U.S. Department of Health and Human Services
Use of Functional Criteria in Allocating Long-Term Care Benefits: What Are the Policy Implications?
Mary E. Jackson, Ph.D., and Brian O. Burwell
This report was prepared under contract #HHS-100-88-0041 between the U.S. Department of Health and Human Services (HHS), Office of Family, Community and Long-Term Care Policy (now known as the Office of Disability, Aging and Long-Term Care Policy) and SysteMetrics. For additional information about the study, you may visit the DALTCP home page at http://aspe.hhs.gov/daltcp/home.shtml or contact the office at HHS/ASPE/DALTCP, Room 424E, H.H. Humphrey Building, 200 Independence Avenue, SW, Washington, DC 20201. The e-mail address is: webmaster.DALTCP@hhs.gov. The DALTCP Project Officer was Robert Clark.
The views and conclusions presented in this. paper are solely those of the authors and do not necessarily reflect the views of the Department of Health and Human Services. The authors alone assume responsibility for the accuracy of the information presented in the report.
TABLE OF CONTENTS
- CHAPTER 1. INTRODUCTION
- CHAPTER 2. THE ALLOCATION OF LONG-TERM CARE BENEFITS IN EXISTING PROGRAMS
- 2.1 Medicare Skilled Nursing Facility (SNF) Benefits
- 2.2 Medicare Home Health Benefit
- 2.3 Proposed Medicare Respite Care Benefit
- 2.4 Medicaid Nursing Home Benefits
- 2.5 Medicaid Home and Community-Based Waiver Services
- 2.6 Medicaid State Plan Home Care Services
- 2.7 Private Long-Term Care Insurance
- 2.8 Public-Private Long-Term Care Insurance Initiatives
- CHAPTER 3. FUNCTIONAL ELIGIBILITY CRITERIA IN CURRENT LEGISLATIVE PROPOSALS
- 3.1 Amendments to the Internal Revenue Code of 1986
- 3.2 Amendments to Title XVI (SSI) of the Social Security Act
- 3.3 Amendments to Title XIX (Medicaid) of the Social Security Act
- 3.4 Amendments of Title XVIII (Medicare) of the Social Security Act
- 3.5 Summary
- CHAPTER 4. ALLOCATING LONG-TERM CARE BENEFITS USING ADLS: WHAT ARE THE POLICY IMPLICATIONS?
- 4.1 No Existing Long-Term Care Program Allocates Benefits Solely on the Basis of ADL Measures
- 4.2 Estimates of the Functionally Impaired Elderly Vary Depending Upon the Data Source
- 4.3 National Surveys May Underestimate the Number of Elderly with ADL Limitations
- 4.4 Estimates of the Cognitively Impaired Elderly Are Even More Difficult to Derive
- 4.5 What Level of Functional Impairment Should Trigger Eligibility for Services?
- 4.6 Should Long-Term Care Benefits Be Rationed Based on Objective Measures of ADL and Cognitive Impairment or Should Benefits Take Into Account "Unmet Needs"?
- 4.7 What Are The Alternatives to Functional Eligibility Criteria?
- 4.8 Should the Same Functional Criteria Be Applied for Institutional Services and Community-Based Services?
- 4.9 Summary
The policy debate regarding the merits of an expanded Federal role in providing financial assistance to elderly individuals with long-term care needs continues. In this debate, there appears to be a growing consensus that if an expanded Federal role is enacted, long-term care benefits should be allocated to the elderly on the basis of objective functional criteria, particularly functioning in the Activities of Daily Living (ADLs). This consensus has evolved from a considerable body of research which has demonstrated that measures of functioning in Activities of Daily Living are an extremely effective approach for assessing the need for long-term care assistance. As articulated by Drs. Robert and Rosalie Kane, two renowned researchers in long-term care:
Functional ability is the key to defining the need for long-term care. An emphasis on functioning taps into the behavioral consequences of chronic disease or ill health, rather than focusing on the disease itself.1
Based upon research which has linked measures of functional performance and the need for assistance, Federal policy initiatives in long-term care generally focus on functional criteria as the basis of allocating long-term care benefits. The eligibility of elderly persons to receive publicly-supported long-term care benefits will be determined on the basis of functional measures of performance, on the assumption that such measures are the best indicator of "need" available. Accordingly, most of the legislation which has been developed to expand the Federal government's role in financing long-term care services for the elderly propose to use functional criteria as the eligibility triggers for allocating benefits.
This paper discusses the policy implications of such an approach. Although research has documented the link between measures of ADL performance and the need for services, developing a long-term care program which allocates resources on the same basis must address an entirely new set of issues. Our paper does not say it can't be done; it's purpose, however, is to move the discussion from a research context into a policy context. If we have such a system for allocating benefits, what will it look like? What problems are likely to arise? How can we estimate the costs of such a program? How might functional criteria interact with other eligibility criteria in the allocation of benefits? As someone once said: "The unfortunate thing about policy is that it has to be implemented. Implementation is like original sin, it can't be avoided." And as program managers are well aware, there is a considerable amount of implementation which takes place between the enactment of a piece of long-term care legislation and the point at which an elderly beneficiary walks in the door and says: "Am I eligible for these services or not?"
In Chapter 2, we begin with a discussion of how existing long-term care programs employ functional criteria in determining eligibility for benefits on the premise that future policy initiatives should naturally draw upon what has been learned in the real world. In Chapter 3, we summarize various strategies for triggering benefits based on functional measures of performance, as proposed by the most current Congressional legislation for financing long-term care services. Finally, Chapter 4 provides a discussion of some of the policy implications of rationing long-term care services based on functional eligibility criteria.
In examining the policy implications of allocating long-term care benefits based on functional criteria, it is obviously useful to review how existing third-party payers of long-term care services determine eligibility for benefits. Long-term care services are currently provided to functionally impaired elderly persons under a wide variety of financing mechanisms, including Medicare, Medicaid, State-funded home care programs, and private long-term care insurance. Each of these financing mechanisms has had to develop eligibility criteria for determining who receives benefits and who does not. All of these financing mechanisms include functional criteria in the allocation of long-term care benefits, although it is extremely important to note that functional criteria are not the sole determinants of benefit eligibility in these programs--other eligibility criteria are used as well.
This chapter presents an overview of the functional criteria presently employed in the allocation of long-term care benefits, and how other eligibility criteria used in the allocation of benefits interact with these functional criteria. Lessons which might be drawn from the "real world" of long-term care benefit management in developing future long-term care eligibility criteria are discussed in Chapter 4.
2.1 Medicare Skilled Nursing Facility (SNF) Benefits
Of all the long-term care benefit packages presently in existence, the Medicare Skilled Nursing Facility (SNF) benefit is probably the most restrictive. In fact, the criteria used in determining eligibility for SNF benefits make it clear that the Medicare SNF benefit is strictly for "subacute or rehabilitative" care, and not for custodial care. The Medicare SNF benefit was originally enacted and is presently administered as "extended hospital care" for patients who are still recovering from an acute illness episode, but who can be more economically served in a skilled nursing facility, rather than an inpatient hospital setting.
To be eligible for Medicare SNF coverage, a Medicare beneficiary must require "on a daily basis, skilled nursing care (provided directly by, or requiring the supervision of, skilled nursing personnel) or other skilled rehabilitation services which, as a practical matter, can only be provided in a skilled nursing facility on an in-patient basis, for any of the conditions with respect to which he was receiving inpatient hospital services."2 The regulations governing Medicare SNF coverage and the Medicare Intermediary Manual are quite specific in defining what type of patients and conditions are considered to require skilled care, and eligible for coverage, as opposed to requiring custodial care only, and not eligible. Custodial care is defined as care "which serves to assist the individual in the activities of daily living--such as assistance in walking, getting in and out of bed, bathing, dressing, feeding, and using the toilet, preparation of special diets, and supervision of medication that usually can be self-administered. Custodial care essentially is personal care that does not require the continuing attention of trained medical or paramedical personnel." Thus, the same functional criteria which are generally used to determine eligibility for other long-term care benefits (deficits in activities of daily living) are specifically cited as not eligible for coverage under the Medicare SNF benefit, if skilled nursing care is not also required.
Other eligibility criteria for Medicare SNF coverage serve to reinforce the distinction between subacute and long-term care. These criteria include:
SNF services are only covered after an individual has been transferred from a hospital in which he was a patient for not less than three consecutive days before discharge; and
SNF services must have been necessitated by the same condition which occasioned the patient's qualifying hospital stay.
The subacute character of the Medicare SNF benefit is also reflected in its benefit provisions, which limit coverage to 100 days per spell of illness, with, a significant coinsurance requirement for days 21 to 100. In 1987, less than one percent of Medicare enrollees received Medicare- covered SNF benefits, and the average number of covered days per admission was 21.5.3
Effective April 1, 1988, HFCA issued new implementing instructions for reviewing SNF claims "for greater clarity and in order to help ensure that the guidelines are implemented in a uniform and consistent manner."4 Although HCFA claims that the new guidelines represent only clarifications of current coverage policies, and not a liberalization of functional eligibility criteria, there was a significant increase in the number of SNF claims submitted by providers after implementation of the new guidelines. One possible "clarification" which may have contributed to the increase in submitted claims was that the new guidelines made clear that "a service that is ordinarily considered nonskilled could be considered a skilled service in cases in which, because of special medical complications, skilled nursing or skilled rehabilitation personnel are required to perform or supervise it or to observe the patient." Thus, the new guidelines made it clear that if the overall management and supervision of a patient require skilled nursing personnel, although the direct services provided to the patient may not require skilled care, that the care nonetheless qualifies as skilled care for the purposes of Medicare SNF coverage.
