Medicare-covered skilled nursing facility services, 1967-88.

The skilled nursing facility benefit under Medicare has been difficult to administer because its intent has been subject to misinterpretation. This article describes the series of legislative and administrative actions taken to align the benefit's use with its intent. Data are presented to show the changes in utilization and program expenditures in response to the actions taken. The 1988 clarifications to the level-of-care requirements seem to have resulted in an increased level of use of skilled nursing facility services.

Introduction

In this article, data are presented on the use of and expenditures for Medicare-covered skilled nursing facility (SNF) services during the period from 1967 (the first full year of Medicare) through 1988. SNF services probably have been subject to more misunderstanding and required greater efforts of administrative interpretation and clarification than any of the other major Medicare benefits. One major reason for this is that the principal loci of these services are nursing homes. The usual perception of nursing homes is that of a place where old people who are no longer able to manage independent living go to receive care and supervision for the rest of their lives, a period that could last many years. SNF services as embodied in the Medicare legislation were conceived in a very different context.

The Medicare benefit structure was developed to provide services to treat acute illnesses. Treatment was conceived as going through stages along a continuum of care. In this model, the framers of the Medicare legislation created a new medical entity to continue treatment of an illness that began with a hospital admission—the extended care facility (ECF). In the Medicare context, the term “extended” referred not to provision of care over an extended period of time, but to the provision of recuperative care in a less intensive facility as an extension of inpatient hospital care. The intent was to provide an alternative to continued hospital care for patients who still required general medical management and skilled nursing care on a continuing basis, but who did not require the constant availability of physician services ordinarily found only in the hospital setting.

Many nursing homes readily made the necessary adaptations to facilities and staff to become eligible to provide ECF services to Medicare beneficiaries, but understanding the purpose and the nature of the services covered by the ECF benefit took longer. Eventually, emphasis on the nature of the services to be provided at this stage of the continuum rather than on the temporal sequence led to changing the name of the benefit to its present name, skilled nursing facility services, in the 1972 Amendments to the Social Security Act.

This article focuses on the changes in legislation and regulations that helped to sharpen the definition of the SNF benefit and assure its implementation in accordance with its intent. I will highlight how Medicare, in the areas of utilization and expenditures, responded sensitively to these changes. Because of the explicit limitations on what services are covered by the Medicare SNF benefit, its place in the provision of long-term care services to nursing home residents will be put into perspective. The data in Table 1 highlight the small role played by the Medicare SNF benefit in the provision of nursing home services in the United States. The use of and expenditures for Medicare-covered SNF services from 1967 through 1988 are shown in Table 2. The discussion of the utilization and expenditure data in this table will highlight the impact of administrative and legislative efforts to define the SNF benefit. The data presented in this table are differentiated by services rendered to the aged or to disabled Medicare beneficiaries. Inasmuch as the Medicare SNF benefit is used mostly by the aged, the discussion of trends and use patterns focuses on the data for the aged.

Table 1

Amount and percent distribution of payments for nursing home services to persons 65 years of age or over, by source of payment: Calendar years 1977 and 1987

Source of payment	1977		1987
	Amount in billions	Percent distribution	Amount in billions	Percent distribution
All sources	$10.5	100.0	$32.8	100.0
Private1	5.0	47.6	19.2	58.5
Public	5.5	52.4	13.6	41.5
Medicare	0.3	2.9	0.6	1.8
Medicaid	4.6	43.8	11.9	36.3
Other2	0.6	5.7	1.1	3.4

Includes expenditures by individuals (i.e. “out-of-pocket” expenditures), private insurance, and philanthropic organizations.

Includes expenditures by the Veterans Administration, State and local facilities, workers' compensation, and other Federal, State, and local programs.

SOURCE: (Waldo et al., 1989.)

