Evaluation of the national swing-bed program in rural hospitals

Abstract

The Health Care Financing Administration (HCFA) implemented a swing-bed demonstration and evaluation program for rural communities in the 1970's. The demonstration substantiated the cost effectiveness of providing long-term care in small, rural, acute care hospitals. As a result, Section 904 of the Omnibus Reconciliation Act of 1980 (Public Law 96-499) authorized the national swing-bed program, allowing rural hospitals with fewer than 50 beds to provide Medicare-and Medicaid-covered swing-bed care. A congressionally mandated evaluation of the program was conducted and the national swing-bed program was found to be cost effective. In this article, HCFA's report and recommendations to Congress are summarized in the context of the evaluation findings. HCFA recommended that the program be continued and that consideration be given to extending the option to larger hospitals. In this regard, the Omnibus Budget Reconciliation Act of 1987 (Public Law 100-203) extended the program to include rural hospitals with up to 100 beds.

Introduction

The rural swing-bed program was enacted by Congress in the Omnibus Reconciliation Act of 1980 (Public Law 96-499). In passing this legislation, Congress envisioned it would encourage efficient and effective use of inpatient hospital beds for the delivery of hospital, skilled nursing, or intermediate care services to Medicare and Medicaid beneficiaries in rural areas. To determine the program's impact, Congress required the Secretary of the Department of Health and Human Services to submit a report describing the program's experiences. Section 904(c) specified that the report consider:

• The extent and effect of the program on the availability and effective and economical provision of long-term care services in rural areas.

• Whether such a program should be continued.

• The results from any demonstration projects conducted under the program.

• Whether eligibility to elect the swing-bed option should be extended to other hospitals regardless of bed size or geographic location where there is a shortage of long-term care beds.

In carrying out this mandate, the Health Care Financing Administration (HCFA) contracted with the Center for Health Services Research of the University of Colorado to conduct an evaluation of the program. Conclusions from the findings and the issues identified in the course of the evaluation are presented in this summary. Some of the issues identified in the course of the evaluation, HCFA's recommendations on the positions to be taken at this time, and plans for further monitoring and evaluation of the program's experiences are discussed. A summary of recent legislative developments is also provided.

Findings and issues

Hospital participation

By July 1986, about 40 percent of the eligible hospitals in rural areas were certified to provide swing-bed care. The total number of certified swing-bed hospitals represented approximately 15 percent of all Medicare-certified, short-stay general hospitals in the United States. Although the national swing-bed program began slowly in the early 1980's, its growth during the 3 years before 1986 resulted in 899 hospitals certified for swing-bed care.

Swing-bed hospitals are predominantly concentrated in the larger rural land areas of the midwestern States, although western, southern, and southeastern States also have relatively high participation rates. Nine of the 11 States with no participating hospitals in July 1986 were located in or near the northeastern section of the country, largely because fewer rural communities and land areas are located in this region. In general, swing-bed hospitals and the communities in which they are located tend to be characterized by lower acute care occupancy rates, a lower ratio of physicians to elderly persons, a larger elderly population, and fewer Medicare skilled nursing facility (SNF) beds per elderly than rural hospitals and communities not participating in the program.

Use of swing beds

In 1986, 97 percent or 872 of the 899 certified swing-bed hospitals were providing long-term care in swing beds. Based on 1985 utilization data, the participating hospitals averaged 50 admissions and 964 days per year of long-term care in swing beds, with an average length of stay of about 20 days. Approximately three-fourths of all swing-bed admissions were from acute care. Only about one-half of nursing home patients in the same communities were admitted from acute care. Of all swing-bed admissions from acute care, about two-thirds of them were from the acute care portion of the swing-bed hospital itself.

Medicare covered 49 percent of swing-bed days in 1985. The average Medicare patient's length of stay was 14 days. Medicare payments are made only in those instances where the services furnished to Medicare beneficiaries meet the same skilled nursing or rehabilitation care criteria required to reimburse SNF's. Medicaid, which covered swing-bed care in 22 of the 39 States with participating hospitals, covered only 8 percent of swing-bed days in 1985, with an average length of stay of 48 days. Medicaid payments may be made for either skilled nursing or intermediate care services, depending on the provisions of the State's benefit structure. Other payers, chiefly private pay, covered the remaining 43 percent of days. These patients had an average stay of 30 days.