Partly as a result of the clarifications of Medicare SNF coverage, Medicare SNF expenditures rose dramatically from 1988 to 1989. However, it has been difficult to disaggregate the independent effects of the new coverage guidelines and the expanded benefits enacted under the Medicare Catastrophic Coverage Act of 1988, which went into effect on January 1, 1989, and which included the elimination of the three-day prior hospitalization requirement. In FY 1988, Medicare SNF reimbursements totalled $883 million. Total FY 1989 reimbursements are expected to total about $2.3 billion, an increase of 260 percent in a single year.5 Of course, should the expanded Medicare SNF benefits enacted under the, Medicare Catastrophic Coverage Act (MCCA) be repealed, Medicare SNF reimbursements will decline by some amount, and it may be possible to better ascertain the independent impact of the new coverage guidelines on utilization and costs.
2.2 Medicare Home Health Benefit
Medicare will pay for home health visits, including nursing care; physical, occupational or speech therapy; medical social services; home health aide services; and durable medical equipment and supplies, for eligible Medicare beneficiaries. To be eligible for Medicare home health services, an individual must be certified by a physician as in need of skilled nursing services on an "intermittent" (as opposed to continuous) basis. To meet the requirement for "intermittent" skilled nursing care, an individual must have a medically predictable recurring need for skilled nursing services, which in most instances will be met if a patient requires a skilled nursing service at least once every 90 days.6 On the other hand, a patient who is expected to need more or less full-time skilled nursing care over an extended period of time would not qualify for home health benefits.
A final functional requirement that is unique to this benefit is that an eligible beneficiary must be "confined to his home." An individual is considered confined to his home if he has a condition which restricts his ability to leave his home, except with the assistance of another person or the aid of a supportive device (such as crutches, a cane, wheelchair, or a walker), or if he has a condition such that leaving the home is medically contraindicated. This "homebound" requirement has been one of the more controversial eligibility criteria for Medicare home health benefits, and it has been necessary for HCFA to clarify that the requirement doesn't necessarily mean that the individual can never leave his home in order to maintain eligibility for benefits.
It is interesting to note that while the Medicare home health benefit can include assistance with Activities of Daily Living (ADLs), there are no specific eligibility requirements concerning the level of ADL impairment required to receive home health aide services, only that the need for home health aide services be certified by a physician, and that the specific services to be provided be determined by a registered professional nurse.
In 1987, approximately 1.6 million Medicare beneficiaries received home health care benefits, or about 5 percent of the total enrolled population. Medicare reimbursement for home health visits in 1987 totalled just under $1.8 billion.
2.3 Proposed Medicare Respite Care Benefit
The Medicare Catastrophic Coverage Act included coverage of in-home respite care for certain chronically dependent individuals. This benefit represented the first time in which impairments in ADLs were specifically used as eligibility criteria for Medicare-covered services. The Act defined a chronically dependent individual as someone who was certified by a physician as being dependent on a voluntary caregiver for assistance with at least two of the five following activities of daily living: (1) eating; (2) bathing; (3) dressing; (4) toileting; and (5) transferring in and out of a bed or in and out of a chair.
It is important to note that in enacting the respite care benefit, however, Congress attached additional restrictions to the benefit in order to reduce its estimated cost. The Medicare respite benefit was restricted to persons who had met either the Part B out-of-pocket limit ($1,370 in 1990) or the prescription drug deductible ($600 in 1991). Once the deductible was met in any given year, a beneficiary would remain eligible for respite care services for 12 months thereafter.
2.4 Medicaid Nursing Home Benefits
Skilled Nursing Facility (SNF) services are a mandatory Medicaid benefit for eligible persons in all States. States may opt to cover services received in Intermediate Care Facilities (ICF), and virtually all States offer this benefit.
Individual States are given considerable leeway in determining functional eligibility criteria for Medicaid coverage of nursing home care. Federal Medicaid regulations only stipulate that a physician must certify an individual's need for care in a SNF or ICF, and additionally, that those receiving care in an ICF must be assessed by an interdisciplinary team of health professionals to determine need for care. These criteria are referred to as "Level of Care" criteria. Initial certification of the need for SNF or ICF care must be made upon admission, and periodic recertifications must be made thereafter. Recertification of SNF services is required every 30 days in the first three months of a stay, and every 60 days thereafter. Recertification of the need for ICF services is required at 60 days, 180 days, 12 months, 18 months, 24 months, and every 12 months thereafter.
Some States have developed more formal Level of Care procedures by requiring Preadmission Screening (PAS) of persons seeking nursing home admission. State approaches vary considerably. Some require only Medicaid-eligible applicants to be screened; others include persons who, once in the facility, are likely to "spend-down" and become Medicaid eligible after their private resources are depleted. A few States even require persons who are likely to remain private-pay residents to be screened upon admission, although the legality of such a requirement has been the subject of litigation.
The most recent national survey of State PAS programs was conducted by Interstudy in 1985.7,8 This survey found that 30 States had implemented some form of preadmission screening for nursing home applicants. Nearly all of these States screen those who are eligible for Medicaid upon nursing home admission. Approximately one-third of these States also require private-pay residents converting to Medicaid be screened at the time of conversion. Of those States requiring screening of the potential "spend-down" applicants, one State stipulates that persons who would spend-down to Medicaid levels in 60 days must be screened; four States require that a PAS be administered to applicants who would spend-down within 180 days; two States require applicants whose income is less than 300% of the SSI level be screened; and one State stipulates that those who would be "eligible soon" must be screened.
Nearly all State preadmission screening instruments collect information on an applicant/resident's physical health, mental health, informal social supports and functional status. Where the States differ, however, is the manner in which they use the information for deciding whether an individual is eligible for nursing home level of care. Some States rely upon the clinical acumen of the screener/assessor or upon the consensus of a review panel to determine eligibility. In these cases, States may also have written guidelines or regulations to guide the clinicians) in making a judgement. The format of these PAS's may consist of standard prompts/open-ended questions to which the interviewer responds through textual descriptions. The other type of instrument is more structured and provides the screener with standard prompts as well as categorical responses.
Other States, fewer in number, use objective decision rules, or algorithms, for deciding the appropriateness of institutionalization. These algorithms are applied to findings from the PAS, and an individual is classified either as eligible or ineligible.9 Examples of the decision rules used in four States (CT, NY, VA, OR) is presented in Exhibit 2-1.10 As seen in the Exhibit, the decision rules in each of these States vary, but all rely upon some combination of ADLs and behavior problems/cognitive impairment to determine eligibility.
|Connecticut Nursing Home Preadmission Screen Decision Rules|
|New York Nursing Home Preadmission Screen Decision Rules|
|Virginia Nursing Home Preadmission Screen Decision Rules|
|Oregon Nursing Home Preadmission Screen Decision Rules|
The differences in these screens are instructive to the extent that they vary in their restrictiveness. By way of relative comparison, the Connecticut and Oregon screens are the least restrictive of the four screens, and the Virginia and New York screens the most restrictive. For example, one would be considered nursing home eligible in Connecticut if one did not have adequate informal supports and had dependencies in bathing and dressing (2 ADLs) [or dependent in shopping, transportation, housework and laundry (4 IADLs)]. On the other hand, in Virginia, one would have to be dependent in the four ADLs of mobility, transfer, toileting and eating as well as have inadequate informal supports in order to be nursing home eligible. In comparing the two approaches, Virginia seems to exclude from eligibility persons with impairments in only the ADLs of bathing and/or dressing. Since it has been demonstrated that dependency in these two ADL items is likely to appear before dependencies in other ADL (i.e. those used in the Virginia screen), the Virginia screen is likely to exclude from eligibility persons who would be determined nursing home eligible on the basis of the Connecticut screen.
Some recent research corroborates empirically this contextual analysis.11 Decision rules derived from each of the four State screens were applied to a 50% random sample of control group participants in the Channeling demonstration study (baseline). The percent of subjects classified as nursing home eligible under each State PAS was then calculated. Table 2-1 shows that the Oregon screen would determine 95.5 percent of the sample population eligible for nursing home placement, while the Connecticut screen would find the vast majority of the sample eligible as well (81.1%). The Virginia and New York screens, on the other hand, would consider only 39% and 37% of this same group eligible for nursing home placement, respectively.12,13
|TABLE 2-1. Eligibility Rates Based on the Connecticut, Oregon, Virginia and New York Nursing Home Preadmission Screens*|
|* Data Source: 50% random sample of Channeling control group subjects at baseline, age 65 and over.|
While the restrictiveness of a screen is clearly one design consideration, especially in containing program costs, another factor to be taken into account is the relative direction of error built into a given screen. By being overly restrictive, a screen is likely to exclude from program eligibility persons who may really need nursing home services. The predominant errors in these screens; occur in not correctly identifying eligible persons. On the other hand, less restrictive screens are likely to err in the direction of deeming individuals eligible for nursing home services who do not require the level of care provided in a nursing home. A good screen tries to minimize both kinds of errors, but one of the other type of error is likely to predominate in any screen. These data underscore the observation that while research may inform about the types of individuals who need services--and thus who should be considered eligible for services, the relative restrictiveness of eligibility criteria also inevitably reflect public policy and the willingness to fund services.