Table 2

Persons receiving reimbursed skilled nursing facility (SNF) services, covered SNF days of care, and Medicare program payments for SNF persons, by type of beneficiary: Calendar years 1967-88

Year	Medicare HI enrollment in thousands	Persons served		Covered days of care			Total Medicare in millions	Program payments
		Number	Per 1,000 HI enrollees	Number	Per person served	Per 1,000 HI enrollees		Total in millions	Percent of Medicare	Per covered day	Per person served	Per HI enrollee
Aged
1967	19,494	354,200	18.2	19,997,000	56.5	1,026	$4,711	$313	6.6	$15.65	$ 883.68	$16.06
1968	19,770	400,700	20.3	21,340,000	53.3	1,079	5,905	402	6.8	18.84	1,003.24	20.33
1969	20,014	393,700	19.7	17,572,479	44.6	878	6,717	371	5.5	21.11	942.34	18.54
1970	20,361	NA	—	10,697,188	—	525	7,279	245	3.4	22.90	—	12.03
1971	20,742	238,700	11.5	7,481,066	31.3	361	8,077	195	2.4	26.07	816.93	9.40
1972	21,115	221,920	10.5	6,628,063	29.9	314	8,862	180	2.0	27.16	811.10	8.52
1973	21,571	250,160	11.6	8,523,049	34.1	395	9,790	231	2.4	27.10	923.41	10.71
1974	21,996	258,260	11.8	8,688,302	33.6	395	11,565	264	2.3	30.39	1,022.23	12.01
1975	22,472	259,520	11.6	8,584,652	33.1	382	14,030	280	2.0	32.62	1,078.91	12.46
1976	22,920	286,020	12.5	9,399,233	32.9	410	16,683	330	2.0	35.11	1,153.77	14.40
1977	23,475	282,540	12.0	9,296,911	32.9	396	19,508	338	1.7	36.36	1,196.29	14.40
1978	23,984	267,300	11.1	8,755,325	32.8	365	22,393	341	1.5	38.95	1,275.72	14.22
1979	24,584	246,910	10.0	8,271,820	33.5	336	26,156	357	1.4	43.16	1,445.87	14.52
1980	25,104	247,800	9.9	8,410,409	33.9	335	31,658	395	1.2	46.97	1,594.03	15.73
1981	25,591	242,840	9.5	8,374,683	34.5	327	37,985	433	1.1	51.70	1,783.07	16.92
1982	26,115	243,860	9.3	8,554,699	35.1	328	44,480	474	1.1	55.41	1,943.74	18.15
1983	26,670	256,500	9.6	9,010,052	35.1	338	50,883	507	1.0	56.27	1,976.61	19.01
1984	27,112	289,820	10.7	9,314,025	32.1	344	57,199	542	0.9	58.19	1,870.13	19.99
1985	27,683	304,360	11.0	8,627,412	28.4	312	61,945	558	0.9	64.68	1,833.36	20.16
1986	28,257	293,700	10.4	7,868,018	26.8	278	67,063	557	0.8	70.79	1,896.49	19.71
1987	28,822	282,640	9.8	7,144,442	25.3	248	72,543	605	0.8	84.68	2,140.53	20.99
1988	29,312	371,240	12.7	10,448,847	28.2	356	79,010	936	1.2	89.58	2,521.28	31.93
Disabled
1974	1,928.1	7,934	4.1	277,025	34.9	144	$1,158	$9	0.8	$32.49	$1,134.36	$4.67
1975	2,168.4	8,386	3.9	289,018	34.5	133	1,578	10	0.6	34.60	1,192.46	4.61
1976	2,392.2	9,420	3.9	315,827	33.5	132	2,078	13	0.6	41.16	1,380.04	5.43
1977	2,619.4	9,520	3.6	334,823	35.2	128	2,601	14	0.5	41.81	1,470.59	5.34
1978	2,793.2	9,177	3.3	320,708	34.9	115	3,173	14	0.4	43.65	1,525.55	5.01
1979	2,910.8	9,221	3.2	321,851	34.9	111	3,841	15	0.4	46.61	1,626.72	5.15
1980	2,963.2	8,504	2.9	319,172	37.5	108	4,599	16	0.3	50.13	1,881.47	5.40
1981	2,999.0	8,272	2.8	310,035	37.5	103	5,480	17	0.3	54.83	2,055.13	5.67
1982	2,954.2	7,788	2.6	296,232	38.0	100	6,516	17	0.3	57.39	2,182.85	5.75
1983	2,917.6	7,920	2.7	304,848	38.5	105	7,015	18	0.3	59.05	2,272.73	6.17
1984	2,884.4	8,912	3.1	314,511	35.3	109	7,396	19	0.3	60.41	2,131.96	6.59
1985	2,906.9	10,292	3.5	305,066	29.6	105	7,772	20	0.3	65.56	1,943.26	6.88
1986	2,958.5	10,264	3.5	295,167	28.8	100	8,199	21	0.3	71.15	2,045.99	7.10
1987	3,030.7	9,920	3.3	272,125	27.4	90	8,390	24	0.3	88.19	2,419.35	7.92
1988	3,101.5	13,000	4.2	389,557	30.0	126	8,883	35	0.4	89.85	2,692.31	11.28

