate Special Committee on Aging, 1984).

Access to nursing home services is a particular problem for two groups: Medicaid patients and those individuals with "heavy care needs" (Greenless et al., 1982; Schlenker, 1984; Scanlon, 1980a,b; Feder and Scanlon, 1980, 1982; Cotterill, 1983; Institute of Medicine, 1981; U.S. Senate Special Committee on Aging, 1984). The National Summary of State Nursing Home Ombudsman Reports for the United States reported in fiscal year 1982 that discrimination against Medicaid recipients or potential Medicaid recipients was identified as a major problem in 21 states and was the fourth most frequently mentioned problem out of 74 problems cited by the ombudsman programs.

In addition to discrimination based on payment source, nursing homes often discriminate against those individuals most in need of nursing home care. The GAO (1983b) summarized 11 studies conducted since 1979 and concluded that severe access problems and discrimination were occurring on the basis of patient "handicap." Individuals who required especially heavy care or substantial "hands-on" care, such as those persons suffering from Alzheimer's Disease and other related disorders, were "backed-up" in hospitals awaiting nursing home placement—even when there were empty nursing home beds in the community.

Access problems axe particularly severe for individuals with dementia. Because such individuals are often defined as not requiring rehabilitative therapies or daily care from an RN, they generally do not qualify for Medicare or for "skilled care," which is reimbursed at a higher level in many states. They do, however, require substantial care and assistance with activities of daily living. This hands-on care and frequent supervision is costly for nursing homes even if provided by aides. As a result, nursing homes have an incentive to select from their waiting lists individuals with the fewest functional disabilities and with the least mental impairment. This incentive is particularly strong under reimbursement systems that provide the same Medicaid rate for every patient—regardless of the degree of physical and mental impairment.

Access problems of Medicaid recipients and heavy-care patients may become quality of care problems after they find nursing home placement. On the basis of an analysis of national data for 1969 and 1973, Scanlon (1980b) posits a "dual" market in which private-pay patients receive desired care (at the market price) while public-pay patients will receive care only after the private demand has been satisfied. The result is not only a dual market in terms of demand but also a dual market in terms of supply, with Medicaid patients having access to those homes least able to compete successfully for private-pay patients. Thus, it is not surprising that, as Schlenker (1984) notes: "Several . . . studies found higher quality . . . associated with a lower Medicaid share and lower overall occupancy rate, both of which suggest access barriers for Medicaid recipients."

The finding that Medicaid utilization rates are associated with lower quality of care is fairly consistent (Bishop, 1980). The Ohio Nursing Home Commission (1979) found that 75 percent of the "low quality" homes had very high percentages (80 to 100 percent) of Medicaid patients, while only one (4 percent) of the "high quality" homes had such a high percentage of Medicaid patients. The commission reported that

Both the Commission and the Ombudsman have received many complaints from relatives, hospital social workers, county (welfare) personnel, etc., about the difficulty of placing Medicaid patients in homes offering high quality care. Sadly, the Commission's study of a sample of homes providing either very good or very. poor care confirmed this testimony . . . . The Commission concludes that Ohio is facing the development of a two-class system of long-term health care, with Medicaid recipients having ready access to care in only a few of the best homes and thus being forced to become patients in the state's worst homes.

In effect, discrimination makes the operation of substandard homes financially viable. With severely limited choice, such patients often have no option but to enter whatever nursing home will accept them, even though it may provide undesirable quality of care, for as Beverly Enterprises noted in an annual report to the SEC, "There is little, if any, competition in price, services, or quality with respect to Medicare and Medicaid patients" (Beverly Enterprises, 1980).

Reasons for Discrimination

Private-pay patients are much more lucrative for nursing home owners because of Medicare and Medicaid payment limits. Some limits have to do with whether particular costs are allowable. In other cases, some states set "ceilings" on reimbursable costs or employ prospective or flat-rate systems to pay homes a set amount per patient day. Often that rate is based on some percentage of the average reported costs of similar facilities and is independent of any one facility's costs of providing care. A "reasonable" payment rate may be set, as in some states, at the 60th percentile or lower of average costs for all facilities in a "class." Thus, Medicare and Medicaid provide nursing homes with relatively little price flexibility, while nursing homes are able to charge much higher rates to private-pay patients (Montgomery Securities, 1983).

Health planning limitations and moratoria on new nursing home bed construction may also facilitate discrimination by contributing to very high occupancy rates in nursing homes in most parts of the country. Occupancy rates of better than 95 percent are the norm. In addition, there are few alternatives to nursing home care for those people who need long-term care on a daily basis. This combination of tight supply and a relatively inelastic demand gives nursing homes much flexibility in setting charges to private-pay patients and, for good quality homes, the ability to discriminate and still maintain high occupancy rates.