2.5 Medicaid Home and Community-Based Waiver Services
Under Section 2176 of the Omnibus Budget Reconciliation Act (OBRA) of 1981, States may provide, under Secretarial waiver authority, certain home and community-based services to persons who without such services are at risk of placement in an institutional setting. Such services are not otherwise eligible for Federal Financial Participation (FFP) under the regular Medicaid program. In FY 1988, States spent approximately $712 million for Medicaid waiver services for the aged and disabled.14
To be eligible for waiver services, an applicant must meet the same Level of Care criteria which are used by States to certify eligibility for Medicaid payment in a nursing home. At their option, States may use additional criteria in targeting waiver services to functionally impaired Medicaid enrollees in order to increase the likelihood that persons receiving waiver services would otherwise be forced to enter a nursing home or other institutional setting. . A survey of 31 out of 36 States with Medicaid waivers for the aged and/or physically disabled conducted in 1986 showed that States used a variety of targeting criteria in determining eligibility for waiver services.15 For example, a few States limited eligibility for home and community-based waiver services to persons who were already placed in an institution. Some States incorporate measures of informal support availability in targeting waiver services, while other States prioritize waiver allocation decisions based on patient safety issues. Further, of the 21 States which had Preadmission Screening (PAS) programs in place for nursing home applicants, 14 (67%) had also incorporated their PAS programs into their assessment process for waiver services. However, more specific data on how States use measures of ADL and IADL functioning in their waiver eligibility assessment procedures are not generally available.
2.6 Medicaid State Plan Home Care Services
In addition to Medicaid waiver programs, States can provide home care services to eligible Medicaid enrollees under their regular State Medicaid plans. In general, the primary distinctions between regular State plan services and waiver services are that: (1) regular State plan services must always be certified by a physician and provided under the supervision of licensed health care professionals, whereas waiver services can be authorized and provided outside the mainstream health care system; and (2) States are provided increased flexibility in allocating waiver services (i.e. rationing) than is allowed for regular State plan services. Nonetheless, States provide a broad array of home care services under the personal care, adult day health, home health, rehabilitation and other service options under their regular State plans. For example, in 1987 the State of New York spent $835 million in Medicaid funds for the provision of personal care services to aged and disabled Medicaid enrollees. Again, specific data on the assessment methods used by States in allocating home care benefits under regular State plan services (beyond the Federal requirement that the services be authorized by a physician) are not generally available.
2.7 Private Long-Term Care Insurance
According to the Health Insurance Association of America there were 1.3 million long-term care insurance policy holders as of mid-1989. Based upon their semiannual marketplace survey they found that 109 companies were offering long-term care policies, up from 103 companies as of December 1988. While the greatest expansion in the market occurred in employer/group coverage, the vast majority of policies (88 percent) were hold by individuals.16
Accompanying the growth in the private long-term care insurance market is a trend toward changing the criteria used to define benefit eligibility. Although many policies still require a prior hospital stay in order to be eligible for nursing home coverage (and a prior nursing home stay for home health benefits), an increasing number of policies now rely upon functional criteria, most notably ADLs, as a means for defining the insurable event. This trend is documented in the most recent survey conducted by Consumer Reports,17 and corroborated by a recent review of eligibility criteria conducted by the HIAA. HIAA reports that 14 out of 43 plans reported by member companies offering long-term care insurance used limitations in ADLs as a trigger for services (both institutional and community-based).18 The number and configuration of ADL items used in plans varies. Thirteen policies, however, include bathing, dressing, and eating. Other ADL items included are: toileting (12 plans); transferring (11), mobility or walking (10), and continence (2). One plan also includes cognitive impairment, defined as "the ability to be properly oriented as to time, surrounding other people and the recognition of basic human needs", as a criterion (p.3). And one policy includes the IADL (Instrumental Activities of Daily Living) items of meal preparation and medication management. Criteria for eligibility ranges from 1 out of 7 impairments to 3 out of 7.
Another survey by HIAA, focusing on home health benefits of long-term care insurance policies sold by HIAA member companies in 1988 and 1989, also indicates a move toward functional criteria for benefit determination.19 Only eight of the 28 plans reviewed rely upon functional criteria for determining eligibility. The others require either a prior nursing home stay or a physician's certification of need. When functional items are used for defining home-health eligibility they are similar to those cited above, with bathing, dressing, and eating used by all plans relying upon functional criteria.
In sum, while the industry seems to be moving toward using functional criteria for benefit determination, there seems to be only minimal agreement on which functional items should be used. Policies vary on the number of impairments used to trigger benefits, and presumably they will differ on how impairments are defined. The long-term care insurance market is still developing; one would expect that as claims begin coming in there will be more information forthcoming on how impairment criteria are actually operationalized in the benefit determination process.
The Robert Wood Johnson Foundation has funded eight States to develop partnerships with insurance carriers to offer long-term care insurance. A task of all States involved in these initiatives is to define the criteria which will be used by insurers to determine eligibility for benefits. For some States one of the challenges of defining the insurable event is to devise criteria which are consistent with that used by their Medicaid programs in determining (non-financial) eligibility for nursing home admission.
Consistency between the two programs is viewed as important since the State will guarantee Medicaid long-term care coverage for persons who have purchased insurance and whose private benefits have expired (or once they have spent-down to the level of assets which they have protected by their policy). That is, these States feel that they cannot operate under a double standard whereby they allow less restrictive criteria for those who are promised Medicaid funding after their private insurance benefits expire than for those who are not insured privately (and must rely on Medicaid for financing their long-term care).
Connecticut is one of the partnership States which has chosen to tie the definition of the "insured event" to its Medicaid long-term care benefit criteria. Connecticut's Medicaid nursing home preadmission screen requires that the individual actively seek nursing home placement and have dependencies in 5-6 ADL's; or actively seek nursing home placement and have at least two out of six ADL dependencies (or dependencies in at least 4 out of 8 IADL or at least 4 out of 10 errors on a mental status questionnaire or evidence of wandering or abusive assaultive behavior) and the absence of a totally intact informal social support system to provide for the applicant's needs (See Exhibit 2-1).20 The major difficulty in setting comparable criteria for the private program is that insurers cannot reasonably discriminate on the basis of informal support in paying for benefits (which Medicaid does do).
The other problem facing the CT partnership is that Medicaid requires persons to be seeking nursing home placement in order to access community-based (2176 waiver) services, whereas there would be no such constraint placed upon beneficiaries of private insurance seeking home health services. Given thedifferences in accessibility to community-based services, Connecticut felt that they needed to be somewhat more stringent in defining the insured event for insurance beneficiaries -- if the State was to assure the policy holder Medicaid coverage once s/he had spent down to the asset protection level defined in his/her policy. Thus, in order to compensate for the informal support and seeking-nursing home criteria in the Medicaid PAS, the Connecticut partnership program and insurers have agreed to the following criteria:
- Dependencies in at least 2 out of 6 ADL;
- Evidence of cognitive impairment indicated by any of the following behavioral problems:
- Unacceptable Hygiene or Habits
- Threats to Health/Safety.
If evidence of any behavior problems exists for which daily supervision is required, a mental status exam (Folstain's Mini-Mental State) would then be administered to validate that the behaviors are due to cognitive impairment, and not to psychiatric impairment.
Some insurers in the Connecticut partnership preferred to rely upon the mental status questionnaire items currently being used by in the PAS rather than using behavior problems and the Folstein to identify cognitive impairment. The definition of the insurable event used by these insurers will be:
- Dependencies in at least 2 out of 6 ADLs;
At least 7 (out of 10) errors on the MSQ.
Planners in Connecticut believe that excluding the IADL criteria and making the cognitive criteria somewhat more stringent will result in the identification of comparable populations eligible for Medicaid long-term care benefits and for private long-term care insurance benefits.
The Massachusetts partnership program, on the other hand, represents a slightly different approach to defining benefit eligibility criteria. Massachusetts has chosen to use Medicaid eligibility criteria as a guide, rather than an as a standard, in developing benefit criteria. Thus, "in keeping with the Commonwealth's Medicaid criteria for benefit eligibility," persons would qualify for benefits based on the following criteria:
- Need for human assistance in two or more ADLs (bathing, dressing, toileting, transferring, continence, eating);
Have a medical need or demonstrate evidence of mental dysfunction.21
Medical need is established by determining need for services which are: (1) "in accord with accepted standards of medical practice for the diagnosis and treatment for the insured's condition"; (2) are "delivered in the least intensive health care setting required by the insured's condition when possible" and (3) are "not solely for the convenience of the insured or the insured's family or health care provider, except for respite care..."22 Massachusetts proposes using the Connecticut partnership's approach to defining "mental dysfunction", as defined above. In contrast to Connecticut, the Massachusetts criteria are more stringent. Whereas Connecticut will provide benefits to persons with documented cognitive deficits, regardless of ADL status, Massachusetts will authorize benefits for a cognitively impaired beneficiary only if s/he also demonstrates impairments in two or more ADLs.
As the other States involved in this project come to define benefit criteria, we are likely to see similar functional criteria emerge. We can also expect differences among the States' criteria to parallel, at least loosely, the restrictiveness of their respective Medicaid nursing home eligibility criteria since Medicaid will assume some amount of risk for those enrolled in the demonstration programs.
The purpose of this chapter is to describe legislative bills filed during the 101st Congress (1st Session) which require the use of functional criteria for determining eligibility for the long-term care services that are proposed by the legislation. Long-term care bills that include functional criteria which were filed during this session fell into four major categories:
- Amendments to the Internal Revenue Code of 1986
- Amendments to Title XVI (SSI) of the Social Security Act
- Amendments to Title XIX (Medicaid) of the Social Security Act
- Amendments to Title XVIII (Medicare) of the Social Security Act.
