Eligibility for Publicly Financed
Home Care
t;M g u
I-ORNK", 4-'l'7'l--'-'l'-----B 1-1. ",
..
Mary E. Jackson, PhD, Brian Burwell BA, Robert F. Clark, DPA, and
Mary Harahan, AM
Introdution sure equitable access to program appli-
cants. As Stone and Murtaugh have dem-
At present, there is no uniform fed-
eral policy for financing home care serv-
onstrated, estimates for the number of
ices for elderly persons in need of long- persons with disability can range widely
term care. Eligibility criteria that depending upon the ADL definition in-
voked.'
determine who is eligible to receive pub- As nonspecific as some of the ADL
licly financed home care services varies
considerably by funding source. To a large eligibility criteria in the expansion propos-
als are, they are vastly more specific than
extent, these criteria are established at the
state level, because states have primary language pertaining to criteria that would
responsibility for allocation decisions un- establish eligibility based on cognitive im-
der many of these funding mechanisms, pairment. This lack of specificity regard-
particularly Medicaid. However, recent ing cognitive criteria is not simply a defi-
congressional proposals to expand federal ciency in policy development, but reflects
coverage of home care have stipulated the state of the art in measuring and
with increasing specificity uniform criteria screening for cognitive impairment in the
for eligibility based on functional disabil- elderly population.
ity. Typically, these criteria rely on mea- To demonstrate how cognitive im-
sures of disabilities in a given number of pairment may be operationalized within
activities of daily living (ADLs). the context of eligibility criteria, we pre-
Eligibility rules that rely exclusively sent estimates of the number of function-
on functional disability criteria have been ally disabled and cognitively impaired
criticized for excluding persons who have community-dwelling elders. We vary both
the physical capacity to perform ADLs, the definition of functional (ADL) disabil-
but not the ability to remember when to ity and the definition of cognitive impair-
initiate a task, and those who need super- ment, and show how sensitive estimates
vision or cuing in order to complete a task. can be to the criteria on which they are
These are typically individuals who have based.
Alzheimer's disease or a related disorder
and who are cognitively impaired. Most
legislative proposals for expanding federal Mary E. Jackson and Brian Burwell are with
financing for home care now include cri- SysteMetrics, Lexington, Mass. Robert F.
teria intended to capture the cognitively Clark and Mary Harahan are with the US De-
impaired population targeted for services. partment of Health and Human Services, Office
of the Assistant Secretary for Planning and
Although nearly all proposals have Evaluation, Washington, DC.
been specific about the number of ADL Requests for reprints should be sent to
disabilities necessary to qualify a person Mary E. Jackson, PhD, SysteMetrics, 24 Hart-
as eligible for services, only a few go so far well Avenue, Lexington, MA 02173.
This paper was submitted to the Journal
as to define how disability in a specific January 18, 1991, and accepted with revisions
ADL would be measured. If imple- November 12, 1991.
mented, any of these proposals would The views and conclusions presented in
this paper are solely those of the authors and do
have to develop more precise criteria for not necessarily reflect the views of the Depart-
defining an ADL disability in order to en- ment of Health and Human Services.
American Journal of Public Health 853
Jackson et al.
Methos All estimates refer to ADL disabilities conditions, the person was considered to
that were chronic in nature, that is, those be cognitively impaired.4
Data that had lasted for 3 months or longer.' The second definition of cognitive im-
Data from the 1984 National Long- Cognitive Impairment pairment, an attempt to operationalize eli-
Term Care Survey (NLTCS) were used to gibility language from some legislative pro-
generate estimates of functional disability The NLTCS was not specifically de- posals, includes evidence of behavior
and cognitive impairment. As part of the signed to estimate the number of elders problems along with the CI criteria de-
NLTCS, in-depth interviews were con- with cognitive problems; thus we were scibed above. Some argue that behavior
ducted of a sample of Medicare beneficia- able to derive cognitive impainment esti- problems should be considered in deter-
riesaged65yearsandolderonApril 1, 1984, mates only for persons who met the func- mining the need for long-term care services
who were found to have either an ADL or tional disability screen for the NLTCS because such problems require supervision
an Instrumental Activity of Daily Living
sampling frame. However, the NLTCS and oversight to ensure the health and
(IADL) limitation.2 sampling frame for identifying disabled safety of the afflicted person.