NOTES: NA is not available. HI is hospital insurance.

SOURCE: Health Care Financing Administration: Office of the Actuary and Bureau of Data Management and Strategy, Office of Statistics and Data Management.

Eligibility (level of care) criteria

A Medicare beneficiary is eligible to receive SNF services when the following conditions are met:

The patient has had a medically necessary stay in a Medicare participating hospital for at least 3 consecutive days prior to admission to an SNF.

The patient is admitted within a short period of time (generally, within 30 days) of discharge from the hospital. If the patient is subsequently discharged from the SNF, he or she may be readmitted to a participating SNF within 30 days of discharge without re-entering a hospital.

The patient receives care for a hospital-related condition (i.e., one for which the patient was treated in the hospital), or for a condition that arose while he or she was being treated in the SNF for the hospital-related condition.

The patient requires skilled nursing services or skilled rehabilitation services (i.e., services that must be performed by or under the supervision of professional or technical personnel).

The patient requires these skilled services on a daily basis.

As a practical matter, considering economy and efficiency, the daily skilled services can be provided only on an inpatient basis in an SNF.

The services furnished must be pursuant to a physician's orders and be reasonable and necessary for the diagnosis or treatment of the patient's illness or injury.

Covered services

If the eligibility criteria are met, the Medicare hospital insurance (HI) program (Part A) pays for up to 100 days of SNF care per benefit period. After prior use of HI inpatient benefits, a new benefit period begins when the enrollee has not been an inpatient in a hospital or a facility that provides skilled nursing care (even if not receiving Medicare-covered SNF services) for 60 consecutive days. For the first 20 covered days in an SNF during a benefit period, the HI program pays the cost of all covered services. From days 21 through 100, the beneficiary is liable for a coinsurance payment equal to one-eighth of the current Part A hospital deductible. In 1988, the Part A deductible was $540 and the SNF coinsurance was $67.50 per day.

If the services required by the SNF patient meet the level-of-care criteria listed earlier, the services covered by the Medicare program include a semi-private room, meals, regular nursing services, rehabilitation services (i.e., physical, occupational, and speech therapy), drugs furnished by the facility during the stay, medical supplies, and the use of appliances. Services considered beyond the scope of Medicare coverage include personal convenience items (such as television and telephone), private duty nurses, extra charges for a private room (unless needed for medical reasons), and the first three pints of blood in a benefit period. Physicians' services furnished to a beneficiary in an SNF are not covered under the SNF benefit, but are covered under supplementary medical insurance (Part B of Medicare).

Medicare policy for reimbursement

SNFs are reimbursed on the basis of reasonable costs incurred subject to limits. Medicare payments to SNFs are based on the determination of:

The SNF's total costs for items and services allowed by Medicare.

The portion of total allowed costs attributable to Medicare patients.

Whether the resultant amount is “reasonable” (whether it exceeds a prospectively determined limit established by a method promulgated in regulations and summarized in the next paragraph).

The amount that Medicare pays to an SNF, based on its reported costs, is determined through a process of cost allocation and cost apportionment. An SNF's overhead costs (e.g., depreciation, housekeeping, etc.) are allocated to the SNF's direct or revenue-producing cost centers (e.g., nursing services, laboratory, therapy services, etc.) through accounting processes such as “step down” or other methods agreed upon by Medicare and the SNF. After the allowable direct and indirect costs of the revenue-producing cost centers are determined, these costs are apportioned between Medicare and other users of the SNF. For routine services (i.e., room, board, and nursing), total allowable routine costs are divided by total inpatient days of care to determine a per diem rate. Medicare reimbursement for routine care services is the product of the per diem rate and the number of covered days used by Medicare patients.