An overview of the content of the legislation within each of these four categories is presented below.23 In instances where there is detailed function-specific eligibility criteria, these criteria are presented. To the extent that it is outlined in the particular legislation, the proposed process of functional eligibility determination is also reviewed.
3.1 Amendments to the Internal Revenue Code of 1986
All of the bills proposing changes in the tax code do so in order to give tax breaks to those who pay out-of-pocket for long-term care services; or to those who purchase long-term care insurance; or to insurance companies offering long-term care insurance; or for employer tax incentives for long-term care coverage. Nearly all of the bills mention functional impairment as the criteria for establishing the need for long-term care services. None of these bills, however, deals with the provision of actual services, and thus do not specify eligibility criteria per se. For this reason, the functional criteria mentioned in these bills will not be described in detail in this report. A summary of these bills appears in Exhibit 3-1.
3.2 Amendments to Title XVI (SSI) of the Social Security Act
During this session three bills were proposed that would amend Title XVI.. One, the SSI Disabled and Blind Children Act of 1989 (H.R.868: Matsui), proposes to take functional limitations into account when determining the eligibility of a child for the entitlement; no specific functional criteria are suggested by the bill, but the Secretary of DHHS must publish guidelines within 18 months of the bill's enactment.
Another proposal, the Supplemental Security Income Reform Act of 1989 (S.665: Heinz), would expand SSI financial eligibility criteria as well as require the Institute of Medicine, or some other entity, to develop clear criteria for determining who is a disabled child.
A third bill amending Title XVI, the National Board and Care Reform Act of 1989 (H.R.2219), was proposed by the late Representative Pepper to establish national minimum standards for board and care facilities. This legislation includes no mention of functional criteria.
Since none of these three bills developed any detailed functional criteria for program participation, they will not be analyzed further in this report.
|EXHIBIT 3-1. Summary of Long-Term Care-Related Amendments to the Internal Revenue Code of 1986|
|Tax Credits for Home Care and Nursing Home Care of a Family Member|
|Tax Credits for the Purchase of Long-Term Care Insurance|
|Employer Tax Incentives for Expanding Health and Long-Term Care Coverage of Retirees|
|Tax Incentives to Facilitate Use of Individual Retirement Plan Funds To Pay for Long-Term Care Insurance Premiums|
|Treatment of Long-Term Care Insurance Contracts the Same as Noncancelable Accident and Health Insurance Contracts for the Purpose of Determining Insurance Company Taxation|
|NOTE: "*" designates legislation which included mention of functional criteria for defining appropriate service recipients.|
3.3 Amendments to Title XIX (Medicaid) of the Social Security Act
Two of the bills proposing changes in Medicaid long-term care services suggest expansion of benefits, but do not include any specifics regarding functional eligibility criteria. The Medicaid Home and Community Quality Services Act of 1989 (S.384: Chaffee et al.) would expand services (social and other services) to the severely disabled to help them attain/maintain maximal potential for independence. H.R.1259 (Panetta et al.) would provide hospice care as Medicaid benefit.
Two other pieces of legislation with nearly identical House and Senate versions, the Medicaid Home and Community Care Options Act of 1989 (S.785: Rockefeller) and the Medicaid Frail Elderly Community Care Amendments of 1989 (H.R.1453: Wyden), would allow States the option of providing community-based long-term care services to the Medicaid-eligible population. They would provide States with an option to the 2176 waiver programs. The bills contain somewhat specific functional criteria for determining eligibility. To be eligible an individual must be at least 65 years of age and unable to perform, without substantial assistance of another person, at least two (out of five) Activities of Daily Living (ADLs). The five ADLs considered are: bathing, dressing, toileting, transferring and eating. ADL disability must be due to physical or cognitive impairment; if a person's ADL disability is due solely to mental illness, then s/he is not program eligible. An individual may also be program eligible if s/he meets the age criteria and has a primary or secondary diagnosis of Alzheimer's disease.
Both H.R.1453 and S.785 require that assessments for determining eligibility be conducted by an interdisciplinary team designated by the State. The legislation further stipulates that the eligibility team cannot be the same entity as the Case Management Agency which is to conduct another assessment to serve as the basis for developing an Individual Community Care Plan (ICCP). Payments for community-based care may not exceed 50% of the average number of persons receiving care in a given quarter, multiplied by the average per them rate for extended care services (institutional), multiplied by the number of days in the quarter. The 50% cost cap refers to an average across all program recipients, and not to individual recipients.
3.4 Amendments to Title XVIII (Medicare) of the Social Security
Two of the proposed amendments to Title XVIII involve regulation of long-term care insurance policies. One, the Long-Term Care Insurance Consumer Protection Act of 1989 (S.142: Durenberger et al.), proposes voluntary certification of policies; and the other, the Consumer Protection for Long-Term Care Insurance Act of 1989 (H.R.1325: Stark), would require mandatory standards for all long-term care insurance policies. S.142 mentions functional impairment as criteria for defining who is chronically ill, and thus eligible for benefits; no detail on the nature of the impairments is given in the bill. There is no mention of functional criteria in H.R.1325.
The remaining four bills seeking to amend Title XVIII all propose expanding the coverage of long-term care services. Only one, the Medicare Adult Day Care Amendments of 1989 (H.R.990: Panetta), does not include any detailed functional criteria. It merely stipulates that to be eligible a person must be 18 years of age or older and have a physical, emotional, mental or neurological impairment that, without the provision of adult day care, would require the level of care provided in a hospital, skilled nursing facility (SNF), or intermediate care facility (ICF). The other three bills which would expand Medicare covered long-term care services are: the Medicare Adult Day Health Care Amendments of 1989 (S.524: Bradley); the Long-Term Care Act of 1989 (H.R.2263: Pepper); and Elder-Care Long-Term Care Assistance Act of 1989 (H.R.3140: Waxman). Each of these is discussed in detail below.
S.524 (Bradley), the Medicare Adult Day Health Care Amendments of 1989, is similar in intent to Panetta's H.R.990; it differs from H.R.990 in that it does outline some function-related eligibility criteria. S.524 stipulates that to be eligible for the adult day health care benefit an individual must be age 18 or older and not able to perform (without assistance), on a daily basis, at least two ADLs out of five. The five ADLs include bathing, dressing, toileting, transferring in and out of bed or in and out of a chair, and eating. A nonfunctional alternative eligibility criteria is also outlined in the bill; to fulfill this alternative criteria an individual must also be age 18 or older and require care in a hospital, SNF or ICF (i.e., needs health maintenance or restorative care) due to a medical or mental impairment.
The eligibility determination process specified in S.524 calls for an assessment conducted by a multidisciplinary team located at the adult day care center; the team is to consist of a physician, registered nurse, social worker, and other consultants as needed. Those receiving benefits under this provision would be required to pay a $5 per day coinsurance fee for services, but the Medicare deductible would not apply to these services.
H.R.2263 (Pepper), the Long-Term Care Act of 1989, seeks to amend Title XVIII to provide protection for long-term home care for certain chronically ill/disabled children and elderly individuals. The eligibility criteria for children stipulates that the child must be under age 19 and meet one of the following three conditions:
Be chronically ill or disabled and unable to perform (without human assistance or supervision) due to chronic illness/injury at least two age-appropriate ADLs (out of five); or
Have a similar level of disability as defined above due to cognitive impairment; or
Require a medical device to compensate for the loss of a vital body function necessary to avert death or loss of body functional capacity and require substantial and ongoing nursing care to avert death or further disability.
Criteria for determining eligibility for the elderly are that the person must be at least 65 years of age and meet one of the following:
Be chronically ill or disabled and unable to perform (without human assistance or supervision) at least two ADLs (out of five); or
Have a similar level of disability (as defined immediately above) due to cognitive impairment.
An elderly person who is also terminally ill must first exhaust benefits for hospice care under Medicare before becoming eligible for home care under this provision. Determination of eligibility is the responsibility of a case management agency.
H.R.2263 specifies that the five ADLs to be considered in determining eligibility are: bathing, dressing, toileting, transferring and eating. This legislation also provides definitions for each of these functional tasks. For example, bathing includes "the overall complex behavior of getting water and cleansing the whole body, including turning on the water for a bath, shower, or sponge bath, getting to, in and out of a tub or shower, and washing and drying oneself." Such specific definitions will prove helpful when it comes time to apply the functional criteria of the bill to national data sets for estimating the numbers of persons eligible for benefits. Legislation which is less specific will require assumptions about the actual definition of ADLs; estimates will be accurate to the extent that assumptions match emerging definitional guidelines of bills which become law. An example of lack of definitional clarity is the absence of guidelines for defining "cognitive impairment" in H.R.2263; the bill merely stipulates that upon passage the Secretary of DHHS must develop regulations for defining "cognitive impairment".
The amount of coverage under H.R.2263 varies as a function of the beneficiary's age and type/degree of disability. Full coverage for home care services would be provided up to 100% of the per diem nursing facility rate in a State for children (under age 19) who require a medical device and ongoing nursing care. For all other beneficiaries, home care services are covered up to 50% of the cost of the average per diem State nursing facility rate for persons with moderate impairment, and up to 65% for those with severe impairment. (The legislation does not include definitions of "moderate" and "severe" impairment; as such, cost estimates for such a bill will have to rely on assumptions as to what constitutes impairment severity.) No copayments are required, but if there is a program deficit at any point, the bill contains a provision to establish copayments which are never to exceed 5% of the national average daily payment rate for home care.