Estimates are based on the commu- community elders was quite liberal in its Wandering, frequent temper tan-
nity-resident (noninstitutionalized) sam- definition of disability. Any type of assis- trums, and compulsive stealing are the be-
ple of the NLTCS. They have been sub- tance, human or mechanical, for anyADL haviors on which the NLTCS collected
jected to a weighting procedure whereby (i.e., core ADLs plus bladder or bowel in- information, and are therefore the behav-
cross-sectional weights were applied to continence, indoor mobility, outdoor mo- iors represented in our behavior problems
adjust for nonresponse and the complex-
bility) or for any IADL (meal preparation, (BP) measure. Because the NLTCS did
laundhy, light housework, groceiy shop- not collect information on other types of
ities ofthe sampling design. Application of ping, money management, medication behavior problems, our results may un-
weights results in nationally representa- management, telephoning) that had lasted, derestimate the prevalence of behavior
tive estimates of the community-dwelling or was expected to last, for a inimum of
population aged 65 years and older. All problems among the elderly.
90 days was included in the definition. The third definition of cognitive im-
estimates presented in this paper meet Given the wide net that the NLTCS cast in
precision standards of relative standard pairment employed in our estimates is
its sampling frame of the elderly disabled, more stringent than the other two. Under
errors of 30% or less, and are based on as well as the fact that most cognitively the "CI+" measure, a person must not
unweighted cell sizes of 400 or larger. Be- impaired elders can expect to have at least only meet the CI criteria but must also
cause the estimates derived from the sur- some limitation in functioning, we are fairly either demonstrate disability (i.e., must
vey reflect the number of persons disabled confident that elders with cognitive impair- receive active or standby assistance for 3
in 1984, we have adjusted 1984 estimates ments are sufficiently represented in the months or longer) in at least one ADL or
to be consistent with 1990 population es- estimates that follow. one of the cognitively oriented IADLs
timates for the population aged 65 years Three definitions of cognitive impair- (money management, medication man-
and older. ment were developed. The first includes agement, telephoning) or exhibit a behav-
persons with an error score of 4 or more on ior problem. For a person to be considered
ADL Disability Measwres a modified Short Portable Mental Status cognitively impaired under this definition,
Estimates of functional disability are Questionnaire (SPMSQ). The SPMSQ is a there must be evidence of the need for
based on the same alternative definitions 10-item screening instrument designed to oversight or some type of assistance. The
of ADL disability as used by Stone and identify elderly persons with probable cog- CI+ criteria is an attempt to operational-
Murtaugh, and the methods used to derive nitive impairment.3 A score of 4 or higher, ize language in proposed legislation that
them are replications of the methods of rather than the customary 5 or higher, was stipulates that a person with Alzheimer's
Stone and Murtaugh.I The estimates were used to indicate cognitive impairnent be- disease or other cognitive impairments
cause one of the SPMSQ questions was must demonstrate needs similar to those
developed from specific questions posed altered in the NLTCS, resulting in fewer
to NLTCS respondents or their proxies of persons meeting ADL criteria in order
errors than the original item typically elic- to be deemed eligible. The CI+ criteria
regarding the types of help the respondent its. The cutoff point indicating cognitive
received in five core ADLs: bathing, operationalize such criteria by requiring
impairment was adjusted downward to ac- some evidence of the need for care be-
dressing, toileting, transferring, and eat- commodate this modification. This mea-
ing. Receipt of services is used in these yond a diagnosis or a score on a cognitive-
sure of cognitive impairment is referred to impairment screen. A diagnosis or a score
analyses as an indicator of disability. as "Cl" in Table 1.
The NLTCS asked multiple questions may not reflect the fact that an impairment
SPMSQ data are missing for those is severe enough to require home care
about functioning in each of the five ADLs. sample members who were not capable of
Disability in ADL was defined in three dif- services.
responding to the questionnaire and
ferent ways, according to the type of as- whose information was provided by a
sistance required to perform a given ADL. proxy. To identify sample members with Estimaes
* Receives active assistance to per-
missing SPMSQ information who were Table 1 presents estimates of the
form an ADL cognitively impaired, alternative criteria number of disabled community-dwelling
* Receives active or standby assis- were developed. If a person with missing
tance to perform an ADL
elderly based on alternative ADL as well
SPMSQ information was judged to be se- as alternative cognitive impairment crite-
* Receives active or standby assis- nile by the respondent's proxy, or if the ria. In each section, the first column rep-
tance to perform an ADL or uses proxy reported that the person had a dis- resents the number of persons estimated
an assistive device in performing ability caused by disorders classified as to be disabled based on functional criteria
an ADL mental retardation or senile psychotic alone. The remaining columns reference
854 American Joumal of Public Health June 1992, Vol. 82, No. 6
Elibity for Publidy Financed Home Care
estimates of the number of ADL-disabled
or cognitively impaired persons; the dif-
ferences in estimates between the last
three columns are a function of the defi-
nition of cognitive impairment invoked.