The Medicare share for the costs of ancillary services is apportioned on the ratio of Medicare charges to all charges applied to total costs. This ratio is determined for each ancillary cost center (e.g., drugs, radiology, laboratory, etc.). The ratio is applied to the total allowable costs accumulated in each cost center during the fiscal year to compute the Medicare share of the costs.

Section 1861(v)(1) of the Social Security Act, enacted as section 223 of the Social Security Amendments of 1972 (Public Law 92-603), authorizes the establishment of limits on allowable costs that will be reimbursed under Medicare. These limits would be based on estimates of the costs for the efficient delivery of needed health services. Under this authority, limits on SNF per diem routine inpatient costs have been published since 1979. Section 1888 of the Act, enacted as section 2319 of the Deficit Reduction Act of 1984 (Public Law 98-369), authorizes the establishment of separate limits on per diem inpatient routine service costs for hospital-based and freestanding SNFs by urban or rural location.

The methodology for establishing the routine service costs limits for 1988 was published in the Federal Register (1987). The freestanding SNF cost limits were set at 112 percent of the average per diem labor-related and nonlabor costs for the urban or rural area in which the SNF was located. The cost limits for hospital-based SNFs were set at the limit for freestanding SNFs plus 50 percent of the difference between the freestanding limit and 112 percent of the average per diem routine service costs of hospital-based SNFs in the same area.

These eligibility, coverage, and reimbursement provisions govern the SNF services furnished in 1988. The SNF provisions went through many legislative and administrative modifications that affected the use of and program payments for SNF services during the period from 1967 through 1988. These will be discussed when trend data are presented, particularly in relation to Table 2.

Long-term care

Medicare covers nursing home services only when furnished by a participating SNF, only when the earlier-noted eligibility conditions are met, and only to the extent of the time and reimbursement limitations allowed by law. In short, the Medicare SNF benefit is neither designed nor intended to pay for long-term or custodial care in a nursing home. After Medicare SNF coverage is terminated because the coverage conditions are no longer met or the benefits have been exhausted, the cost of continued care must be met by the patient, the family, or other third party payers.

The Medicaid program, which pays for the costs of health care services to the poor, is the largest source of third-party payments for nursing home services. During the period covered by this article, Medicaid covered two levels of nursing home care: skilled nursing care (analogous to the Medicare SNF benefit) and intermediate nursing care, which provides a less intensive level of care. In neither case have the States placed limits on the number of days they would cover under Medicaid.

The distribution of payments for nursing home services by source of payment is shown in Table 1 for the years 1977 and 1987. The Medicare share was only 2.9 percent in 1977, and it decreased to 1.8 percent in 1987. During the same period, the Medicaid share, while still remaining the largest source of public funds for nursing home services, decreased from 43.8 to 36.3 percent. The share contributed by private sources, mainly “out-of-pocket” payments by the nursing home residents or their families, had a significant increase during this period from 47.6 to 58.5 percent.

If one were to apply the age distribution of nursing home residents as found by the 1985 National Nursing Home Survey (Hing, 1987) to the Medicare enrollment of 1987, approximately 1.42 million aged persons were nursing home residents in any given month during 1987. If one were to round out the Table 1 estimate of 1987 out-of-pocket payments for persons 65 years of age and over at $18 billion and assume an equal monthly distribution of such payments, the average monthly out-of-pocket payment was about $1,056 per aged resident.

At the time of admission, about one-half of the residents relied on their own or family resources to pay for their care. Thus, for self-paying new admissions, out-of-pocket payments amounted to about $2,100 per month. At this rate of payment, it would not take long for many elderly residents to exhaust their capability to continue paying for care from their own resources. At that point, Medicaid becomes the principal source of payment for continued care. It should be further noted that Medicaid is already the principal source of payment for 40 percent of newly admitted elderly. Medicare is the primary source of payment for only 5 percent of new nursing home admissions (Hing, 1987). In short, then, Medicare is neither a major source of nursing home payments nor a significant source of relief from the cost of long-term care of most nursing home residents. When private resources are exhausted, Medicaid is the more likely source of funds for the continued maintenance of long-term care.