A portion of H.R.3140 (Waxman), the Elder-Care Long-Term Care Assistance Act of 1989, seeks to amend Title XVIII to expand coverage for both community-based and institutional care. All Medicare Part A participants would be eligible for expanded benefits provided they meet the functional eligibility criteria. To be eligible for community care a person must have either a severe or moderate impairment. Operational definitions of "severe" and "moderate" are not given in the legislation, but are to be defined by the Secretary of DHHS upon passage of the bill. One may receive up to 20 hours of community-based care per week for moderate impairments, and up to 30 hours per week for severe impairments. A copayment of 20% would be required of those receiving services.
Nursing facility eligibility requirements under Waxman's proposal state that the person must be:
Chronically dependent, defined as being unable to perform, without substantial assistance of another person, because of physical or cognitive impairment, at least two ADLs (bathing, dressing, toileting, transferring, eating); or
Have similar functional disabilities due to cognitive impairment such that they require substantial direction, instruction, or supervision of another person in order to perform two or more ADLs; or
Require substantial direction, instruction or supervision of another person in order to remain in the community without causing harm to self/others because of inappropriate behavioral problems.
Eligibility for coverage under the nursing home provision also requires that an individual must need nursing home level of care for a continuous period of at least 60 days.
Waxman's nursing home benefit covers the "back-end" of institutional stays; the individual is fully responsible for the first 60 days of nursing home care. Beneficiaries are responsible for coinsurance of 33.3% of the national average per them rate beginning on day 61 through the second year of a stay; coinsurance would be reduced to 10% for stays beyond two years. (Coinsurance would not apply to skilled nursing care under Medicare Part A.)
The process of eligibility determination for both community and institutional services is similar. A multidisciplinary team, which must include a registered nurse, from a CARE (community assessment, review and evaluation) agency would assess individuals seeking services/admission. Eligibility determination for those seeking institutional care would occur before admission to a facility.
Finally, Senator Kennedy has drafted, but not yet filed, legislation which would also expand coverage for community-based and institutional long-term care. The tentative title of this legislation is the Security Care Act. While Waxman's proposal would cover the "back-end" of nursing home stays (but covers an increasingly larger proportion of the cost of longer stays), Kennedy's bill would fund the "front-end." The Kennedy bill would fully cover the first six months of a nursing home stay. Coverage for stays beyond six months would be the responsibility of the individual. This legislation, however, would also establish an optional federal long-term care insurance program for persons age 45 and older through which one could insure against nursing home stays longer than six months duration. Insured stays beyond six months would be reimbursed at 65% of the cost of the per them rate for an unlimited time, provided eligibility criteria continue to be met.
Community-based long-term care under Kennedy's plan would be covered in full as long as the cost of care does not exceed, per individual, 65% of the cost of the average amount for care in a nursing facility multiplied by a severity index weight. The only other constraint is that respite care may not exceed 30 days, or 720 hours, during a calendar year.
Like some of the other expanded-care bills, Kennedy's specifies function-based criteria for determining program eligibility. To be eligible for community-based services a person must fall into one of the following three categories:
* Age 65 or older* Medicare Part A eligible* Completely dependent (does not participate) in at leas one ADL (out of five) ORPartially dependent in 2 or more ADLs (requires physical assist of supervision ORSo cognitively impaired as to require substantial supervision from another individual because s/he engages in inappropriate behavior that poses a substantial health/safety hazard to self or other (i.e., cognitively impaired and exhibiting dangerous behaviors)
- * Under age 19* Dependent in one or more age-appropriate ADLs (includes complete and partial dependence -- may include supervision, some human assistance as well as does not participate) ORRequires device necessary to avert death or maintain functional capacity and requires ongoing nursing care to avert death or further disability
- * Would be eligible for benefits under Title XVIII of the Social Security Act, except for required 24 month waiting period* Completely dependent (does not participate) in one or more age-appropriate ADLs ORIs unable to perform two or more age-appropriate ADLs without some physical assistance or supervision OR* Has a medical prognosis of life expectancy of 12 months or less.
Nursing home eligibility criteria are identical to those for community-based care, and criteria for respite care are also the same as those for community-based services, with the additional following stipulation:
* Is dependent on a primary caregiver for performance of at least two age-appropriate ADLs and without such assistance the individual would not perform the ADL ORHas dementia or other cognitive impairments.
Senator Kennedy's legislation contains very specific definitions of what constitutes ADL impairments. The five ADLs which are included are: bathing, dressing, toileting, transferring, and eating. An example of the high degree of specificity in defining ADLs in this legislation is the definition of toileting: going to the bathroom for bladder/bowel function; transferring on/off the toilet; cleansing after elimination; and arranging clothing.
Persons applying for benefits under this proposal would be judged eligible based upon the outcome of a screen administered by a Long-Term Care Screening Agency. The screening process would be a two-pronged activity. First, an initial screen would be administered over the phone or a questionnaire would be completed by the applicant/proxy. On the basis of this initial information, an in-person screening would be conducted for those having passed the first screening component. The second screen would be administered by a team that must include a physician, nurse practitioner, or registered nurse. Once deemed eligible by the Screening Agency, a Long-Term Care Management Agency would conduct a comprehensive needs assessment and develop a care plan. The only exception to this eligibility determination process would be in the case of those applying for respite care services; eligibility for these applicants would be conducted by the Long-Term Care Management Agency.
Long-term care-related bills filed during the first session of the 101st Congress fall into four general categories: amendments to the Internal Revenue Code of 1986 or amendments to either Title XVI (SSI), Title XIX (Medicaid), or Title XVIII (Medicare) of the Social Security Act. The purpose of bills proposing changes in the tax code is to render tax breaks to purchasers and offerors of long-term care insurance (insurance companies, employers). Proposed amendments to the Social Security Act seek, for the most part, to expand long-term care benefits.
Nearly all of these bills include some mention of physical functioning as a criteria for determining eligibility for benefits. Some bills are more specific than others in defining criteria. Most of the bills, however, also specify additional eligibility criteria other than physical functioning (ADL). These include indicators of cognitive impairment (e.g., Alzheimer's diagnosis, behavior problems indicative of cognitive deficits) and the need for medical devices or ongoing nursing care.
Besides criteria relating to the condition of the beneficiary, the expanded care bills also stipulate other criteria that serve to limit the beneficiary pool. These other factors include waiting periods, co-payments for services, and limits on the amount and duration of care. In this regard, the proposed programs are similar to existing federal programs (e.g., the Medicare SNF and respite benefits), which place analogous restrictions on utilization.
|EXHIBIT 3-2. Summary of Eligibility Criteria101st Congress (1st Session) Bills|
|(2/5; unable w/o substantial assistance)|
|(2/5; unable w/o assistance daily)||(requires hospital, SNF or ICF for medical or mental impairment)|
|(2/5; unable w/o assistance or supervision)||(Needs medical device & ongoing nursing care)|
|(2/5; unable w/o assistance or supervision)|
|(Moderate or severe impairment)|
|(2/5)||(requires substantial direction, instruction or supervision in 2+ ADL)||(Needs >60 days NH level of care)|
|S.???||Kennedy||65+Part A Eligible||CBS & NH||X||OR||X|
|(1/5 complete or 2/5 partial dependence)|
|(1/5)||(Needs medical device & ongoing nursing care)|
|Medicare Eligible but in 24 mos. waiting period||CBS & NH||X||AND||X|
|(1/5 complete or 2/5 partial dependence)||(Life expectancy <12 months)|
|All above categories||Respite||X||OR||X||AND||X|
|(2/5)||(Dependent on primary caregiver for ADL assistance)|
There is considerable discussion in the policy arena regarding an expanded Federal role in financing long-term care services for functionally impaired elderly and disabled persons, and it is commonly assumed that under this expanded Federal initiative, long-term care benefits will be allocated on the basis of functional measures of performance in the Activities of Daily Living. This chapter discusses the implications of this assumption, drawing upon what can be learned from the research on ADL measurement and from existing programs which already allocate long-term care benefits using functional eligibility criteria.
4.1 No Existing Long-Term Care Program Allocates Benefits Solely on the Basis of ADL Measures
The review of long-term care eligibility criteria presented in Chapter 3 shows that while measures of ADL performance are often included in the eligibility determination process under existing long-term care programs, ADL impairments are never the sole eligibility criteria. Indeed, one general observation is that the broader the entitlement nature of the program, the less likely that benefit allocation decisions are based on ADL measures.
Medicare, our broadest entitlement program, has strictly shied away from using ADLs as eligibility criteria in the allocation of benefits. Medicare's Skilled Nursing Facility benefit and home health care benefit are still primarily based upon a physician's certification of the need for care, and upon a determination that the individual is need of skilled, rather than custodial, care. In fact, care needs related to deficits in ADL performance are specifically cited as needs which are not covered by the Medicare benefit package. This exclusion of coverage for assistance in ADLs reinforces the position that Medicare is strictly an acute care insurance program, and that chronic care needs are not addressed by the Medicare benefit package. It is interesting to note, however, that the proposed respite care benefit, enacted as part of the Medicare Catastrophic Coverage Act, did propose to use ADLs as eligibility criteria in allocating benefits, although policymakers were sure to include other criteria to limit access to the benefit (i.e. limiting the benefit to persons who had met the out-of-pocket limit for Part B or prescription drug costs) and expected costs.
