In 1994 and 1995, the National Health Interview Survey included a Disability Supplement (NHIS-D) to collect extensive information about disabilities among individuals sampled as part of annual census-based household interview surveys. Here we describe the development and application of operational definitions of mental retardation and developmental disabilities to items in the NHIS-D to estimate prevalence. In our analyses, we estimate the prevalence of mental retardation in the noninstitutionalized population of the United States to be 7.8 people per thousand (.78%); of developmental disabilities, 11.3 people per thousand (1.13%); and the combined prevalence of mental retardation and/or developmental disabilities to be 14.9 per thousand (1.49%). Differences in prevalence estimates for mental retardation and developmental disabilities and among people of various ages are explored.
In this study we used the National Health Interview Survey's Disability Supplement (NHIS-D) to estimate the prevalence of mental retardation and/or developmental disabilities (MR/DD) among the noninstitutionalized population of the United States. Prevalence refers to the proportion of persons in a population who have a particular condition, illness, or status. Mental retardation is defined by the American Association on Mental Retardation (AAMR) as the co-occurrence of “significantly subaverage intellectual functioning . . . with related limitations in two or more adaptive skill areas: communication, self-care, home living, social skills, community use, self-direction, health and safety, functional academics, leisure and work” with such limitations manifested “before age 18” (Luckasson et al., 1992, p. 1). A similar definition is posed by the American Psychological Association (APA) and comparable diagnostic practices are outlined by both AAMR and APA for the classification of people as having mental retardation (American Psychological Association, 1996; Luckasson et al., 1992). Developmental disability is defined by the Developmental Disabilities Act as a “severe, chronic disability” attributable to “mental” and/or “physical” impairments that are “likely to continue indefinitely”; resulting in substantial functional limitations in three or more “major life activity areas”: self-care, receptive or expressive language, learning, self-direction, capacity for independent living and economic self-sufficiency; manifested by age 22 and requiring care, treatment or other services of lifelong or extended duration” (P.L. 98–527). Unlike mental retardation, classification of developmental disabilities does not benefit from a set of standardized assessment instruments, procedures, and professional training programs to operationalize its basic concepts.
There are many reasons to want accurate estimates of the prevalence of mental retardation and developmental disabilities. Because definitions of these disorders presume the need for assistance, prevalence estimates help approximate the number of people who may be expected to need assistance. When such definitions are applied to eligibility for specific programs, they help in estimating the number of persons who might claim benefits from such programs, and, when combined with service use statistics, they help estimate the number of eligible persons who are not being served. Prevalence estimates for subgroups (e.g., ethnic, age-related, or socioeconomic groups) assist in identifying the status and needs of such groups. Prevalence estimates provide an important statistical context to efforts within a society to plan and provide for groups of interest. The need for prevalence estimates for policy purposes was addressed by Senator Durenberger at a 1990 Senate hearing:
At the national level, there is somewhere between 1.7 million and 4.6 million [people with developmental disabilities]. There is no current method to determine the exact figures. This is crazy, that we have this kind of wide variance in our estimates. . . . I hope. . . we are going to look for ways to measure who those people are and where they are, so that we can best suit policy to their needs (S. HRG. 101–847, p. 3).
Although defined and recognized in federal laws and regulations, the term developmental disabilities has not been operationally and procedurally defined by the professional organizations that establish and certify diagnostic and classification technologies. As a result, prevalence estimates of developmental disabilities have depended on relatively unrefined procedures.
Aggregating “component conditions.”
Lacking formal diagnostic criteria, professionals have often estimated the prevalence of developmental disabilities by combining the prevalence estimates for conditions associated with developmental disabilities. Most often, these conditions include mental retardation, cerebral palsy, autism, spina bifida, and epilepsy. Such approaches have typically yielded estimates of 16 to 20 cases per 1,000 people; however, functional limitations criteria have not been applied in conjunction with the condition criteria. As a result, they include persons whose level of impairment would be less than the “substantial functional limitations standard” specified in federal definitions of developmental disabilities. These approaches have also faced the complexities of adjusting estimated prevalence to account for duplicate counting of people having more than one of these component conditions (Lubin, Jacobson, & Kiely, 1982).