Legislative history and benefit trends

The ensuing discussion is based on the data for the aged shown in Table 2. During the period from 1967 through 1969, the use of SNF services far exceeded the level projected by actuaries during the development of the Medicare legislation (U.S. Senate Committee on Finance, 1970). Actuarial projections for 1967, the first year that the benefit would be available, estimated total Medicare payment for SNF services at about $30 million. Instead, SNF payments in 1967 were $313 million, ten times the projection. In 1968, SNF payments rose to $402 million. It was apparent that the SNF benefit was being misinterpreted to provide long-term nursing home care beyond the scope intended by the legislation. In April 1969, the Bureau of Health Insurance issued Intermediary Letter No. 371, “Determining coverage of care in an extended care facility” (Social Security Administration, 1969). This intermediary letter re-emphasized the legislative intent of the SNF benefit as previously described and, to promote uniformity among intermediaries in making coverage determinations, provided a list of services that would (under normal circumstances) be covered.

The effects of the letter were immediate. All measures of SNF benefit utilization and program payments decreased significantly. In 1970, the days of SNF care used were one-half of that in 1968. Reimbursements dropped from $402 million in 1968 to $245 million in 1970. The SNF share of total Medicare program payments was cut in half, from 6.8 percent in 1968 to 3.4 percent in 1970.

The downward trends in SNF benefit utilization and payments began with the issuance of the intermediary letter and continued until the SNF provisions in the Social Security Amendments of 1972 reversed them. In addition to extending Medicare coverage to disabled persons who have been receiving social security cash benefits for more than 24 months and to persons with end stage renal disease, the 1972 Amendments provided for a uniform definition of SNFs under both Medicare and Medicaid and liberalized the Medicare concept of SNF care, effective January 1, 1973 (U.S. Department of Health, Education, and Welfare, 1973).

Prior to 1973, a patient had to need skilled nursing care on a continuing basis to qualify for SNF benefits. The 1972 Amendments made the following changes in the conditions required for Medicare coverage:

Altered the required frequency of skilled nursing care from “continuing” to “daily.”

Recognized a need for skilled rehabilitation services as a basis for coverage of care.

Accepted the concept of skilled management being needed for an “aggregate” of unskilled services.

Introduced the “practical matter” concept; that is, the availability of alternative health care facilities and services and the patient's condition are taken into account to determine if the patient's need for care and supervision justifies SNF care.

The use of SNF services increased significantly from their 1972 levels and remained relatively stable from 1973 to 1977. The user rate hovered around 12 per 1,000 aged enrollees, up from the rate of 10.5 per 1,000 enrollees in 1972. The total days of care used per 1,000 enrollees ranged from 395 to 410, up from the rate of 314 days per 1,000 enrollees in 1972. However, as a share of the Medicare dollar, SNF program payments decreased from 2.4 percent in 1973 (after rising from 2.0 percent in 1972) to 1.7 percent in 1977 as the costs for other services rose more rapidly, particularly costs of hospital services.

The 1972 Amendments also contained provisions that affected the administration of the SNF benefit. Their impact was not apparent until 1978, when they set off a downward trend (with intermittent fluctuations) in the rate of use of SNF services and an accelerated rate of decrease in the SNF share of the Medicare dollar that were not definitively reversed until 1988. The 1972 Amendments authorized the development of periods of minimum coverage by diagnosis; that is, Congress permitted Medicare to determine prospectively the number of days for which a patient with a particular diagnosis could expect coverage in an SNF. Stays beyond the minimum period were to be subject to the regular review procedures. This provision, known as “presumed coverage,” was never used to any significant extent and was formally rescinded in 1980. However, protection from retroactive coverage denials has been provided through the policy of “waiver of liability.” The 1972 Amendments waived a beneficiary's liability for payment for services received but deemed not covered when the beneficiary was unaware that the services would not be covered. Liability would pass to the provider unless the provider, also, had no reason to know that the services were not covered, in which case Medicare would pay for the services.