Medicaid is the primary public financing mechanism for long-term care, and many States include assessments of ADL performance in determining eligibility for nursing home and/or home-based long-term care services. However, our review of State long-term care eligibility criteria shows that service allocation decisions are only loosely tied to standardized measures of functional performance in most States. In most States, a combination of physician authorization, clinical judgment, and functional assessment are used in determining eligibility for Medicaid payment for nursing home care. More importantly, we cited recent research which demonstrates that even within those States which rely more heavily on functional assessment, that objective measures of "need" in ADLs vary significantly from State to State depending upon the overall availability of Medicaid resources to pay for long-term care services. There is little available data on how States use functional criteria in allocating home and community-based care services under Medicaid.
ADLs are used in many private long-term care insurance policies as eligibility triggers for nursing home and home care benefits, although it is interesting to note that only 8 of 28 plans which cover home care services rely on ADLs. Others require either a prior nursing home stay or a physician's certification of need. We believe this reflects the skittishness of insurers to promise benefits strictly based on ADL measures. Since benefit eligibility in the long-term care insurance market is more hypothetical than actual at this stage, we believe the subject of functional eligibility criteria for benefits will become a much larger issue once claims for covered services begin to increase.
4.2 Estimates of the Functionally Impaired Elderly Vary Depending Upon the Data Source
Although there is considerable agreement that functional measures are superior to diagnostic or other approaches for assessing the need for long-term care services, it is difficult to develop accurate estimates of the number of elderly persons who are functionally impaired.24 At the same time, any Federal policy initiative in long-term care will require the development of reasonably accurate estimates of the costs of expanded benefits, so that financing mechanisms can also be developed.
To address this question, the Committee on Definitions of Functional Limitations, under the direction of the Forum on Age-Related Statistics and comprised of representatives from various government agencies, reviewed estimates of the functionally impaired elderly population based upon several national surveys. Initial examination of these estimates revealed considerable discrepancies. One major reason for the differences is that definitions of functional impairment varied somewhat from survey to survey. In an attempt to reconcile the differences among the estimates the Committee applied a uniform definition of impairment to each survey. Prevalence estimates of impairment were generated for the five ADL items common to the surveys (bathing, dressing, toileting, transferring and eating). Receipt of human assistance in the performance of an ADL was taken as evidence of impairment.
Even after the estimates of ADL functioning were standardized by number, type and definition of impairment, discrepancies remained. For example, data from the 1984 Supplement on Aging show that that 5.0% of the non-institutionalized elderly (1.32 million) have at least one ADL impairment, while data from the 1987 National Medical Expenditures Survey show that approximately 8.1% (2.25 million) are impaired, a difference of slightly over 70%.25
Since most of the current legislative proposals discussed in Chapter 3 propose to trigger long-term care benefits when someone is impaired in two or more ADLs, estimates of the elderly population dependent in two or more ADLs are of greater policy interest. While the Committee on Definitions of Functional Limitations did not report these estimates, it is possible to develop estimates based on prior research of the hierarchical nature of ADLs. Research has shown that loss of function occurs in an orderly and hierarchical fashion whereby loss of ability to bathe oneself is very likely to occur before loss of function in other ADLs.26 Loss of function in bathing is followed in sequence by inability to dress, toilet, transfer and feed oneself. For example, a person dependent in dressing is likely to be dependent in bathing as well. Persons with disabilities further on down the hierarchy are very likely to also be dependent in bathing and dressing.27 Using this approach with the reported data from the 1984 Supplement on Aging and the 1987 National Medical Expenditures Survey, estimates of the number of non-institutionalized elderly with two or more impairments range between 2.9% and 4.4% (between 771,000 and 1,228,000 persons).28
Another policy option is to limit benefit eligibility to persons who require ADL assistance more than once daily, i.e. in the ADLs of transferring, toileting, and eating, or using the hierarchical method, persons who have three or more ADL impairments. Persons with three ADL impairments are likely to need assistance in toileting, the ADL limitation which is most strongly associated with increased future needs for assistance.29 A three ADL benefit trigger would ration benefits to persons with more heavy care needs. Somewhere between 2.6% and 4.2% of the community-dwelling elderly would be eligible for benefits under the 3+ ADL criteria.30
4.3 National Surveys May Underestimate the Number of Elderly with ADL Limitations
Even if all discrepancies between data sources could be resolved, there will continue to be uncertainty about estimates that are derived from national survey data. Estimates derived from survey data may underestimate the true prevalence of impairment in the population. There is some evidence which suggests that the elderly may underreport the extent of their impairments, perhaps in order to present themselves in the best possible light, i.e., unimpaired or less impaired.31,32 However, if a national long-term care program was implemented which rationed benefits on the basis of functional impairments, it is reasonable to expect some amount of exaggeration of reported impairment in order to increase the probability of accessing services. In developing estimates of the eligible population under an ADL-triggered benefit program, it is important to account for these factors in projecting program utilization and costs.33
4.4 Estimates of the Cognitively Impaired Elderly Are Even More Difficult to Derive
In the long-term care policy debate, there is an emerging consensus that eligibility criteria for long-term care benefits should include cognitive as well as physical impairments. This trend can be observed in both legislative proposals and in private long-term care insurance policies. While loss of ADL functioning is considered the "final common pathway" of decline for conditions associated with both physical and cognitive etiologies, persons in the early to middle stages of Alzheimer's disease (and related disorders) may not exhibit any dependence in ADLs, but may not safely reside in the community without some level of supervision. Cognitively impaired persons may be able to bathe and dress themselves, but many should not be left unattended for long periods of time because they may wander off, get lost, or put themselves or others at risk, e.g. leaving on the stove, not remembering-to eat, ingesting harmful substances, etc. Eligibility criteria based on physical functioning alone will exclude some individuals who are in need of less intense human assistance but who may require fairly consistent supervision, and who would otherwise be at risk. Recent research estimates this population to be approximately one-half million persons.34
There is no standard measure for identifying cognitive impairments. The measures and approaches which do exist are not universally accepted in the way ADL impairments are accepted as measures of physical functioning. This dilemma is reflected in the Federal legislative proposals, which detail the number and types of ADL eligibility criteria, but which are notably non-specific regarding indicators for cognitive impairment.
One approach proposed in both the Rockefeller (S.785) and Wyden (H.R.1453) bills is to qualify individuals for benefits based on a primary or secondary diagnosis of Alzheimer's Disease. This is not a terribly satisfactory approach since a definitive diagnosis of Alzheimer's Disease can only be made post mortem. But more importantly, most persons in the very early stages of Alzheimer's do not need long-term care. These individuals may have some difficulty in recall, but in the early stages of the disease process are not likely to present a safety hazard to themselves or others. Relying upon a diagnostic approach may result in some misspecification of the target population, i.e., deeming eligible persons not currently needing assistance, unless other criteria are employed as well.
Surveys of the elderly have typically used the Short Portable Mental Status Questionnaire, a ten-item scale tapping orientation and memory, or some variant of the measure, to identify persons with cognitive impairments. The difficulty with this strategy is that such measures provide only gross approximations of impairment level. They are able to identify persons in the later stages of decline with a good deal of certainty, but not necessarily those in the early to middle stages. And it is precisely those persons in the middle stages of decline who are not as likely to also be ADL dependent, but who may need some amount of assistance and supervision. Depending upon the specific cut-off points employed, use of an MSQ-type approach could result in considerable targeting inefficiencies.35
Another method of identifying cognitively impaired persons in need of assistance, which has promise, is the "behavior problems" approach. Many nursing home preadmission screening programs utilize this approach, and a few of the legislative proposals suggest this approach as well, although the content of the bills provide no guidance regarding actual measures.36 Behaviors which are sometimes used as manifesting cognitive impairments which place an individual at risk include:
Wandering: Aimless, potentially dangerous movement within or outside the home;
Abusiveness: Physically causing harm to self or others; verbal assaults such as threatening physical attack or menacing in other ways;
Unacceptable Hygiene or Habits: Gross and unacceptable hygiene or eating habits, such as throwing or smearing food or excrement; disrobing in inappropriate situations; screaming; making dangerous or inappropriate sexual advances;
Threats to Health or Safety: Inability to follow medication or dietary regimens without supervision; creating fire hazards; exhibiting poor judgment which is potentially harmful to self or others.37
In order to determine eligibility based on these criteria, the need for daily supervision due to these behavioral problems must also be considered, since what is really appropriate for the the policy process is the need for oversight, which ultimately translates into need for services.
One problem with the behavioral approach to cognitive impairment assessment is that persons displaying the types of behavioral problems described above may suffer from mental illness rather than cognitive impairment. There are instruments which can be used in differential diagnosis, but since we don't know how frequently potential beneficiaries may exhibit behavior problems due solely to psychiatric problems, it seems premature to suggest further assessments to validate the finding of cognitive impairment based on behavioral indicators.38 Another potential disadvantage of using a behavioral approach over an MSQ approach is that it relies upon reported observations of the applicant which may or may not be accurate. Also, some applicants may not have a caregiver or family member who could provide this type of information. In these instances, the more objective approach exemplified by the MSQ may be a preferred method for ascertaining cognitive impairment.
4.5 What Level of Functional Impairment Should Trigger Eligibility for Services
A major goal in the implementation of a national long-term care benefit program will be to assure uniformity in the eligibility determination process. Eligibility systems will be structured, regulations will be promulgated, assessment tools will be developed, and screeners/assessors will be trained, all with the objective of attaining an acceptable level of uniformity in allocation of benefits.
One major problem in the development of this eligibility system is that within the potential target population (the impaired elderly), functional impairment is always a continuous variable, while the eligibility determination process is always a binary decision. One cannot be more eligible or less eligible--one is either eligible or not eligible. But one can be more or less functional. An individual's functioning level can range from total independence, to slight impairment, to moderate impairment, to severe impairment, to so impaired that someone else must take care of all of one's basic daily needs. In developing a long-term care eligibility system, policymakers will essentially be deciding the "optimal" point on the functional continuum for separating out the eligible from the non-eligible.