Projecting school-age identification across the life span
School is a universal experience. It occurs during the years in which developmental disabilities are by definition manifested (i.e., birth to age 22). As a result, school assessments have been viewed as having substantial potential for the identification of the prevalence of developmental disabilities. Such estimates typically combine the number of students counted within certain types of the categories reported by states to the Department of Education (most often mental retardation, orthopedic disabilities, multiple disabilities, autism, deaf/blind, and traumatic brain injury). For example, child-count statistics reported by the states to the federal government for the 1997 school year (U.S. Department of Education, 1998) estimated that among children and youth 6 to 21 years old in the United States, 13.2 per 1,000 were in school-related categories suggesting developmental disabilities. Such estimates suffer from the limitations noted earlier, and the growing tendency to report “mental retardation” by the less pejorative label of “learning disabilities.” Evidence of this tendency can be seen in the child counts reported by the U.S. Department of Education (see Table 1). The proportion of students receiving special education services who were labeled as having mental retardation declined from 24% in 1977 to 16% in 1986 to 11% in 1995 (U.S. Department of Education, 1980, 1989, 1998). Estimates of the prevalence of developmental disabilities based on child count statistics is also limited because it may overestimate the prevalence of developmental disabilities among adults because people with developmental disabilities have lower than average survival rates as they age, although this may not be as much of a factor today as it was in the past.
Developing estimates from larger disability surveys
Numerous national and state population based household surveys have been designed to estimate populations of persons with disabilities. Notable among such surveys are major recurring national surveys, such as the Survey of Income and Program Participation conducted by the U.S. Bureau of the Census and the National Health Interview Survey (National Center on Health Statistics, 1994), for which tens of thousands of household members are interviewed annually. On the state level, New Jersey conducted a major disability survey in the early 1990s that included nearly 15,000 households. These surveys all included several measures associated with the types and levels of functional performance identified in the definition of developmental disabilities, but none has provided comprehensive coverage. This, in turn, limits the precision of estimates made by researchers. Often, the estimates provided vary depending on the operational definitions employed. For example, in an analysis of Survey of Income and Program Participation data, Thornton (1990) estimated a “narrow definition” prevalence of developmental disabilities of 0.8%, a “middle definition” prevalence of 1.1%, and a “broad definition” prevalence of 2.8%. In the New Jersey study, Hill, Harvey, and Berkovitz (1992) concluded that based solely on conditions associated with developmental disabilities, the prevalence would be 4.5%. When persons with associated conditions who did not report themselves to have any functional limitations are excluded, the prevalence decreased to 1.3%.
Although benefiting from standardized operational definitions and diagnostic procedures (American Psychological Association, 1996; Luckasson et al., 1992), estimates of the prevalence of mental retardation are also fraught with social and diagnostic complexities. For example, although approximately 2.3% of the population might be expected statistically to have substantially subaverage intelligence (by convention an IQ of at least two standard deviations [SDs] below the mean), only some of those individuals would also have concurrent substantial limitations in adaptive behavior. A second complexity is that “mental retardation” as defined is not necessarily a permanent condition. That is, people may be identified as meeting the criteria for the diagnosis in one phase of their lives but not in another (Luckasson et al., 1992). For example, children who are diagnosed as having mental retardation are often not labeled as such in adulthood. A third and related complexity is that identification and diagnosis of mental retardation is situationally influenced. Children of school age are particularly susceptible to the diagnosis because of the academic demands of schooling and the necessity of being labeled to receive needed remedial education services (Grossman, 1973). After leaving school for the day (President's Committee on Mental Retardation, 1969) or following graduation (Susser, 1968), many of these students no longer exhibit the “adaptive behavior” limitations that resulted from the demands of the school curriculum. A fourth complexity derives from the stigma associated with the diagnosis of mental retardation. This is manifested in at least two ways: (a) the label is decreasingly applied to individuals generally, with preference for less pejorative labels, most commonly “learning disabilities” (U.S. Department of Education, 1998); and (b) when applied, it is more likely to be applied to individuals of lower socioeconomic status and ethnic, racial, and cultural minority groups (Andrews, Pittman, Goldberg, Struening, & Wellen, 1995; McDermott & Altekruse, 1994; Mercer, 1973b; Murphy, Yeargin-Allsopp, Decoufle, & Drews, 1995). A final complexity, related to the fourth, is that given the stigmatizing mental retardation label, people have an aversion to acquiring it and tend to voluntarily submit to it only when notable advantages derive from it (e.g., remedial assistance in school, cash assistance through the Supplemental Security Income program, specialized long-term care, and other services). In the absence of perceived benefits, and often despite perceptions of certain benefits, many people resist being recognized as having mental retardation (Edgerton, 1967).