General purposes served

Although a few hospitals provided more than 2,500 days of swing-bed care in 1985, more than 90 percent used swing beds to provide care to relatively short-stay, long-term care patients. For the most part, swing beds are used to provide subacute long-term care to patients who are more difficult to place in community nursing homes owing to their intense needs for medical and highly skilled nursing care. These patients are either discharged home or to community nursing homes after relatively brief stays in swing-bed hospitals.

In most instances, swing beds serve as holding beds until patients are sufficiently rehabilitated to return home or until nursing home beds become available in the community. At times, however, swing beds are used to fill other gaps in the long-term care delivery system in rural communities, especially when community nursing home beds are fully occupied. In such situations, swing-bed hospitals provide more traditional long-term care, such as that commonly found in nursing homes. Even in these circumstances, swing-bed stays appear to be considerably shorter than those of nursing homes.

Community retention and access

Earlier data from the Utah swing-bed demonstration program in the 1970's indicated that a per capita increase in the number of Medicare SNF patients receiving care in their home communities occurred as a result of the swing-bed approach. More than 75 percent of the nurses and more than 90 percent of the physicians in swing-bed communities believed that the program enhanced the retention of long-term care patients in their communities.

Although nursing home administrators and representatives of the nursing home industry have expressed concern that swing beds compete directly with nursing home beds, a partial exploration of this issue found no evidence of statistically significant decreases in nursing home occupancy rates. In fact, in one of the States with the largest number of swing-bed hospitals, nursing home occupancy rates actually increased in swing-bed communities between 1982 and 1985.

Cost to hospitals

Under the assumption that hospital beds exist primarily for the provision of acute care services, their use for long-term care takes advantage of the declining acute care occupancy rates and the surplus in hospital capacity. As such, based on special analyses of hospital cost reports, the cost of swing-bed care to long-term care patients was calculated as an incremental cost. This calculation was based on the additional cost of providing long-term care given that the beds and the associated resources already exist for the provision of acute care. The incremental costs for routine and for ancillary long-term care provided in swing beds in 1984 were estimated as $33 to $34 and $19 to $21 per day, respectively. These costs were below the estimated average per diem swing-bed revenues of $44 for routine care and $32 for ancillary services.

Cost to Medicare and Medicaid

It appears that a portion of Medicare use of hospital swing beds has been caused by the prospective payment system (PPS). The more specific issue of whether, under PPS, rural swing-bed hospitals have “gamed the system” by transferring patients to the SNF level of care to gain additional revenue cannot be definitively answered at this time. Available evidence does not indicate that this is a widespread practice. However, this issue is the subject of a study of the impact of PPS on the swing-bed program. The findings will be incorporated into the series of annual reports to Congress on the impact of the Medicare hospital prospective payment system.

There is consenus, however, that nursing home case mix did not change substantially in swing-bed communities before and after the implemenation of the swing-bed program. Site visits have tended to confirm that many of the SNF patients receiving care in hospital swing beds would have remained for longer periods as acute care patients prior to PPS and would have been discharged to urban SNF's at the present time had swing beds not been available. The use of swing beds to provide Medicare SNF care results in a per-day saving to Medicare of approximately $16. This estimate is based on 1984 data and the assumption that swing-bed patients would have been placed in equal numbers in freestanding and hospital-based rural SNF's in the absence of this program. The saving was greater to the extent that swing bed patients would have otherwise gone to urban SNF's or to newly constructed or expanded facilities.

The overall cost (including routine and ancillary services) to the Medicare program of providing SNF care in hospital swing beds in 1985 was $26 million. If the SNF care had been rendered in places other than rural hospitals' swing beds (i.e., freestanding SNF's or in a distinct-part SNF of a hospital), routine care would have cost an average of $16 more per day. The total Medicare cost for providing such care in rural communities in 1985 would have been $5 million more. Because Medicaid covered a substantially lower portion of swing-bed care, its 1985 annual cost for long-term care services provided to swing-bed patients is estimated at slightly over $2 million.

