Robert Newcomer, PhD

Professor and Chair, Department of Social & Behavioral Sciences
University of California. San Francisco

Monitoring Trends and Case Mix in Residential Care Facilities and Other Forms of Group Housing

Mr. Chair and other members of the committee.

I appreciate the opportunity to speak to you about a major component in the long term care/chronic care delivery system, and the inadequacy of current data systems for monitoring the case mix and operational performance. Potential changes in public policy with respect to vendor and client identification offer the opportunity for a simple change in data reporting and a vast augmentation of public health data. My initial comments include a brief on why this is an important public policy issue, and the limitations of existing data systems.

• About a million persons live in nursing homes, and at least a half a million live in licensed assisted living, residential care facilities, or board and care homes in the U.S. An equal number live in unlicensed boarding homes. Both licensed facilities (herein referred to as RCFs) and unlicensed facilities include the provision by a nonrelative of shelter, meals, and such other tasks as cleaning the residents' room and laundering linens. Licensed facilities are also usually responsible for helping with transportation and shopping; supervising residents' medication; assisting in obtaining medical and social services; and on a limited basis assisting with dressing, grooming, eating, bathing, and transfers. Some RCFs provide assistance for special needs, such as those using oxygen and assistive devices, or with cognitive impairments.
• State governments, in statutes and regulations, differentiate the levels of care that can be provided in RCFs versus nursing homes. However, the operational boundary between levels of care within RCFs, or between RCFs and nursing homes, is more easily established in regulation than in practice. One approach to such problems are licensing standards that directly specify the level and type of personal care available. Some states have personal care definitions that are tied to "hands-on care" or assistance in activities of daily living; while others expressly restrict this concept to housekeeping or general protective oversight--excluding care of incontinence or help with transferring. Common regulatory solutions are limits on the number of days of "higher levels of care", and requirements that the facility obtain any needed health and personal care services from outside licensed home health care agencies.
• In part because of the ambiguity between who can be appropriately served in RCFs, a number of states have begun to modify their RCF regulations to facilitate the use of this level of care in order to reduce or replace time as a nursing home resident (Mollica, et al., 1992; Newcomer & Lee, 1994). The trend is for much fewer condition exclusions. The net result, that at least for a subset of facilities, is that the level of care distinguishing a nursing home from RCF resident (patient) is being bed bound or needing 24-hour skilled nursing.
• Proponents for an expanded role and scope of services for residential settings argue the cost savings of a "non-medical" approach to care, and the necessity of "normalizing" the living situation for the disabled older person, including many who may qualify for admission to nursing homes. Critics of expanded and enriched levels of care in RCFs raise concern with resident safety and the adequacy of care in such settings (e.g., US GAO, 1989; Feder, et al, 1988). They fear that RCF settings may become less well staffed versions of nursing homes, with the result that health conditions may deteriorate even further.
• In spite of the public policy concerns about the quality of residential care, and the seemingly rapid evolution of state policy and industry practices, relatively little is known about who is being served by RCFs within specific states and market areas, and how the resident mix is influenced by state nursing home and residential care regulation and reimbursement policies.
• One reason for the absence of information is that the licensing, regulation, quality assurance, and financing of this industry remains largely outside the established management systems for long term care. Consequently, case assessments and case management data are not compiled into data systems.
• Another reason is that the information potentially available, such as from hospital discharge abstracts, Medicare and Medicaid claims files, and national survey monitoring sytems generally do not identify the population living in residential care or other forms of group housing.
• In many situations this may be problematic. Ambulation-impaired RCF residents illustrate the complexity of the policy response to these perspectives. Can one provide an appropriate level of care to such individuals simply by upgrading staff, service mix, and environmental standards of RCFs? Is the answer a ban on entry or occupancy by persons with mobility or cognitive disabilities? How does this work for short-term disability? What effect do preadmission screening criteria and financing for nursing home care have on the access to such care? If access is more limited, how does "demand" spill over into RCFs or community care?
• The emphasis of my comments will be on data systems. Studies of RCF residents have largely relied on case studies and convenience samples of facilities and residents (see for example Hawes, et al., 1995; Newcomer, et al, 1994; Mor, et al; 1986; Dittmar & Bell, 1983; US DHHS, 1982). As a consequence health status, functional and cognitive ability among RCF residents is not well documented across the country. This contrasts with the substantial data base on nursing home residents. These data are compiled as part of the nursing home minimum data set (MDS). Instruments used for the MDS cover topics ranging from functional abilities, to medical problems, and emotional states (such as depression and behavior problems). The MDS is collected on all residents at or near the time of admission, upon readmission from a hospital, if there is a significant change in status, and quarterly.
• There are short term and long term strategies for rectifying this. Among the simplist is the use of hospital discharge abstracts. The tracking of hospital discharges associated with such conditions as skin ulcers, nutrition, dehydration, injuries, drug or medication poisoning could be a basic first order indicator system for problems in the long term care delivery system. All that is needed in these data systems is information connecting the hospital patient to their address or location prior to hospital admission.
• Presently discharge records are much better at indicating the post discharge destination than in identifying where someone came from. The most common prior location is the emergency room. This is a solvable problem.
• Nursing home MDS+ data systems can similarly be used to identify prior residence and targeted conditions at the time of admission. Distinguishing direct admissions to nursing homes from hospital transfers would provide a pretty complete picture of inpatient stays associated with breakdowns in chronic care delivery.
• Data systems of this sort provide end point information only. Getting beyond that to more subtle changes in the delivery system is also desirable for monitoring changes in case mix and staff performance. Without such data, it is difficult to isolate problems, identify trends, and forecast or evaluate the results of policy changes.