For example, the estimate in the first
column of section A for 2+ ADLs repre-
sents the number of elders estimated to
have two or more ADL disabilities, where
disability is defined as the need for the
active assistance of another person to per-
form the ADLs in question.
The estimate in the second column
references the number of elders who
would be eligible for services if the crite-
rion were 2+ ADLs or CI or BP. This
decision rule targets individuals with two
or more ADL disabilities (active assis-
tance only) or those who are cognitively
impaired (those who score 4+ on the
SPMSQ or who have a cognitive impair-
ment-related diagnosis) or those who ex-
hibit one or more behavior problems
(wandering, frequent temper tantrums, or
compulsive stealing). The rationale for
considering those with behavior problems
within the cognitive impairment group is
that behavior problems are considered an
indicator of cognitive impairment, and
those with behavior problems are thought
to need oversight and supervision to en-
sure their health and safety.
Slightly more than 2 million elders
would meet the 2+ ADLs (active only) or
CI or BP criterion in 1990. Approximately
1 million additional persons (or 3.5% more solely in terms of evidence of cognitive bility for services be evidenced by an
of the elderly population) would be eligible impainnent, as exemplified by the CI cri- SPMSQ score indicating cognitive impair-
for services if this criterion were used in- terion. With this criterion, fewer individ- ment (or a cognitive impairment-related
stead of comparable criteria that exclude uals would fall into the eligibility pool. Al- diagnosis) and some evidence of the need
the cognitively impaired from consider- though using the cognitive impairment for care or oversight. Such care or over-
ation. criterion instead of the CI or BP criterion sight might be required because of (1) one
The "Cl or BP" criterion is the most does lower the estimates, the decrease is or more ADL disabilities or (2) a disability
liberal of the three cognitive impairment only on the order of 0.2% to 0.5%, or in one of the three cognitively oriented
criteria included in the estimates; it clearly 60 000 to 150 000 fewer eligible persons IADLs (money management, medication
yields the highest esimates of all the mea- nationwide, depending upon the exact management, or telephoning) or (3) dis-
sures. However, this particular criterion ADL criteria invoked and the definition of play of behavior problems.
may result in an overestimate of the cog- disability used. The CI+ criterion is undoubtedly the
nitively impaired population because ofthe This approach, though more conser- most restrictive of the three cognitive im-
way in which the behavior problem factor vative, is still subject to criticism. A score
figures into the criterion. For example, un- on a cognitive impairment test, or even a pairment criteria. For example, 6.8% of
derthe 2+ ADLsorCIorBP rule a person diagnosis of Alzheimer's disease, does not the nation's elderly would be service-eli-
without any ADL disabilities, who is not necessarily correspond to a need for home gible under the 2+ ADLs (active) or CI or
considered cognitively impaired according care. The target population for home care BP criterion, and 6.5% under the 2+
to the aI criterion, but who compulsively services should probably not include per- ADLs or cognitive impairment criteria,
steals, would be deemed eligible. A person sons for whom organicity has not yet im- but only 4.8% under the 2+ ADLs or CI+
with this particular profile may not be an peded independent functioning of ADLs criterion. Yet this 4.8% estimate is sub-
appropriate candidate for home care serv- or presented a threat to the safety of the stantially higher than would result from
ices, and including such persons in an eli- individual. This criticism has led to the the same criterion without the cognitive
gibilitypool mayresult in misspecfyingthe suggestion that eligibility for home care impairment component added. This trans-
targeted population, even under very ib- services for cognitively impaired persons lates into an additional 450 000 eligible
eral targeting policies. be linked to more specific evidence of the persons under the 2+ ADLs or CI+ cri-
A more conservative approach need for care. The CI+ criterion responds terion who are eligible because of the cog-
would be to define cognitive impairment to this suggestion by requiring that eligi- nitive impairment component alone.