The prevailing opinion, as evidenced from the legislative proposals stlymarized in Chapter 3, appears to be that impairment in two or more ADLs is the optimal eligibility trigger. There seems to be little rationale, however, in any of the literature for selecting two or more ADLs as the optimal eligibility cut-off point. Obviously, one reason for selecting two or more ADLs as the benefit trigger, rather than one or more, is to limit the eligible population and thereby program costs. However, this approach is likely to exclude from services persons who require assistance with bathing or dressing.39 For persons with such impairments, lack of assistance on a daily basis may not be life-threatening. This is not to say however, that those with one ADL impairment are not in need, or would not benefit from assistance.
Whatever level of impairment is decided upon as the trigger for benefits, there will inevitably be assessment, measurement, and clinical decision errors around that point. Given the political and regulatory controversies over decision rules regarding Medicare payment for "medically necessary" hospital care, eligibility for SNF services, and other Medicare-covered services, the enactment and implementation of a Federally-financed long-term care benefit based upon functionally-based eligibility criteria will inevitably lead to additional controversies regarding approved and denied applications for services, particularly for that population which falls extremely close to the designated "cut-off point." Indeed, the clear demarcation of eligible and non-eligible applicants for benefits is even more problematic in a chronically ill population than in a population whose needs are more acute in nature. Policymakers need to consider the controversies that will inevitably surround the designation of an eligibility cut-off point in the continuum of functional impairment levels.
4.6 Should Long-Term Care Benefits be Rationed Based on Objective Measures of ADL and Cognitive Impairment or Should Benefits Take Into Account "Unmet Needs"?
Policymakers concerned about the costs of a Federally-financed long-term care benefit package are naturally motivated to target resources to persons who are in the most need of public assistance. Given concerns about the number of elderly persons who may be eligible for benefits based on objective measures of the need for assistance, alternative strategies for restricting the benefit population are being considered. One alternative is to base eligibility on evidence of unmet need, rather than total needs. Unmet need is generally defined as the difference between the amount of care which an individual's support system is currently providing, or can provide, and the individual's total need for care.
Some States incorporate the concept of unmet need in their nursing home screening and waiver programs by including the availability of informal caregivers in their decision rules. That is, to be eligible, persons must not only have documented impairments, but must also be receiving inadequate assistance from family, friends, or other available avenues of assistance.
It is important to note that the concept of unmet needs as a strategy for rationing benefits is inimical to the entitlement concept exemplified by the Medicare program.40 Under Medicare, if one has a demonstrated need, then one qualifies for benefits. Although Medicaid is also an entitlement program, States are more likely to impose criteria beyond need in determining eligibility for benefits. It is interesting to note that although Medicaid regulations do not require that informal supports be taken into account when determining eligibility for nursing home admission (ICF), they do indicate that the assessment preceding admission must include an "evaluation by an agency worker of the resources available in the home, family and community..."41 Thus, at least implicitly, Medicaid regulations suggest that unmet needs, rather than total needs per se, be considered as criteria for long-term care benefits.
Some States clearly limit Medicaid services to persons with unmet needs. Connecticut, for example, invokes the notion of "practical matter" in determining Medicaid eligibility for nursing home care. If an individual is impaired, but all service needs are being supplied by a family member, the State may not provide assistance because, as a "practical matter," the individual does not require additional assistance.42 New York also operationalizes the concept of unmet needs in its PAS decision rules by stipulating that functional impairment must be accompanied by inadequate informal supports.43
Even though States may use the concept of unmet needs in rationing long-term care resources under Medicaid, such a rationing approach may be more problematic within a Federal policy initiative, since Federally-administered programs generally require more formal decision rules regarding the allocation of benefits. Although the concept of unmet needs has intuitive appeal as a strategy for targeting long-term care resources, it is a difficult concept to operationalize in a formal program.44 For example, it is difficult to develop uniform decision rules for evaluating both the capacity and the willingness of the informal support system to provide services, particularly when a support system may be available but unwilling. Similarly, it is difficult to develop decision rules regarding the competing demands on informal caregivers, or regarding their physical and emotional capacity to fulfill caregiving functions. Thus, existing long-term care programs which incorporate assessments of informal caregiving capacity into their allocation decisions tend to rely heavily on clinical judgment, rather than formal decision rules.
The unmet needs concept has also been criticized as being inherently discriminatory towards women. Women's advocates correctly point out that such a policy unfairly reinforces the cultural expectation that wives, daughters, and other female relatives are responsible for the care of elders. In the absence of a national long-term care benefit program, the current caregiving burden falls disproportionately on women, and an unmet needs policy in a Federal long-term care initiative would institutionalize this uneven distribution of responsibility. Thus, a long-term care allocation policy which incorporates the concept of unmet need is likely to encounter significant political resistance, and might also be challenged in the courts.
4.7 What are the Alternatives to Functional Eligibility Criteria?
Although functional criteria are likely to be a central component of any Federally-established eligibility system for long-term care benefits, the inherent difficulties involved in the implementation of a system based solely on functional criteria has led to the consideration of additional criteria or program features which be incorporated into a long-term care eligibility system which could make the allocation of benefits simpler, "cleaner," more efficiently targeted, and less subject to controversy.
One idea is to impose a deductible or waiting period period on a Federally-supported long-term care benefit. One example of this approach, commonly referred to as "back-end" coverage, is Congressman's Waxman's proposal for nursing home coverage. This bill (H.R.3140) would require individual liability for the first 60 days of nursing home coverage. For days 61 through to the end of two years, the beneficiary would be responsible for a 33.3% co-payment on the cost of nursing home care. For nursing home stays beyond two years, this co-insurance requirement would be reduced to 10%.
Under this approach, prior utilization of long-term care services, paid for out-of-pocket, becomes the primary criterion for allocating publicly financed benefits. This approach assumes that a person who pays for long-term care services out of their own resources must be truly in need of assistance (i.e. sufficiently impaired) and therefore, should qualify for public assistance after the deductible period has been met.45
One major criticism of this approach is that benefits which require a substantial out-of-pocket deductible tend to favor persons with higher income and assets. Persons of more moderate means may forego purchasing long-term care benefits, despite high levels of impairment, and therefore never qualify for public benefits.
Another criticism of this approach is that it assumes that once an individual is impaired, he or she will either remain at the same level of impairment or deteriorate further. Empirical evidence shows, however, that there is as considerable improvement, as well as decline, in functional status over time among the elderly population.46,47 The approach also assumes that an individual's prior level of service utilization was not in excess of what was actually needed.
In effect, Medicaid is an example of a "back-end" long-term care coverage program in which the deductible is defined as "everything you have in excess of $2,000." However, even for nursing home patients who enter as private pay residents and impoverish themselves to Medicaid eligibility levels after spending-down, States still employ functional criteria in establishing eligibility for Medicaid payment. A small percentage of spend-down residents do not meet these functional criteria at their point of conversion to Medicaid, and become at risk of being discharged. In other cases, those who have spent down their assets in a Skilled Nursing Facility do not meet Medicaid criteria for SNF coverage, and must be transferred to a lower level facility (i.e. an ICF). Thus, even under a Federal program which rations long-term care benefits by imposing a significant deductible or waiting period, there will still be a need to apply functional criteria at the point of eligibility. However, given the prior utilization of long-term care services, the application of functional criteria at this eligibility point may be significantly less problematic than in a program which attempts to apply criteria for "front-end" coverage.
4.8 Should the Same Functional Criteria be Applied for Institutional Services and Community-Based Services?
A number of Federal long-term care policy initiatives have addressed the issue of whether it is possible to ration home and community-based services to persons who would otherwise enter nursing homes. These initiatives were implemented with the hope that long-term care service benefit packages could be expanded to include home-based services at no increased cost to the taxpayer, if services could be efficiently targeted to persons who would otherwise become utilizers of publicly-financed nursing home care.
The answer to this policy issue has been a fairly definitive "no" due to the fact that there are a large number of persons in the community who are as functionally impaired as persons in nursing homes (some evidence that the proportion may be as high as three community-impaired elders to every nursing home resident), but who will never enter nursing homes. Thus, it has become increasingly acknowledged that an expansion in public coverage of long-term care services to include home and community-based services will require increased investment of public resources.
However, as the policy debate goes beyond the "cost-effectiveness" of home and community-based services, the issue will remain whether the eligibility criteria for home-based services should be the same as those applied for institutional services. On a conceptual level, it is reasonable to expect that a Federally-financed long-term care program would ration benefits according to various levels of need among those deemed eligible for benefits. Thus, persons who receive a higher level of benefits (e.g. nursing home care) would be expected to meet a more restrictive set of eligibility criteria than persons in need of lower level services (e.g. respite care). Nursing home case-mix reimbursements systems are an example of how long term care systems are becoming increasingly more sophisticated in allocating long-term care resources according to levels of need, and similar systems may be developed for home care services. Thui, in enacting a Federal long-term care benefit, policymakers will not only have to develop "threshold" functional criteria to determine eligibility for benefits, but criteria for rationing benefits among the eligible population.
This chapter discusses some of the policy issues related to the enactment of a Federal long-term care benefit package which allocates services on the basis of functional criteria among the elderly population. These issues include:
Existing long-term care programs do not allocate benefits solely on the basis of functional criteria. Clinical judgment and physician authorization are generally used, in addition to functional criteria, in allocating benefits.