-- One approach to this, an expensive alternative, would be to replicate the MDS+ process in which all RCF admissions would be given assessments.

• - A second approach would be to better connect the state's long term care management structure to this industry--assigning residents or facilities to individual case managers who monitor the clinical performance of the facility. In effect this is being done on a non-systematic basis by using home health agencies and hospice providers as providers for the eligible residents, but information on the quality of this care is not being compiled.
• - A third approach would be to use Medicare and Medicaid claims data to monitor the diagnoses, and health care utilization of RFC, nursing home, home and community based care recipients. Doing so requires much more accurate information on the resident's address than currently available, and it requires that managed care systems provide patient level claims data--something missing in current data systems.
• - A fourth type of approach could be to use national, state, or community based surveys for general trends monitoring.
• - These approaches are not mutually exclusive.

· In my remaining time I will concentrate my comments on the current inadequacies of national data systems. My first concern about these data systems is that they yield only national or regional estimates. This misses the wide local variation in resident case mix and quality of care that may result from the competition among levels of care (i.e., supply of nursing home beds, residential care beds, and home based care), and the effects of policies and local policies (e.g., per diem rates for nursing home and RCF care, regulatory standards affecting quality of care monitoring, and placement eligibility for RCF care) within the study's market areas.

• Getting community level estimates can be done with samples of communities, such as used in disease registries or public health surveillance system. Communities could be selected as "indicator" sites to track the effects of policy and economic shifts as they affect this industry.
• A second, yet fundamental limitation of current national surveys is the definition of group housing used. The term "Group housing"encompasses many types of living situations. Among them are institutions such as nursing homes, mental health hospitals; and noninstitutions such as rooming homes, communes, residential care facilities, homes for the aged and disabled, halfway housing. To help simplify matters, local zoning, fire and safety ordinances, and state regulations typically define group housing using a standard of five or more unrelated persons living together. Fewer than five unrelated individuals living together or any number of related persons living together, is usually considered as an individual household. The level of services provided by the housing facility or operator may also determine if this living arrangement needs to be licensed as a form of special care facility. States vary substantially in the criteria used to determine whether a facility needs to be licensed.
• The somewhat arbitrary decision rules that differentiate independent living from group housing have had major implications for how the U.S. Census is conducted. This in turn has affected the sample design and generalizability of many other surveys of the aged and disabled population. The net result is that both national and community level information about the housing and living arrangements of persons with disabilities may be substantially biased--particularly among persons living in the so called group quarters.
• Both licensed and unlicensed facilities vary in their architectural design. This further complicates how such units are classified for many survey samples or how the data are compiled. Facilities may offer independent apartment units, with shared or common services available on site. Alternatively, they may offer only single or shared bedrooms, with full access to kitchens, bathrooms and other features common to any home. The former case may be defined as independent living or as an institution, depending on the licensing status, but it would likely not be defined as group housing. The latter situation would almost invariably be defined as group housing, if there were five or more unrelated individuals present. Identical facilities with fewer than five residents would be defined as independent housing (unless the facility were considered an institution).
• These definitional complexities have reduced the effectiveness of all the major US surveys of health status and housing, resulting in less than perfect information about the living situation of persons with disabilities.
• Over 75 survey instruments were evaluated and an extensive literature review was conducted to identify potentially relevant data sources. These materials were supplemented with discussions with agency personnel involved with the major data sources to further clarify the survey methodologies. Limitations in sample size, sample frame, frequency of collection, and measurement of living arrangements or disability status precluded most of the surveys from extensive consideration. Appendix 1 provides a synopsis of the principle data sets reviewed.
• Appendix 2 is a synopsis of the subset of surveys which received indepth evaluation. Five of these are reviewed more fully in Section 2 of this paper, because of time I will not comment on them in these remarks. In combination, these surveys offer excellent options for the measurement of disability, sample designs for identifying alternative living arrangements, approaches for monitoring housing choices, and measures of housing type and living arrangements.
• Five factors contribute to the current inadequacies in these data sets:

-- Surveys which provide extensive housing data have little or no information on disability, health status, and the services available within the housing.

• - Surveys that have extensive health and disability data provide little or no housing or living arrangement data.
• - Surveys either attempt to exclude institutions and some forms of group housing to limit their sample frame to the noninstitutionalized population; or they rely on proportionate weighting of the population in "group settings".
• - Both these sample frame designs produce either an under enumeration of the disabled population, or a sample of specialized living situations that may be too small for detailed subgroup analysis.
• - The definitions used for the enumeration and selection of eligible housing units are problematic, leading to likely problems in inappropriate sample exclusions.

Recommendations

• First priority is to make recommendations for improving the prior or current address and provider code data on Discharge Abstracts, and Medicare Claims files.
• Recommend to DHHS that NCHS or the Bureau of the Census organize and convene a Policy Forum to discuss how to modify sample frame rules used to classify institutional and noninstitutional group housing and specialized living arrangements for the disabled and aged in the US Census, the National Health Interview Survey, the American Housing Survey, and the Medicare Current Beneficiary Survey;
• This committee or forum should also discuss possible measures for classifying special supportive living arrangements within the American Housing Survey and the Medicare Current Beneficiary Survey, and describing the services provided or available to residents; and
• Discuss possible measures for health, and disability status items which could be incorporated into the American Housing Survey
• Discuss the feasibility of organizing survey tabulations or files so that living arrangements of the disabled can be more fully and explicitly analyzed. For example, given the trend in the disabled community toward less restrictive living arrangements (that is, away from institutional settings and toward informal models such as shared housing) it seems appropriate that sample frame rules (and data tabulations) should include any housing situation in which two or more disabled individuals, or a single disabled person and an unrelated caregiver, are housed irrespective of services offered or legal structure pertaining to the facility's licensing status.

This concludes my prepared remarks.

American Housing Survey

The American Housing Survey (AHS) is conducted by the American Housing Survey Branch of the Bureau of the Census. It is conducted every other year. In 1995, the sample consisted of approximately 61,000 units. The sample areas, called Primary Sampling Units (PSUs) are stratified by region, and urban/rural locations. Both national estimates and estimates of selected metropolitan areas are available. Census Enumeration Districts are sampled from within the PSUs. Housing units form the sampling unit. An important feature of these surveys is that generally the same housing units remain in the sample year after year, and it is the housing unit rather than its occupants that is followed.

Strength

This data set has extensive information on housing stock, including type of housing, living arrangements, and quality of housing. Its housing typology is particularly useful. Its coding includes categories for "non-institutional, special places" (see Appendix 3 ,Section A, pages C2-63 through C2-80). Sub-categories are quite detailed and have extensive descriptions. Types that may be relevant to analysis of supportive housing for the disabled include Boarding House (code 76); Combination Boarding and Rooming House (code 76); Commune (code 92); Non-transient Hotel or Motel (code 71); and Independent Group Residences for senior citizens, the handicapped and functionally disabled (code 94).