Volume thresholds

The incremental cost of routine long-term care provided in swing beds was found to increase as a function of total swing-bed patient days. The volume threshold was defined as the point at which the incremental cost of swing-bed care exceeds swing-bed revenues (i.e., it is no longer cost effective to retain hospital beds as swing beds, or where it would appear to be more reasonable to convert them to permanent nursing home beds). This volume threshold was found to range between 1,500 and 3,000 days of swing-bed care. (In this study, swing-bed routine cost for the average swing-bed hospital began to exceed swing-bed revenues at approximately 2,000 days of swing-bed care per year.) Although hospital circumstances and State-specific reimbursement rates may increase the break-even point even beyond 3,000 days, such thresholds would appear to exist for any hospital. If a hospital's swing-bed experience indicates a strong and stable demand for long-term care, such a demand clearly indicates a hospital has moved into the nursing home business. Its cost structure, staffing needs, and the care orientation associated with swing beds are basically the same as those for a nursing home when it reaches this threshold.

Case mix in swing beds

Swing-bed patients have substantially shorter stays and greater rehabilitation potential than do nursing home patients. They are less frequently characterized by typical long-term care problems, such as incontinence, impaired cognitive functioning, dependence in activities of daily living (ADL's), and related psychological or social problems. However, swing-bed patients tend to have more subacute problems, such as recovery from surgery, hip fractures within the past 6 weeks, shortness of breath, and the need for intravenous catheters.

In general, the long-term care needs served by swing-bed hospitals are substantially different from those served by community nursing homes. Swing-bed hospitals tend to treat patients with subacute problems that need more intense medical and skilled care, while nursing homes tend to treat patients with problems that are more typically seen in institutional long-term care settings. The average swing-bed patient appears to be at least 20 percent more costly to care for per day than the average nursing home patient. This case-mix difference appears to be one of the reasons why swing-bed care is not regarded as a substitute for (or in competition with) nursing home care in most locations.

Relative to home health patients, swing-bed patients are more dependent in ADL's such as bathing, dressing, and using the telephone. In addition, swing-bed patients are characterized by a somewhat more intense set of medical and skilled nursing needs than are those of home health patients in areas such as hip fracture, stroke, and conditions requiring intravenous catheters. Swing-bed patients are also more dependent than are home health patients in areas such as incontinence and mental status problems. The differences between swing-bed and home health patients in terms of subacute needs are not as substantial as those differences between swing-bed and nursing home patients. The greater dependency in physical and cognitive functioning on the part of swing-bed patients renders their need for continuous skilled nursing and medical care greater than the intermittent needs of home health patients.

The results, therefore, suggest a continuum of dependency and subacute problem intensity in case mix that, respectively, characterizes swing-bed, home health, and nursing home patients. Overall, nursing home patients tend to be more dependent in traditional measures of functioning than do swing-bed patients, who, in turn, are more dependent in functioning than are home health patients. Subacute care needs appear to be strongest among swing-bed patients; although when a certain level of rehabilitation has been reached, such patients can and should be discharged to home health care. In home health settings, certain types of medical services are more frequently provided on an intermittent basis to patients with subacute needs than are services associated with more traditional nursing homes. Nonetheless, home health care also relies on a reasonable degree of independence in physical and mental functioning. Therefore, especially in rural communities where the distances that home health nurses have to travel can be substantial, swing beds offer the opportunity for continuous medical and skilled nursing care for a relatively short institutional stay to subacute patients. Often, such patients are subsequently discharged to their homes with intermittent home health care to continue the rehabilitation process.

Quality of care

Adjusting for the case-mix differences just noted, swing-bed patients were discharged home sooner and more frequently than were nursing home patients. (The frequency of written discharge plans was found to be greater for swing-bed than for nursing home patients, based on samples of patients discharged from each setting.) This discharge pattern seems to reflect the stronger rehabilitation philosophy that accompanies the provision of acute and subacute care.

Based on the criteria used in this study, subacute nursing services were found to be provided moderately better in swing-bed hospitals relative to nursing homes. However, more traditional long-term care nursing services were found to be better in nursing homes. These differences tended to persist after adjusting for case-mix differences. Physician visits, X-rays, laboratory tests, and intravenous medications all occurred more frequently for swing-bed than for nursing home patients after adjusting for case-mix differences. In all, it appears that long-term care to subacute patients is provided at least as well and probably better in swing-bed hospitals than in community nursing homes in rural areas. However, it also appears that the care typically required for longer-stay, chronically ill, or disabled long-term care patients is provided better in community nursing homes.