June 1992, Vol. 82, No. 6 American Joumal of Public Health 855
Jacksn det aL
Diussion ployed in legislation to expand federally simple matter. In this context the question
financed home care services will be major must be raised whether it is possible to
In this study, we have demonstrated factors in determining the ultimate costs of enact federal legislation that would ensure
how various eligibility criteria can be used such an expansion. A second implication that benefits are allocated in an equitable
to derive estimates of the disabled elderly is that unless the eligibility criteria desig- manner according to true need. Unlike ac-
population living in the community, taking nated in the legislation are specific in their cess to publicly financed acute care serv-
into account those in need of services be- definitions, the number of persons who ices, which is determined primarily by
cause of cognitive impairment. We have will actually qualify for benefits will de- physicians, access to expanded federal
shown how these various criteria affect pend largely upon how the general lan- home care benefits will require policy
the size of the population that would be guage of the legislation is interpreted in makers to establish relatively specific de-
made eligible for benefits under alterna- regulatory and administrative procedures cision rules about who is eligible for ben-
tive definitions of functional disability and efits and who is not.
cognitive impairment. during program implementation. This
It must be recognized that these es- could lead to considerable discrepancy
timates are derived from survey data, and between the intent of the legislation and its Acknowledgents
incentives for survey respondents to over- actual implementation. Funding for this study was provided under con-
state their disabilities are minimal. Survey Definitions of disability using mea- tract HHS-100-88.0041 from the Department of
sures of ADL impairment have become Health and Human Services, Office of the As-
respondents, unlike program applicants, sistant Secretary for Planning and Evaluation,
receive no benefit from exaggerating their popular mechanisms for allocating home Division of Disability, Aging, and Long-Term
disability. In fact, there is some evidence care resources under proposed federal Care.
that survey respondents may underesti- legislation. However, the need for pub- We would like to acknowledge the assis-
licly financed home care depends upon a tance of William D. Spector, PhD, and Peter
mate their disability levels.5 Thus we Kemper, PhD, from the Agency for Health
would expect some increase in self-re- variety of factors, not just one's level of Care Policy and Research for providing some of
ported disability if persons were applying disability. Financial need is an obvious the analytic files used in these analyses. We
for services and knew that their receipt of factor, and one that is also built into many would also like to thank Sidney Katz, MD, for
service was contingent upon a disability legislative initiatives. However, it is also his review of an earlier draft of this manuscript.
measure. Additionally, when service eli- true that the elderly adjust to their disabil- The technical assistance provided by Karen
Koch of Social and Scientific Systems and Sean
gibility is predicated on the need for hu- ities through a multitude of adaptive be- Kennedy of SysteMetrics is also greatly appre-
man assistance in performing ADLs, haviors, just as nonelderly persons with ciated.
there may be a disincentive to substitute disabilities do. In addition, many elderly
assistive devices for human help, thus ren- persons have access to support from fam- References
dering the impaired more dependent than ily and friends that can augment the formal 1. Stone R, Murtaugh C. The elderly popula-
tion with chronic functional disability: im-
necessary and also jeopardizing the devel- provision of services. Given the variety of plication for home care eligibility. Gerontol-
opment of new technologies for fostering factors that contribute to the need for ogist. 1990;30:491-496.
independence. home care, should a federal entitlement 2. Manton KG. A longitudinal study of func-
Nonetheless, the estimates presented program for home care services, based tional change and mortality in the United
in this paper demonstrate how substan- States. J Gerontology. 1988;43:S153-161.
solely on standardized measures of dis- 3. Pfeiffer E. A Short Portable Mental Status
tially different estimates of the disabled ability, be implemented? What role, if Questionnaire for the assessment of organic
elderly population with and without cog- any, should the availability of family and brain deficit in elderly patients. JAme,ican
nitive impairment, can be generated by friends, as well as financial status, play in Genatrics Society. 1975;23:433-441.
employing varying definitions of disabil- 4. Spector W. Cognitive impairment and dis-
eligibility determination and service allo- ruptive behavior among elderly persons liv-
ity. The use of more stringent or more cation? ing in the community: implications for tar-
liberal eligibility criteria for home care In any case, it is apparent that the geting long-term care. Gerontologist. 1991;
services will have dramatic effects on the establishment of criteria to determine 31:51-59.
number of elders who would qualify for which elderly persons with disabilities 5. Ford S, Folmar SJ, Salmon RB, Medalie JH,
services. The most obvious implication is should be eligible for publicly funded Roy AW, Galazka SS. Health and function
in the old and very old. JAmencan Geriat-
that the specific eligibility criteria em- home care and which should not is not a rics Society. 1988;36:187-197.
856 American Journal of Public Health
June 1992, Vol. 82, No. 6