The AHS has further advantages in its nationwide coverage, stratified, random sampling, frequency of application, consistency with census tracts (from which sampling is done), and detailed definition of household membership. Further, information on households with SSI and disability income programs is collected in detailed categories. In addition to housing data, the AHS currently solicits a significant amount of demographic, financial and geographic data, including education, marital status, cost of housing, and transportation use. A third possible strength of the AHS is that the sample (approximately 61,000 units) is large enough to identify a statistically significant sample of non-aged disabled residents, but it may require an over sampling of special units to permit a generalization to this subset of housing types.

Weakness

The most important limitation of the AHS instrument is that it does not contain any items on the disability status of residents. A second problem stems from their definition of "non-institutional special places", and a definition of group housing based on the criterion of five or more unrelated persons living together. The special places classification includes potentially relevant categories of housing, such as "rooming house", "commune", "rest home", and residential care, but the delineation and differentiation of housing units into these various classes is affected by interviewer interpretation of a particular facility's fit with the operant definition, and state variation around the classification of facilities. The absence of a individual housing unit within one of these special places would render the housing unit ineligible for the survey, thus affording no information on the inhabitants. There is not comparable survey information on these ineligible special places. This decision rule applies only to facilities of greater than five beds.

Possible Adaptations

• Add a health and disability status section to the AHS interview.
• Add a section to the instrument to obtain information on supportive services in special units.
• The definitional and data collection techniques of the AHS and the Decennial Census need to be refined and coordinated.
• Develop a subsample of specialized housing units to permit the national profiling of trends in these settings.

National Health Interview Survey and Disability Supplement

The National Health Interview Survey (NHIS) is a continuous national survey sponsored by the National Center for Health Statistics. It provides data on acute conditions, limitations of activities, injuries, occupational disability days, physician and dental visits, and selected chronic conditions. It also provides standard demographic information and modest housing data.

NHIS uses a probability sample of civilian non-institutionalized households. Sampling is done by the Bureau of Census, it consists of approximately 50,000 households representing 130,000 individuals. The sample is stratified to reflect regions and metropolitan areas. There are 210 Primary Sampling Units (PSUs), from which Census Enumeration Districts are selected. Small land areas or groups of addresses are assigned for interview. Data is collected on all household members within this area probability sample of housing units. Some sample units are located in places with special living arrangements such as dormitories, institutions, convents, or mobile home parks. These types of living quarters are specified as special places. Units defined as institutions are excluded from the sample.

Most NHIS survey waves include supplemental surveys for special groups identified during the primary survey. In 1994-95, a supplemental survey on disability was conducted to obtain additional information on the living arrangements and caregiver resources of this population. The public data is expected to be available in 1996. The survey instrument for the supplement includes questions on ADLs and IADLs, mental retardation and developmental disabilities, physical, emotional and cognitive impairments, transportation, employment barriers, home access and accommodation, and services utilization. (Refer to Appendix 3, Section B.)Strength

The NHIS in general and the Disability Supplement in particular, offer excellent data on health and disability status and many related areas. Because of the main surveys' continuous sampling, it is useful for the cross sectional monitoring of changes in condition prevalence for a wide range of conditions and among various ethnic and economic subgroups. The large sample size used in this survey assures a minimal standard error for the prevalence estimates for major conditions and subpopulations.

Weakness

While the basic instruments have some questions on housing type, the categories do not capture the variations in supportive living arrangements, such as persons living with others having a disability, or in specialized housing. Like the American Housing Survey, there is variation among enumerators in the inclusion of licensed supportive housing units. This, occurs, in part, because of differences in state licensing requirements and by the discretion afforded the enumerators in defining eligible and ineligible housing units. The instrument used for Supplemental Survey of Disability provides extensive information on support services attendant to those with impairments, but the standard instrument does not. Even in the special instrument, many of the quality of care items apply to the age 70+ population only.

Possible Adaptations

• Implement a regular schedule for a fielding of the disability supplement.
• Add a more complete housing component, perhaps using the AHS format for categorizing special places housing within the disability supplement.
• Extend the quality of care questions to all ages among the disability group.

