The sample size and broad coverage of mental retardation facilities and residents in the Institutional Population Component of NMES will yield much information of value regarding the residential living arrangements of persons with mental retardation and related conditions. At the same time the study's sample frame limitations, and thereby its sample limitations, demonstrate the importance of assuring that any sample survey of persons in residential settings begins with the most comprehensive "universe" of facilities feasible. It is axiomatic, but certainly not trite, to observe that a sample can be no better than the sample frame. There is no way in the latter stages of a sample survey to compensate for inadequate efforts to understand and identify the universe of facilities being studied. Based on work with the NMES sample frame (the Inventory of Long-Term Care Places) as well as with the sample data themselves, the following general observations seem important lessons to derive from this study to guide other future studies of a similar nature.
Sample frame construction must begin at the state level and permit tailoring to the idiosyncracies of each state.
States differ in the components and organization of their service systems. A specific survey of each state to understand the different out-of-home services offered to persons with mental retardation and related conditions and to identify the individuals/agencies that can describe the necessary methods and key contacts for identifying and surveying all the settings within the services system is an essential first step to sample frame development. In states with decentralized service systems, often the only alternative to working with multiple state agencies and/or regional or county agencies is to accept an incomplete identification of facilities.
Inclusion of some types of community-based residences in certain states will sometimes require the direct involvement of state agencies in the research.
States are often reluctant to and/or directly prohibited from providing listings of their smallest residential settings, particularly those of a foster care model. To include such residential options in a national survey may require specific recruitment of the agencies controlling access to such facilities. For example in the 1982 University of Minnesota census survey of residential facilities for persons with mental retardation, Now York State's Office of Mental Retardation/Developmental Disabilities directly distributed over 1,000 surveys to family care homes; Delaware distributed, collected and forwarded questionnaires from its 71 special foster care settings. The ILTCP did not utilize such a strategy, which probably contributed to the undercounting of small facilities. For example in New York the ILTCP included a total of 1,484 mental retardation facilities, which was 911 (or 38%) fewer than surveyed in the 1982 University of Minnesota study. In Delaware the 1986 ILTCP counted 22 facilities, which was barely a quarter of the 80 surveyed in 1982.
The inclusion of residential options must be as comprehensive as feasible.
In recent years there has been increasing attention to the thousands of facilities generically referred to as "board and care homes." Board and care is a generic term which generally is taken to mean out-of-home, community-based living arrangements which are not Medicaid certified and/or do not provide medical services, but which do provide care, protective oversight and often training to people living in them. These operate under a range of different labels in different states, (e.g., adult foster care, domiciliary care, group homes, semi-independent living arrangements). Such places are becoming more common as services become more community oriented, and as states try to respond to increasing demands for community services with limited funds. The movement away from the majority of residents being in one or two models of care, state institutions and/or ICFs-MR, has led to much greater complexity and variety in residential services. Study designs must attend to the importance of procedures that assure systematic and comprehensive inclusion of all forms of residential settings.
Data on residents' service utilization and need, experiences and functional and behavioral characteristics are at least as important as population estimates.
One of the strengths of NMES was its gathering of data on a large sample of persons in a wide range of residential settings (i.e., mental retardation, nursing and related care homes). However, it is important to assure not only that data collection include sufficient samples of individuals and settings, but also that the data collected respond directly to the contemporary issues in providing services. Despite NMES' being the largest ever study of individuals with mental retardation and related conditions in all types of residential settings, it did not directly respond to data needs in areas such as functional and dysfunctional behavior of residents, specific services provided and needed, daily living experiences and relationships, community participation and resource use, and other topics which are important to understanding the current status and changing patterns of residential and related services.
Careful attention must be given to the quality of instrumentation, especially in creating unambiguous, objective and quantifiable questions regarding important characteristics and experiences of persons with mental retardation and related conditions.
In a number of areas the NMES attended to important topics, but did so in general and sometimes ambiguous ways that were often much less useful than they might have been. In many instances the specificity in the instrumentation was consistent with other federal health related surveys, but was considerably less than is typically used and generally expected in studies of persons with mental retardation and related conditions. For example, in the important area of challenging behavior, which is very frequently a critical characteristic, NMES asked, "Does (subject) sometimes disturb you or anyone else by [a list of potentially disturbing behaviors]?" These questions were presented with no severity or frequency indicators, making the intensity of the problem behavior impossible to assess. Examples of question terminology used in NMES that might have been ambiguous to respondents when applied to subjects with mental retardation are whether respondents "socialize" with friends or attend "meetings." Whether the interactions of persons with severe cognitive impairments would be interpreted as "socializing" or whether the group situations in which they find themselves would be seen as "meetings" is left to the interpretation of the respondent in many areas throughout the study, specific behaviorally defined items would have greatly improved the clarity of questions asked and data obtained.
The operational definition of "residential facility" used in NMES and in previous national studies needs to be expanded.
In the NMES, as in previous census studies which served to identify facilities for it, a "facility for the mentally retarded" was defined as:
- A place or unit certified as an Intermediate Care Facility for the Mentally Retarded (ICF-MR) by Medicaid.
- A place or unit that is formally state-licensed, or contracted living quarters (a) with three or more beds for clients who reside there, (b) providing to mentally retarded persons either personal care (ADL or IADL) or protective oversight, i.e., 24 hour-a-day, seven-day-a-week supervision, (c) not a licensed hospital unless a hospital for the mentally retarded, and (d) not a family providing services exclusively to relatives. In the case of an MR unit within a hospital, only the MR unit of the hospital is eligible.
Increasingly the service providers for persons with developmental disabilities are attempting to provide the residential supports needed by individuals rather than exclusively focusing on the development of supervised congregate care settings. As this important shift is taking place increasing numbers of people with developmental disabilities can be expected to be living in places that provide less than full-time protective oversight within the living unit. Such an orientation is supported by professional attitudes and program philosophies, as well as by federal policies such as Medicaid Home and Community Based Services.
NMES shows the benefits of simultaneous gathering of identical data sets on persons in mental retardation facilities, as well as other nursing and related care facilities.
A true strength of the NMES Institutional Population Component is that it included persons in mental retardation facilities and in nursing and related care homes and gathered comparable data on persons with mental retardation and related conditions in both classes of facility. The importance of this was further supported by the very small number of people (1%) identified as having related conditions, but not mental retardation who were living in mental retardation facilities. Quite apparently a national survey of persons with developmental disabilities in residential settings must include attention to facilities outside the traditional mental retardation programs.