Hospital and program administration

The reasons most frequently cited by hospital administrators for joining the program were to meet a community's needs for long-term care and to provide better continuity of care. Increased revenues and more efficient use of staff resources were also cited frequently as reasons for joining. Generally, the problems and difficulties associated with implementing the swing-bed approach at the hospital level declined in importance as hospitals gained experience with the program. Staff resistance was often cited as a major start-up problem, particularly from the nursing staff who were concerned that the hospital would become primarily a nursing home. Dissatisfaction with reimbursement was viewed as a significant start-up and ongoing problem by many swing-bed hospital administrators. Despite the fact that incremental cost appears to be covered by the current reimbursement structure, most hospital administrators felt payment was inadequate.

Other administrative entities such as HCFA regional offices, State certification agencies, State planning agencies, Medicare Part A intermediaries and Part B carriers, Medicaid fiscal agents, and peer review organizations (PRO's) generally required few resources to incorporate the swing-bed program into their operations. Most agencies shifted personnel to the swing-bed program, and usually no new employees were added. Few agencies reported any start-up or ongoing problems associated with the administration of the program. One of the most frequently mentioned problems by agencies was the misunderstanding of regulations and program requirements by hospital personnel.

Demonstration projects

No demonstration projects to test alternative arrangements for implementing the swing-bed concept have been carried out. There has been a great deal of interest in developing a swing-bed demonstration project in urban areas. HCFA and representatives of hospitals wanting to participate in such a demonstration project were unable to reach agreement on key issues. Although the evaluation study recommended experimentation in urban areas, because of this, further discussions were terminated. Based on the evaluation, however, HCFA plans to investigate the potential for developing new approaches toward the payment of acute and post-acute care in urban hospitals.

Policy issues and recommendations

This section addresses the questions specifically posed by the Congress in mandating this evaluation and discusses issues identified in the course of conducting the evaluation. HCFA's recommendations concerning the issues are listed. Following the list is a discussion of the rationale underlying the recommendations.

• The rural swing-bed program should be continued.

• At this time, eligibility to elect the swing-bed option should not be extended to urban hospitals. HCFA plans to explore alternative models for testing the payment of acute and post-acute care in urban hospitals.

• The current method for determining the rate of payment for routine long-term care in a swing-bed hospital should be retained.

• Ancillary services to patients receiving long-term care in a swing-bed hospital should continue to be reimbursed at cost.

• HCFA will continue to monitor the cost behavior of hospitals and nursing homes and the use of ancillary services in swing-bed hospitals. Alternative payment arrangements will be explored.

• HCFA will review current visit screens for physician services by place of service with the intent of developing consistent criteria specific to physician visits to patients receiving long-term care in swing-bed hospitals.

• Swing-bed hospitals furnishing more than 1,000 days of long-term care to patients with stays of 60 days or more should be required to meet all conditions of participation for SNF's.

• HCFA will undertake a review of the desirability of conducting regular surveys of long-term care services furnished in swing-bed hospitals.

• HCFA will draw on the growing experience of PRO's to determine the feasibility of developing guidelines governing the transition of patients from acute to post-acute care in swing-bed hospitals.

• Consideration should be given to extending the swing-bed option to larger rural hospitals; for example, to those hospitals with fewer than 100 beds.

Continuation of the program

Despite certain weaknesses and disadvantages of the rural swing-bed program that are usually restricted to individual communities or providers, the weight of the evidence gathered as part of the evaluation study supports a continuation of the national swing-bed program in rural hospitals. On balance, its most important attribute is that the program has increased access to cost-effective, long-term care in many rural communities throughout the country. It has been accepted by residents of rural areas and by health care professionals. Relatively few administrative difficulties have been encountered in its implementation. To eliminate the opportunity for small rural hospitals to provide long-term care in swing beds would be detrimental to many rural residents. It is doubtful that more than a small portion of Medicare SNF days of care provided in hospital swing beds would be eliminated by virtue of abolishing swing beds. Because of the incentives embedded in PPS, SNF admissions of Medicare patients to swing beds would probably translate into longer SNF stays in existing or newly constructed/converted SNF beds in rural and in more distant urban communities. This would be more costly to Medicare.

Eligibility to elect swing beds

Bed-size limits

At the present time, whether the hospital meets the current bed-size limit for eligibility to participate in the swing-bed program is largely determined by applying the prior year's acute care occupancy rate to the total number of licensed beds that are not special care unit or newborn beds. Although this approach is somewhat generous in the eyes of some regulators, the resulting marginal increase in the number of swing-bed hospitals compared with a more stringent method appears to be inconsequential. It does not appear appropriate to mandate a uniform method of determining eligiblity at the regional or State level. The individual circumstances of each region or State, in terms of the supply of and demand for subacute long-term care beds, can more readily be taken into consideration without a strictly enforced guideline for determining the applicable number of beds.