Decennial Census of Population and Housing

The Decennial Census encompasses the entire US population. It provides data cumulated at the census tract level on household demographic (e.g., age, education level, race, ethnicity, sex, marital status, migration/mobility), income, income source, labor force participation, occupation, and type of dwelling, (e.g., home or apartment, mobile home, trailer, house, apartment, and qualified group housing). The number and relationships of persons in each household is also compiled. Included among the many tabulations and tape files is a summary statistics file representing a 100% count of persons in "group" quarters. Group quarters include persons in institutions and noninstitutions.

Strength

The major value of the Decennial Census for the present purposes is its 100% enumeration of the civilian population. This yields a basis for community level analyses and trends monitoring, and the application of synthetic estimates of condition prevalence, and service use.

Weakness

The living arrangement categories, especially for specialized housing, preclude extensive application of census data for those in special living circumstances. For example, it the Decennial Census uses two major housing categories: regular housing units and Group Quarters. The Group Quarters category is further in divided into institutional and non-institutional components. Institutional Group Quarters includes "skilled nursing facilities", "homes for the aging", "ICF/MR" facilities, "halfway houses/correctional", "long term care rooms", and "residential treatment centers" among other. These can be of any size. Non institutional Group Quarters include "rooming/boarding homes", "MR", "mentally ill group homes" and "non transient hotel/motel" and must have ten or more unrelated residents. Non institutional Group Facilities with less than ten unrelated individuals are categorized as Regular Housing Units. This leads to a significant under-counting of supportive housing facilities as many , especially of the unlicensed variety, are known to be small in size. (Refer to Appendix 3, Section C.)

Further complicating this issue is that the methodology for identifying facilities into group housing is less than rigorous. A facility with nine unrelated residents may be listed by one enumerator as a "home for the aged", but considered a boarding home with less than ten persons and thus put in regular housing units by another. Similarly, an apartment in a supportive housing complex is considered as independent housing, while residents with private or shared rooms in similar facilities would likely be defined as in group housing. Some of this confusion could be reduced if the services provided or available in Group Quarters were specified.

Another limitation of the Census is that it collects no significant direct information on health, disability or functional status.

Possible Adaptations

• The most practical change would be to clarify the enumeration process for the multitude of specialized living arrangements, and to create data files and tabulations that are specific to these settings.
• The housing or living arrangement categories used here should be consistent with those in the American Housing Survey. This The Bureau of Census conducts the AHS, the HIS and the Decennial and Current Population surveys, reducing the barrier to the creation of common definitions and cross-identification.
• Consideration should also be given to adding questions applicable to specialized living settings that describe facility physical and staff characteristics, as well as the service provided.

Medicare Current Beneficiary Survey

The Medicare Current Beneficiary Survey (MCBS) is sponsored by the Health Care Financing Administration as an ongoing, multipurpose survey conducted among a probability sample of Medicare beneficiaries. Also collected are data on health conditions, functional status, health care utilization and demographics, including income and vital statistics (Refer to Appendix 3, Section D). To date, there has been only one survey round in which there was a supplementary interview conducted with a particular set of survey recipients, in this case for those receiving home health care.

The sample is derived from 107 Primary Sampling Units (PSUs) consisting of metropolitan areas and clusters of non-metropolitan counties. Separate strata are compiled for the disabled under the age of 65 and for five cohorts over the age of 65. The strata are of equal size, producing a total baseline sample of twelve thousand persons. All cases receive an initial interview and a reinterview every four months through a total of twelve interviews over thirty-six months of tracking. The original panel was first interviewed in 1991. Since 1994, there has been a phased supplemental panel of new beneficiaries. Phasing is designed to replace one third of the panel annually, thus assuring maintenance of a longitudinal panel and an ample sample for cross sectional analysis.

Strength

The MCBS provides longitudinal data on multiple age cohorts. Being focused on health care utilization, it provides extensive data on health and disability status as well as some types of service utilization. Unlike most other national surveys the MCBS includes both community and institutional populations. These are selected proportionate to their distribution among all eligible beneficiaries. Living arrangement data are collected for both those in the community and those living in institutions.