It is difficult to estimate the number of elderly persons who will be eligible for benefits on the basis of functional criteria because data sources differ in their estimates of the functionally impaired elderly, depending upon how impairments are defined and measured.
The number of functionally impaired elderly persons who will be eligible for benefits may be underestimated from existing data sources since survey respondents may tend to underreport their impairments while program applicants tend may tend to exaggerate their disabilities in order to receive services.
Most long-term care initiatives recognize the need to also extend eligibility for services to persons who may be physically capable of functioning independently, but whose cognitive impairments result in a need for monitoring supervision. However, objective measures of cognitive performance, except at very high levels of impairment, are not well refined.
The considerations discussed above make the estimation of the costs of a long-term care benefit package based on functional criteria extremely problematic.
In developing eligibility criteria for long-term care benefits, policymakers are essentially designating a level of "need" that determines eligibility for benefits. For some reason, policymakers seem to have focused on "impairment in two or more ADLs" as the level of need warranting publicly supported services, without strong justification. Further, any eligibility cut-off point is likely to encounter controversy as persons of relatively equal impairment near the cut-off point are determined eligible or not eligible for services.
Although the concept of unmet needs has proved a useful mechanism for allocating services in smaller programs, it will be a difficult concept to operationalize in a more formal and uniform Federal program, because it is difficult to develop formal decision rules about the availability and willingness of informal care providers to supplement publicly-financed services. Also, the concept of unmet needs may encounter resistance as being inherently discriminatory towards women, who make up the vast majority of informal caregivers.
Alternative functional criteria for allocating benefits, other than objective measures of impairment at the point of application, are being considered. For example, a "back-end" long-term care benefit package, with a significant "front-end" deductible, has the effect of using "prior utilization" of long-term care services as a functional eligibility criterion. This approach is not without its own disadvantages, however, such as its eligibility bias towards persons who can afford the costs of the deductible period.
We anticipate that until the issues discussed above can be adequately addressed by the policy process, that they will serve as barriers to enactment of a Federal long-term care benefit package. Given the legitimate concerns about the costs of a Federal long-term care initiative, any proposal which cannot define the eligible population in fairly specific terms, and therefore estimate the costs of the proposal with a relatively high level of certainty, does not stand much chance of enactment. We believe that the legislation currently being proposed to expand Federal coverage of long-term care services is not adequate in this regard, and will have to become much more specific in proposing eligibility criteria for program benefits. The recent rise in Medicare SNF utilization and expenditures associated with a simple "clarification" of coverage criteria (not even an expansion) underscores the considerable impact which even a minor change in criteria can have on program costs. As well, objective measures of "need" and the allocation of benefits will always be affected by the availability of resources, so that it is important for the policy process to develop functional eligibility criteria for long-term care benefits in close coordination with the development of financing mechanisms, rather than along separate policy tracks.
Kane, R.A. and Kane, R.L. Long-Term Care: Principles, Programs, and Policies. Springer, NY, 1987, p.113.
Commerce Clearing House: 1988 Medicare Explained, Chicago, Illinois, 1988.
Latta, V.B. and Keene, R.E.: "Use and Cost of Skilled Nursing Facility Services Under Medicare, 1987." Health Care Financing Review 11(1):105-116, Fall 1989.
Health Care Financing Administration, Medicare Intermediary Manual, Part 3-Claims Process, Section 3132, Skilled Nursing Facility Level of Care--General, Transmittal 1365, December 1987.
Congressional Budget Office. CBO Reestimate of Medicare Outlays for Skilled Nursing Facility (SNF) Benefits, September 1989.
Commerce Clearing House, 1988 Medicare Explained, Chicago, Illinois, 1988.
Iverson, LH, A Description and Analysis of State Pre-Admission Screening Programs, InterStudy, Center for Aging and Long-Term Care, Excelsior, MN, March 1986.
Iverson, LH, Summary Descriptions of State Screening Programs, InterStudy, Center for Aging and Long-Term Care, Excelsior, MN, 1986.
The potential for a clinical override of the decision-rule outcome is built into most of these approaches.
Jackson, ME & Eichorn A, Feasibility of Conducting Predictive Validity Analyses on State Nursing Home Preadmission Screens, report prepared for the Health Care Financing Administration, Office of Research and Demonstrations under Cooperative Agreement No. 18-C-99213, October 1988.
Data supplied by ME Jackson. This research is presently being conducted as part of HCFA Cooperative Agreement No. 18-C-99213, Efficacy of Nursing Home Preadmission Screening.
The decision rules from each State were not able to be replicated exactly given the constraints of the Channeling database; in some instances proxy variables were relied upon. There is thus some unknown margin of error built into these results.
These findings reflect the percent of individuals (with characteristics similar to the Channeling population) who would be deemed eligible by the State screens based on their respective decision rules only; they do not take into account how eligibility is actually determined, or the nature/extent of clinical overrides.
Burwell, B.: "FY 1988 Medicaid Expenditures for Home and Community-Based Waivers." Memorandum, June 12, 1989.
Clinkscale, R.M. and Ray, S.S.: Survey of Medicaid Home and Community-Based Care Waivers: FY 1986. Medicaid Program Evaluation Working Group, La Jolla Management Corporation, December 1987.
Health Insurance Association of America, News Release: Long-Term Care Market Continues to Grow: Group Market Triples in Size, Washington, D.C., October 4, 1989.
Paying for a Nursing Home, Consumer Reports, October 1989, pp.664-667.
Health Insurance Association of America, Summary of Long-Term Care Plans Using Activities of Daily Living, Washington, D.C., September 1989.
Health Insurance Association of America, Long-Term Care Insurance Home Health Benefit Summary, Washington, D.C., August 1989.
The intactness of the informal support system is not considered when the applicant exhibits abusive/assaultive behavior and otherwise qualifies on functional grounds.
Commonwealth of Massachusetts, Executive Office of Elder Affairs, Long-Term Care Project, Demonstration of a Public-Private Partnership for Long-Term Care Insurance: Request for Proposals, October 30, 1989.
op cit., Commonwealth of Massachusetts, Long-Term Care Project, 1989.
Note that several of the legislative packages reviewed in this document may also include other than long-term care-related proposals; this review focuses exclusively on the long-term care components of the bills.
Kane RA & Kane RL, Assessing the Elderly: A Practical Guide to Measurement, Lexington Books, Lexington, MA, 1981.
- Weiner, JM & Hanley RJ, Measuring the Activities of Daily Living Among the Elderly: A Guide to National Surveys, Preliminary Report of the Interagency Forum on Aging-Related Statistics, September 1989. [http://aspe.hhs.gov/daltcp/reports/guide.htm
Katz S, Ford AB, Moskowitz RW, $ Jaffee MW, The Index of ADL: A Standardized Measure of Biological and Psychosocial Function, Journal of the American Medical Association, 185:94-99, 1996.
Individual constellations of ADL impairments do depart at times from the hierarchy. Therefore, there will be some small proportion of individuals who will display impairment in dressing, but not in bathing. For this reason, our estimate of persons with two or more impairments represents a lower-bound estimate.
op cit. Weiner & Hanley, 1989.
Katz S, Downs M, Cash M, et al., Progress in the Development of the Index of ADL, The Gerontologist, 10:20-30, 1970.
Estimates derived from tables in Weiner & Hanley, 1989, op cit.
Rubenstein L, Schairer K, Wieland GD & Kane R, Systematic Biases in Functional Status Assessments of Elderly Adults: Effects of Different Data Sources, Journal of Gerontology, 39:686-691, 1984.
Epstein AM, Hall Ja, Tognetti J, Son LH & Conant L, Using Proxies to Evaluate Quality of Life, Medical Care, 27:S91-S98, 1989.
Equally as important in developing program cost estimates will be the expected rate of participation among those eligible.
Rowland D, Help at Home: Long-Term Care Assistance for Impaired Elderly People, Commonwealth Fund Commission, May 1989.
An additional difficulty in using the MSQ approach in eligibility determination is the uneasiness which many interviewers, even trained clinicians, experience in administering the questions contained in such instruments. Their uneasiness stems from the perceived intrusiveness of the questionnaire items. If interviewers are uncomfortable with this approach the information which they gather may prove unreliable.
For example, H.R.3140 (Waxman) and the draft of Kennedy's bill.
These definitions provided by Kevin Mahoney, director of the Connecticut Public-Private Partnership for Long-Term Care Insurance funded by the Robert Wood Johnson Foundation; these behavior problems are included in the definition of the Partnership's "insured event." See Section 2.8 above.
For example, Folstein's Mini-Mental State Examination.
Based on the hierarchical nature of the Index of ADL.
Of course, the services provided under the Medicare benefit package (i.e. skilled medical and professional services) are also less likely to be provided through other sources of assistance.
Code of Federal Regulations, Section 456.370.
Burwell, Brian: Shared Obligations: Public Policy Influences on Family Care for the Elderly. Medicaid Program Evaluation Working Paper 2.1, Health Care Financing Administration, Office of Research and Demonstrations, May 1986.
In contrast, a "front-end" program would cover the cost of services from the beginning of a utilization episode. Out of necessity, a "front-end" design calls for some mechanism to determine eligibility for benefits so as to insure that those receiving services have needs commensurate with the assistance being provided.
Manton KG, A Longitudinal Study of Functional Change and Mortality in the United States, Journal of Gerontology, 43:S153-161, 1988.
Branch, LG & Ku L, Transition Probabilities to Dependency, Institutionalization, and Death Among the Elderly Over a Decade, Journal of Aging and Health, 1:370-408, 1989.