On the basis of this bed-size eligibility criterion, 2,236 rural hospitals were eligible to participate in the swing-bed program prior to the Omnibus Budget Reconciliation Act of 1987. There were an additional 1,023 hospitals in areas defined as rural by the U.S. Bureau of the Census. Many of these hospitals are located in rural communities where swing beds are not available and an unmet need for long-term care services appears to exist. Consideration of extending the swing-bed option to rural hospitals with fewer than 100 beds would, therefore, be appropriate. If the bed-size eligibility criterion was to be increased to include rural hospitals with fewer than 100 beds, it would increase the pool of hospitals eligible to elect the swing-bed option by 640 hospitals and increase the availability of long-term care services in rural areas.

Urban hospital eligibility

At the present time, urban hospitals are ineligible to elect the swing-bed option, regardless of size. The swing-bed approach has embedded in it the opportunity to “game the system” by discharging acute care patients to the skilled nursing level of care, thereby, gaining additional revenues beyond that provided by the diagnosis-related group (DRG) payment rate. However, the evaluation's findings suggested that the differences between urban and rural communities may lead to differential use of the opportunity to “game the system.” In urban areas, distances between patients' residences and providers are usually substantially shorter and consumers usually have a larger number of providers to choose from. Further, urban hospitals generally have a greater capacity to maximize revenues through more sophisticated means.

Although the evaluators recommended experimentation on the swing-bed approach in urban areas, HCFA felt that the previously stated considerations suggested a conservative posture in extending the swing-bed option to urban hospitals. For these reasons, HCFA recommends that the swing-bed option not be extended to urban hospitals at this time. HCFA plans to develop demonstration projects that test alternative methods of paying for acute and post-acute services, including combining both levels of care under one payment arrangement. These demonstrations would permit an assessment of the utility and cost effectiveness of expanding the swing-bed approach to urban hospitals.

Many urban hospitals, under the impetus of PPS, are converting part of their facilities to distinct-part SNF's to create “transitional” beds; so-called because they are used to facilitate the patient's transition between acute care and SNF and/or home care. From the viewpoint of Medicare's program costs, the conversion to distinct-part SNF's may be more costly than providing the swing-bed option, assuming that conditions to control “gaming” can be developed.

Reimbursement arrangements

Routine care

At the present time, Medicare per diem reimbursement for swing-bed routine care is determined separately for each State as the average statewide Medicaid reimbursement rate for the applicable level of care for the preceding year. On average, this rate covers the incremental cost of providing routine long-term care in swing beds. The current method of paying for routine care should be retained. HCFA will continue to monitor the relationship of payment rates based on the current methodology to changes in the relative behavior of nursing home and hospital incremental costs to determine whether, in the interest of equity, modifications of the current methodology are indicated. Concurrently, HCFA will continue to explore alternative payment arrangements through demonstration projects or through experience gained from nursing home reimbursement systems used in several States. These arrangements may include payments based on case mix, on total stay rather than on a per diem basis, or combining the cost of acute and post-acute care.

Ancillary services

At present, ancillary services to swing-bed patients are cost reimbursed by Medicare. As with reimbursement for routine care, the present reimbursement for ancillary care was found to cover the incremental cost of ancillary services provided to swing-bed patients. A potential problem that has surfaced, albeit in only a few settings, is the abuse of the present reimbursement methodology through the excessive provision of ancillary services. Medicare reimbursement for ancillary hospital services is a direct function of the volume of the services. Consequently, discharging patients as rapidly as possible from acute care to swing-bed care, and providing the ancillary services largely after discharge from acute care, results in maximizing ancillary reimbursement.

As mentioned, this post-acute overloading of ancillary services does not appear to be taking place in swing-bed hospitals except for a few individual facilities. Because abuse appears minimal at this time, HCFA recommends retention of cost reimbursement for ancillary services. However, HCFA will continue to monitor ancillary service use in swing-bed facilities and, concurrently, explore alternative arrangements for paying the cost of ancillary services. These arrangements may include setting limits per day or per year on payments for ancillary services to swing-bed patients. The feasibility of developing and testing alternative arrangements for combining routine and ancillary services in one payment scheme is an option that could be explored in demonstration projects.