Strategies for Estimating Prevalence
Related to the preceding complexities are the methodologies for assessing the prevalence of mental retardation and the influences that these variations may have on observed prevalence rates. Methods used to determine the prevalence of mental retardation or developmental disabilities have included registries, birth cohort or prospective studies, administrative prevalence, case census or case finding, population-based screening, and household surveys.
Registries are a common method of estimating prevalence in Western countries other than the United States. Registries tend to be complete in their identification of persons with mental retardation through mandated reporting or other reporting mechanisms (Lilienfield & Stolley, 1994). A major limitation with registries is that they may overestimate the prevalence of mental retardation if persons are not removed from the register due to death, changes in diagnostic status (e.g., no longer exhibiting both significantly subaverage intelligence and limitations in adaptive behavior) or changes in geographical location.
Many researchers who have estimated the prevalence of mental retardation followed a birth cohort prospectively. The benefit of studies such as these is that they eliminate the likelihood of duplication found in registries. They are useful in identifying factors that may be associated with differing prevalence rates among different population groups (e.g., people living in poverty, low birthweight children). They are also useful in attending to changes in diagnosis and status over time. The main limitation to the cohort method is attrition (Lilienfield & Stolley, 1994). The number and characteristics of persons lost to the study scope can affect prevalence estimates.
A common method of estimating the prevalence of mental retardation is administrative prevalence. In this method, researchers survey service agencies (or schools within a particular geographic area) to obtain an unduplicated count of all persons receiving or qualifying for the agency's services. However, this method tends to identify only persons who are receiving services and overlooks persons who are unserved (Lakin, Hill, & Bruininks, 1985). It also fails to count those who may be receiving “generic” services, such as food stamps (Lakin et al., 1985). Case finding or case census methods are an extension of administrative prevalence. This approach seeks to find cases by expanding the circle of persons surveyed for possible cases. For example, information may be sought from physicians and others who provide generic supports, such as food or housing, to persons in the target group. One limitation to case-finding methods beyond those associated with administrative prevalence is the lack of response from those individuals being surveyed for information (Lilienfield & Stolley, 1994).
Researchers conducting population-based or cross-sectional prevalence surveys select a random sample of persons from a population in a specific geographical region (e.g., county, state, national). The sample is screened either through interviews or diagnostic examinations to identify those with the condition of interest (Lilienfield & Stolley, 1994). National household surveys are conducted by agencies such as the U.S. Bureau of the Census and the National Center for Health Statistics (e.g., The National Health and Nutrition Examination Survey, the National Health Interview Survey). Such approaches are notable because they are representative of the total population of the United States. As with any sampling method, however, there are certain limitations to this approach. Such approaches are affected by refusals of sampled individuals to participate, although for national population-based surveys, refusal rates tend to be remarkably low. Of greater concern is the potential for respondents to fail to reveal information about themselves or a family member that may be seen as highly sensitive or stigmatizing (e.g., having mental retardation) (Lakin et al., 1985).