Weakness

A primary limitation for purposes of supportive living arrangement monitoring is that the housing categories used in the survey do not fully delineate the range of alternative facilities, both licensed and unlicensed, that characterize specialized living arrangements. Sample size, particularly of the non-aged disabled and age groups under age 80, is not large enough to capture a sizable number of specialized living arrangements using the current proportionate weighting methods. This limits the ability to analyze living arrangements by housing subtypes. Also, since this survey covers only Medicare beneficiaries it excludes the disabled population that has not qualified for this program.

Possible Adaptations

• The addition of a more delineated housing classifications would enhance the utility of MCBS for tracking supportive housing arrangements.
• Similarly, the ability to organize living arrangement information by the health and functional status of household members would be useful in differentiating licensed from unlicensed supportive housing arrangements.
• Periodic supplemental surveys of specialized housing should be considered as a means of expanding knowledge of the population living in such settings.
• Given the primary intent of the MCBS in dealing with its service population the problem of excluded disabled populations does not seem easily rectifiable.

Social Security New Beneficiary Survey

This survey provides longitudinal data on disability and aging for new beneficiaries under Title 11 (workers with disabilities). It is sponsored by the Social Security Administration. Personal interviews with a random sample of non-institutionalized beneficiaries and their spouses are conducted. The size of the sample in the first wave was 13,962 including 12,128 original sample persons and 1,834 widows/widowers of original sample persons. The first wave was collected in 1982 by SMSA The second wave was conducted in 1992, and was collected nationally. Data collected include demographics, employment and income, health conditions that limit ability to do work, and limitations of activities (ADL/IADL). Also provided are data on long term care in residential facilities and financial data on recipients.

The Social Security Administration conducts a number of other surveys on its recipients, such as the Continuous Disability History and the Supplemental Security Income Stewardship Survey. These all may have some utility in studying the supportive housing needs of the disabled but have sample design or measurement flaws that are similar or worse than those of the New Beneficiary Survey (Refer to Appendix 3, Section E).

Strength

This survey is focused strictly on the disabled population. It provides a relatively large sample size of almost 13,000 individuals. The survey provides excellent detail on disability limitations and health status.

Weakness

Like the other major surveys described earlier, the New Beneficiary Survey does not adequately distinguish among the alternative specialized living arrangements and service categories. Because it surveys only qualified beneficiaries, it excludes all the disabled who have not qualified for benefits under Title 11. Finally, the survey is designed to track the experience of individuals when they first come into benefit. Consequently, this survey does not necessarily reflect the situation of the full disabled population.

Possible Adaptations

• The housing, living arrangements, and services section of the instrument should be brought into alignment with whatever changes are made in the Medicare Current Beneficiary Survey.
• As with the Medicare Current Beneficiary Survey, the Social Security Beneficiary Survey is directed only at a specific service population. Therefore addressing the problem of excluded individuals cannot be easily addressed, nor are we recommending that this issue be addressed.

Bibliography

Bishop, C.E., Pollakowski, H.O., and Weisbrod, G.E., "Living with Others as an Alternative to Nursing Home Use," Journal of Applied Gerontology, vol. 10, no. 3, September 1991. p. 274-293.

Corder, L.S. and Manton, K.G. "National Surveys and the Health and Functioning of the Elderly: The Effects of Design and Content" Journal of the American Statistical Association, June 1991, vol. 86, no. 414. p. 513-525.

Dittmar, N. & Bell, J. (1983) Board and Care for Elderly and Mentally Disabled Populations: Final Report, Volume 2. Denver, CO: Denver Research Institute, Social Systems Research and Evaluation Division, University of Denver.

Elkind, J. "The Incidence of Disabilities on the United States" Human Factors, 1990, vol. 32(4). p. 397-405.

Feder, J. and Scanlon, W. "People with Disabilities: A Review of Current Knowledge," Center for Health Policy Studies, Washington DC, August 1992.

Feder, J., Scanlon, W., Edward, J. & Hoffman, J. (1988) Board and Care: Problem or Solution? Washington, DC: Center for Health Policy Studies, American Association of Retired Persons.

Gioglio, G. and Jacobson, R. "Demographic and Service Characteristics of the Rooming Home, Boarding Home and Residential Health Care Facility Population in New Jersey," Trenton: Bureau of Research, Evaluation and Quality Assurance, Division of Youth and Family Services, Dept. of Human Services, 1984.