Physician services

During the evaluation it was found that physicians visit their swing-bed patients with far greater frequency than they do their nursing home patients. Less than half the physician visits to long-term care patients in hospital swing beds appear to be covered by either third-party payers or by their patients. However, it appears that the greater attentiveness on the part of physicians had a significant positive impact on rehabilitation of post-acute patients. A pattern of wide variation was found in the number of physician visits to post-acute swing-bed patients allowed by Medicare carriers. In practice, the number of visits allowed to different types of patients are based on screens promulgated by Medicare for physician visits in nursing homes (generally, intermediate care facilities), SNF's, and hospitals. The evaluation found that the limits on swing-bed physician visits are more generous than those of physician visits for Medicare patients in certified SNF's. This appears reasonable in view of the greater intensity of care required by post-acute swing-bed patients. Limits on swing-bed physician visits that are closer to acute care physician visits appear warranted. However, the wide variation found in the limits may be inappropriate. HCFA will undertake a review of the situation, possibly adding the category of post-acute swing-bed patients to those categories for which routine screens for physician visits are used.

Quality of care and standards

Volume thresholds

Swing-bed hospitals and nursing homes in rural areas have evolved into serving two distinct, but partly overlapping, long-term care markets. Generally, following the acute hospitalization phase, swing-bed hospitals tend to treat those patients who might be characterized as “subacute.” These patients require more intense and skilled nursing care services to further their recovery and rehabilitation from illness. They are discharged, on average, within 20 days. The tendency for swing-bed hospitals to avoid traditional nursing home care was found to be rather pronounced. At the subacute phase, the quality of services furnished by hospitals was found to be better overall than those services furnished by nursing homes. On the other hand, nursing homes provide higher-quality, traditional, long-term care services.

The differences found between the two types of facilities in case mix and the ability to care appropriately for the different types of patients were anticipated (on the basis of the swing-bed demonstrations in the 1970's) in the regulations that implemented the swing-bed legislation. The conditions of participation for swing-bed hospitals were predicated on the assumption “… that patients in swing-bed hospitals are less likely to become long-term residents.” Accordingly, the regulations attempted to avoid imposing significant burdens on rural hospitals by requiring adherence only to those SNF standards that are necessary and appropriate to SNF patient care and do not duplicate existing hospital requirements, do not require extensive structural modifications, and are unnecessary in what is primarily a general routine inpatient hospital setting. In short, the regulations did not contemplate the swing-bed hospital as providing a significant amount of the traditional type of long-term nursing home care.

A few swing-bed hospitals provided care to persons who remained in the facility for 60 or more days. The evaluators found that these patients' needs are more akin to the “traditional” long-term nursing home patient. Swing-bed hospitals were found to be less capable of meeting these needs than were nursing homes. The evaluators, therefore, proposed the establishment of volume thresholds or levels of long-term care stipulating when a swing-bed hospital provides a significant amount of such care. At that level, special measures to assure quality of care need to be taken. HCFA agrees with this assessment and recommends that a swing-bed hospital that provides more than 1,000 days of long-term care (either at the skilled or intermediate levels of nursing care) to patients with stays of 60 days or more in 1 year should be required to meet all SNF conditions of participation.

To meet all conditions of participation as an SNF, the hospital may elect to create a distinct-part facility. If this option were chosen, the hospital's reimbursement for SNF services in the distinct-part facility would be based on incurred costs. However, despite the increased costs involved in meeting all SNF conditions of participation, some hospitals may prefer to retain the flexibility to use all of their beds for acute care although, as a swing-bed facility, they would still be paid on the basis of the previous year's Medicaid rates for “subacute” services. The hospital should be allowed to decide the course it elects to take.

A further threshold might be established at 2,000 days of long-term care to all patients. However, if the swing-bed option were extended to larger hospitals, these hospitals might be more likely to furnish 2,000 days or more of appropriate subacute care services to a larger number of patients. Extension of the threshold to 2,000 days is not being recommended at this time, although it can be reassessed in light of further experience.