Price data for private-pay patients, as compared to Medicaid and Medicare patients, is difficult to determine (Lane, 1984). If charges reported on Medicare cost reports are indicative of private-pay rates, the differential may be between 13 and 18 percent, although some estimates of the differential are as high as 20 percent (Lane, 1984). Montgomery, Securities, using more current data, estimates that private-pay rates, on average, range up to 30 percent above Medicaid and Medicare rates. Based on its survey of nursing homes, Montgomery Securities (1983) reports the 1982 rates as shown in Table 12.

TABLE 12

Nursing Home Rates by Payee for 1982.

The analysts conclude private-pay patients are central to a nursing home's profitability. Thus, nursing homes have a strong financial incentive to discriminate against Medicaid patients in favor of private-pay patients. The Montgomery Securities analysts comment The higher the percentage load of private pay patients in the home's mix, the greater the nursing home's ability to cost and price shift, and the greater the prospects of increased profits (Montgomery Securities, 1983).

Ownership and Discrimination

Evidence suggests that all nursing homes that can attract private-pay patients discriminate in their favor and against Medicaid patients. Nonprofit homes and facilities that are part of some of the major proprietary nursing home chains seem most successful in achieving a relatively high percentage of private-pay patients. Most empirical analyses have found that nonprofit facilities have a lower than average Medicaid utilization rate. While Winn (1974) found no significant relationship between type of ownership and rate of public assistance patients in Washington state nursing homes, most other analyses have found an association between nonprofit status and a lower Medicaid share. Gottesman (1974) found that proprietary facilities were more likely to serve Medicaid patients. Brooks and Hoffman (1978) found no difference in Medicaid utilization rates in SNFs based on ownership, but there was a significant difference in ICFs. Among the nonprofit ICFs in this study, the Medicaid utilization rate was 57 percent while in the proprietary homes the rate was 71 percent. Studies in Ohio (Ohio Nursing Home Commission, 1979) and California (Vladeck, 1980) show the proportion of Medicaid patients to be higher in for-profit than in nonprofit facilities. In Texas the pattern was the same, as Table 13 indicates.

TABLE 13

Average Percentage of Medicaid Days of Service by Owner Type, Texas (percentage).

Several of the major nursing home chains also have relatively low percentages of Medicaid recipients. For some, this is the product of a deliberate strategy. Care Enterprises (1984), for example, notes in its report to the SEC that

Care believes that rehabilitative services at its SNFs are relatively more profitable than other nursing services and have enhanced its ability to attract private and Medicare patients.

This chain reported that in 1982 the company was able to increase its revenues quite substantially over the prior year via increases in billing rates and changes in the mix of private, Medicare, and Medicaid patients. In 1982, Care Enterprises had an average Medicaid utilization rate of 60 percent, with 22 percent private pay and 16 percent Medicare.⁴²

Care Enterprises is not alone in seeking to increase its private-pay patient load and decrease Medicaid utilization. Hillhaven/National Medical Enterprises is also concentrating on keeping a "favorable" mix of private-pay patients (Punch, 1984). However, the most aggressive large chain in this regard is Manor Care. In Ohio, Manor Care attempted to evict individuals whose care was paid for by Medicaid during the late 1970s (Ohio Nursing Home Commission, 1979). More recently Manor Care has "targeted" its acquisitions, purchasing facilities that already have high private-pay utilization rates and selling off homes with too many Medicaid patients. In 1982, for instance, Manor Care sold 14 facilities because they did not have sufficiently high ratios of private to Medicaid patients. As Manor Care (1981) noted in its report to the SEC, "As a general rule, the margin of profits is higher with private patients." The report continued,

Manor Care attempts to locate and operate its nursing centers in a manner designed to attract patients who pay directly to the facilities for services without benefit of any government assistance program.

Evidence of its success in attracting private-pay patients and the greater price flexibility they have with these patients can be seen in the data displayed in Table 14.

TABLE 14

Proportion of Inpatient Days to Revenues by Payee—Manor Care (percentage).

Summary and Conclusions Regarding Access

There is substantial evidence that Medicaid recipients—some two-thirds of all long-stay nursing home patients—and those with heavy care needs experience discrimination in access to nursing homes and that such discrimination places Medicaid patients disproportionately in homes that provide the lowest quality of care. The evidence is also clear that nonprofit homes provide the fewest beds on the average to Medicaid patients and that some nursing home chains systematically seek to reduce the proportion of Medicaid patients in their facilities.

Further, there is evidence that suggests discrimination and that a facility's patient mix has quality implications. Caswell and Cleverley (1978, 1983) argue that there is a differential effect of ownership control status on the relationship of Medicaid utilization to cost and quality. They found that, in general, increasing Medicaid utilization in a facility is associated with lower costs, particularly in patient care expenses. However, they found that for-profit facilities react more strongly than non-profits to increasing Medicaid utilization. The nonprofits appear to reduce patient care expenditures far less than do for-profits as the proportion of Medicaid patients in the facility increases. Thus, while the evident discrimination is a matter of concern in terms of freedom of choice for consumers, it may also negatively affect the health, well-being, and quality of life of vast numbers of the aged and disabled.