A provider could qualify for a favorable waiver presumption if it usually made accurate coverage decisions. Since 1978, an SNF would be granted favorable waiver status as long as no more than 5 percent of the days billed by the SNF as covered in the previous quarter were found by the intermediary to be noncovered. Prior to 1978, a 10-percent denial rate was permitted without loss of the favorable waiver presumption.

It has been asserted that the more stringent criterion begun in 1978 led some SNFs, out of a concern to maintain favorable waiver status, to avoid filing Medicare claims for which coverage may be uncertain. In those instances, the SNF billed the patient directly or another third party, such as Medicaid. As shown in Table 2, for aged enrollees, the user rate decreased from 12.0 per 1,000 enrollees in 1977 to 11.1 per 1,000 enrollees in 1978. The days of care rate decreased from 396 days per 1,000 enrollees in 1977 to 365 days per 1,000 enrollees in 1978. Except for a brief upsurge in 1983 and 1984 following the introduction of the prospective payment system (PPS) for hospital services, these two measures of SNF utilization continued their general downward trend through 1987, as did the SNF share of the Medicare dollar.

The introduction of PPS provided an incentive to hospitals to discharge patients as soon as their medical condition no longer required the continuous availability of physician services, because no additional revenue would be derived from retaining the patient. This was expected to result in an increased rate of admissions to SNFs for post-acute care. At the same time, however, the incentive could induce abuses resulting in premature discharges and inappropriate transfers to SNFs. Intensified review by intermediaries constrained further increases in Medicare-covered SNF admissions, contributing, after 1985, to decreasing numbers of Medicare enrollees using covered SNF services and, in 1987, to the lowest days-of-care rate in the history of the program—248 per 1,000 enrollees.

There was, however, another important factor affecting the use of Medicare-covered SNF services during this period. During the late 1970s and the 1980s, the increasing costs of hospital services affected the use of Medicare-covered SNF services through the rise in the HI deductible they induced. The HI deductible, paid upon the start of a benefit period, is based on the average Medicare cost of an inpatient day of care in a hospital. The SNF coinsurance payment beginning with the twenty-first day of Medicare coverage in an SNF is equal to one-eighth of the HI deductible. In 1988, as an example, the SNF coinsurance was equal to $67.50 per day. This amount was frequently greater than an SNF's normal charge. In such cases, the patient or the patient's family prefer to pay the full charge rather than seek continued Medicare coverage with the accompanying coinsurance liability.

For whatever combination of reasons, the decreasing rate of use of Medicare-covered SNF services led to concern that Medicare beneficiaries were being deterred from needed access to their SNF benefits. In addition, evidence of wide geographic and intermediary variations in the application of coverage criteria were cited by several courts, which ruled against intermediary denials of SNF coverage. To address these concerns, the Health Care Financing Administration, in early 1988, promulgated revisions to the Intermediary Manual governing SNF services. The revisions to the custodial care section were the first in 20 years, and the revisions to the SNF level-of-care section were the first in 12 years (Health Care Financing Administration, 1987).

The revisions represented a clarification of the guidelines to present more clearly the requirements for coverage in an effort to assure greater accuracy and consistency of coverage determinations among intermediaries. The guidelines emphasized the importance of clear documentation to justify a denial of a claim for coverage. The guidelines expanded the number of examples illustrating covered care to emphasize what is covered rather than what should be denied. Extensive training of intermediary personnel in the application of the new guidelines was undertaken.

The data for 1988 in Table 2 show the immediate impact of the revised guidelines and the augmented training of intermediary personnel. The number of persons using SNF services and the user rate were the highest since 1969. The number of SNF days of care was the highest since 1970, and the days of care rate was 44 percent greater than in 1987 and the highest since 1978. Program payments for SNF services were 55 percent higher in 1988 than in 1987. The SNF share of the Medicare dollar increased for the first time since 1973, to 1.2 percent from 0.8 percent in 1987.

It is believed that the experience of 1988 represents a new level of use of SNF services by aged Medicare beneficiaries. The anticipated effects of the expanded SNF benefits under the Medicare Catastrophic Coverage Act of 1988 (Public Law 100-360) were obviated by its subsequent repeal by the aptly titled Medicare Catastrophic Coverage Repeal Act of 1989 (Public Law 101-234).