Periodic review

The evaluators found greater State-to-State variations in certifying swing-bed hospitals as long-term care providers than they did in certifying nursing homes. The evaluators recommended that principles and guidelines should be established for conducting surveys in swing-bed hospitals that provide a significant amount of traditional nursing home care. HCFA will undertake a review of this issue in conjunction with the earlier recommendation on the establishment of volume thresholds and the requirement that PRO's review transfers between acute and long-term care in swing-bed hospitals as discussed in the next section. Thereafter, it will institute arrangements that will assure that the needs of long-term care patients are appropriately met in swing-bed hospitals.

Peer review organizations

HCFA's contracts with PRO's include a review of swing-bed hospitals; particularly, transfers between acute and post-acute levels of care. This review provides a mechanism for assuring that the transfer is medically appropriate. In addition, it provides a control on any “gaming” that might take place in making transfers to maximize total revenues derived from the DRG payments for the acute care phase and the payments for routine and ancillary services rendered during the post-acute phase.

At this time, the indicators for the appropriateness of these transfers are not always clear. As such, the determinations of appropriateness are made on the basis of clinical judgments. It is expected that more experience will increase the clinical base for developing guidelines governing the transition of patients from acute to subacute care in swing-bed hospitals. An ongoing study, which examines the impact of PPS on the swing-bed program, will address this issue further through analysis of PRO acute care admission, readmission, and transfer denial rates. Also, comparisons will be made of acute care readmissions from swing-beds relative to nursing home or SNF beds.

Recent legislative and regulatory developments

Several significant changes in the swing-bed program were enacted under the Omnibus Budget Reconciliation Act of 1987 (Public Law 100-203). The program was extended to hospitals with up to 99 beds. However, the newly eligible hospitals with between 50 and 99 beds have two restrictions imposed on them. First, no Medicare payment may be made for skilled nursing facility (SNF) services provided to a patient in a swing bed more than 5 days after an SNF bed becomes available in the locality of the hospital, unless the patient's physician certifies that the patient's transfer to that facility would not be medically appropriate. In the absence of such certification, the hospital's designation of the swing bed as an SNF bed would become ineffective at the end of the 5 days. The hospital could not charge the Medicare beneficiary for continued care thereafter unless it gives the beneficiary a readmission notice of noncoverage. Second, a hospital may not be paid for swing-bed services to Medicare beneficiaries after the number of Medicare covered days of extended care services in a cost reporting period exceeds 15 percent of the licensed bed days available at the hospital during the reporting period.

In addition, Congress mandated a report by February 1989 on peer review organization denials of swing-bed care. The report is to include recommendations on how to encourage participation in the swing-bed program by eligible (but not participating) hospitals that have low occupancy rates and are located in areas with an unmet need for long-term care. HCFA has contracted for the collection of data for this report.

Data sources

A number of data sources were used in this study. Although not all sources contributed directly to the results summarized in this article, they are included because they contributed indirectly by providing contextual information on issues related to the swing-bed concept. Selected data sources are discussed in several articles in the list of references.

Patient-level primary data were prospectively collected on-site at swing-bed hospitals and community nursing homes for five different samples of patients. These data were used to assess case mix, process and outcome measures of quality, resource consumption and service-use patterns for admission, discharge, and cross-sectional cohorts of patients.

Medicare cost reports were obtained for several years for 75 swing-bed hospitals and 75 comparison hospitals in 27 States. These were used to analyze routine and ancillary service costs and revenue data for swing-bed hospitals relative to comparison hospitals. Approximately 20 different types of surveys were administered by phone, mail, or during on-site visits throughout the course of the study. These surveys involved State hospital associations, swing-bed and comparison hospital administrators, State Medicaid agencies, State fiscal agents, State planning agencies, certification agencies, nursing home administrators, home health agencies, swing-bed physicians, other swing-bed staff, comparison hospital physicians, hospital directors of nursing, directors of nursing in nursing homes, Medicare Part B carriers, and Medicare Part A intermediaries.

A number of secondary data sets were used. The more important ones consisted of American Hospital Association survey tapes; the American Medical Association Physician Masterfile; the National Center for Health Statistics Master Facility Inventory; several of HCFA's files from its Medicare Statistical System (including Medicare enrollee data, the Medicare provider of service file, Medicare hospital claims data, and the Medicare SNF claims data); and U.S. Bureau of the Census population tapes. At 6-month intervals, the survey and certification branches of HCFA's regional offices were contacted to obtain the number of certified swing-bed hospitals in each State. Data collected as part of other studies (conducted by the University of Colorado Center for Health Services Research) of nursing home, swing-bed, and home health care were used for comparative purposes.