Hawes, C., Mot, V., Wildfire, J., Iannacchione, V., Lux, L., Green, R., Greene, A., Wilcox, V., Spore, D., Phillips, C. (1995) An Analysis of the Effect of Regulation on the Quality of Care in Board and Care Homes: Executive Summary. Research Triangle Park, North Carolina: Research Triangle Institute.

Kane, R. A. & Wilson, K. B. (1993) Assisted Living in the United States: A New Paradigm for Residential Care for Frail Older Persons? Washington, DC: American Association of Retired Persons.

Liu, K., McBride, T.D., and Coughlin, T.A. "Costs of Community Care for Disabled Elderly Persons: The Policy Implications" Inquiry, vol. 27, Spring, 1990. p. 61-72.

Litvak, S. . Presentation to the National Council on Independent Living, 1990, from the World Institute on Disabilities

Manton, K.G., Corder, L.S., and Stallard, E., "Estimates of Change in Chronic Disability and Institutional Incidence in the U.S. Elderly Population From the 1982, 1984 and 1989 National Long Term Care Survey" Journal of Gerontology:SOCIAL SCIENCE, vol. 48, no. 4, p. S153-166

McCoy, J.L. and Conley, R.W. "Surveying Board and Care Homes: Issues and Data Collection Problems," Gerontologist, vol. 30, no. 2, 1990. p. 147-153

Mollica, R. L., Ladd, R. C., Dietsche, S., Wilson, K. B. & Ryther, B. S. (1992). Building assisted living for the elderly into public long term care policy: a technical guide for states. Portland, ME: The Center for Vulnerable Populations, National Academy of State Health Policy.

Mor, V., Sherwood, S. & Gutkin, C. (1986). A national study of residential care for the aged. The Gerontologist 26 (August): 405-417.

Newcomer, R., Breuer, W., & Zhang, X. (1994) Residents and the Appropriateness of Placement in Residential Care for the Elderly: A 1993 Survey of California RCFE Operators and Residents. San Francisco: Institute for Health & Aging, University of California.

Newcomer, R. & P. Lee. (1994) State Innovations in Residential Care for the Elderly: A Profile of Best Practices in Range of Care, Placement Control, and Financing. San Francisco: Institute for Health & Aging, University of California.

Newcomer, R.J, and Grant, L.A. "Residential Care Facilities: Understanding their Role and Improving their Effectiveness," In Aging in Place, Chapter 5, D. Tilson, ed. Scott, Foresman, Glenview, IL

Newman, S. J. "Housing Policy and Home-Based Care," Milbank Quarterly, vol. 73, no. 3, 1995. p.407-441

Prosper, V. "A Review of Congregate Housing in the United States" New York State Office for the Aging, 1987 p.254-281

Research Triangle Institute. "Analysis of the Effect of Regulation on the Quality of Care in Board and Care Homes" Executive Summary, July 10, 1995 , Research Triangle Park, NC

Scott, C.G., "Living Arrangements of SSI Recipients," Social Security Bulletin, July 1990, vol. 53, no. 7. p.14-25

Speare Jr., A. , Avery, R., and Lawton, L. "Disability, Residential Mobility and Changes in Living Arrangements", Journal of Gerontology, Social Sciences 1991, vol 46, no. 3, p. S133-142.

Spector, W., Reschovsky, J. & Cohen, J. (1996). Appropriate placement of nursing homes residents in lower levels of care. Milbank Memorial Fund, in press.

US Department of Health and Human Services, Office of the Inspector General (US DHHS). (1982) Board and Care Homes: A Study of Federal and State Actions to Safegxuard the Health and Safety of Board and Care Home Residents. Washington, DC: Office of the Inspector General, DHHS.

US General Accounting Office (US GAO). (1989) Board & Care: Limited Assurance That Residents Needs Are Being Met. Washington, DC: General Accounting Office (GAO/HRD 89-50)

U.S. House of Representatives, Select Committee on Aging, "Board and Care Homes in America: A Nation of Tragedy" a Report by the Chairman of the Subcommittee on Health and Long-Term Care. House of Representatives 1989, pub. no. 101-711