Estimates of the Prevalence of Mental Retardation
The classic study estimating the prevalence of mental retardation in the United States was by Mercer in 1973. Mercer (1973a) reported a prevalence of 9.7 per 1,000 (.97%) using a variety of methods, including administrative prevalence, household surveys, case finding, and clinical evaluations, to confirm diagnosis. In more recent U.S. studies, the range of prevalence estimates for mental retardation and related conditions varied depending on the study location, the demographic characteristics of the sample, and the methods and definitions of disability used. For example, Kirby, Brewster, Canino, and Pavin (1995) reported a prevalence rate of 3.2 people with mental retardation per 1,000 children less than 14 years of age (.32%) looking retrospectively at a birth cohort in Arkansas, whereas Murphy et al. (1995) reported a prevalence rate of 12.0 per 1,000 in a population based cross-sectional study of 10-year-old children in Atlanta. Such differences reflect the important effects of methodology on the estimated prevalence of mental retardation. Murphy and colleagues (1995) used an administrative prevalence approach (i.e., identified individuals who were receiving services) and included children with all levels of mental retardation. They identified a rate of mild mental retardation of 8.4 per 1,000. This was larger than the prevalence rate for all persons with mental retardation found by Kirby and colleagues (1995), who used case finding to discover the prevalence of mental retardation within a birth cohort. Although they used many of the same sources as did Murphy and colleagues (e.g., hospital, physician, or school records), their prevalence estimates were lower because their wider age range included young children who may not be diagnosed until they begin school. Reschly and Jepson (1976) conducted a study based on a stratified random sample of children tested using the Wechsler Intelligence Scale for Children-Revised (WISC-R). Their results were much closer to the Kirby study (3.5 per 1,000).
Other investigators using different methods and age ranges have reported prevalence rates that were as much as two to three times as large as the studies by Kirby and colleagues (1995) and by Reschly and Jepson (1976), but still lower than the estimates by Murphy and colleagues (1995). Katusic and colleagues (1996) found an overall prevalence rate for mental retardation of 9.1 per 1,000 in a birth cohort study of children less than 9 years old with mild mental retardation, accounting for 4.3 cases of mental retardation per 1,000. Boyle, Decoufle, and Yeargin-Allsop (1994), using a registry system, found an overall prevalence rate for mental retardation of 8.7 per 1,000 with persons who have mild mental retardation accounting for 5.7 cases per 1,000.
Researchers who have included people of all ages have reported prevalence estimates that tend to be lower than those including only children. This reflects the tendency for children identified with mild mental retardation while in school to disappear into society in their adult years. The prevalence estimates for mental retardation in the United States for all ages ranged from .43 per 1,000 to 7.6 per 1,000. The Centers for Disease Control and Prevention (1996) estimated a 7.6 per 1,000 prevalence rate for mental retardation for all ages using an administrative prevalence approach with records from the Department of Education and Social Security Administration. Jacobson and Janicki (1983) found a prevalence rate of 7.2 per 1,000 in New York through the creation of an integrated administrative registry, including persons participating in a wide range of preschool, school, and adult programs; residential and home-based services; and generic health, social, and leisure/recreation services. Jones (1979) found a prevalence rate of .43 per 1,000 using the U.S. Bureau of Census 1976 Survey of Income and Program Participation of noninstitutionalized persons, in which people were asked to self-identify themselves or a family member as having mental retardation as a reason for limitations in age-appropriate activities. LaPlante's (1990) integration of the 1983–1986 sample of the National Health Interview Survey (NHIS) of noninstitutionalized persons yielded an estimated prevalence of 4 per 1,000 for all ages, using essentially the same procedure as the Survey of Income and Program Participation. With both of these surveys, an additional approximately 1.4 persons per 1,000 with mental retardation would be added if “institutionalized” populations were included (Prouty & Lakin, 1999). These studies illustrate the differences that can result when different methodologies are used to develop prevalence estimates for mental retardation.
In at least two studies investigators looked at the prevalence of a broader group of disabilities. Boyle et al. (1994) estimated the prevalence of developmental disabilities in people less than 18 years of age to be 1.7 per 1,000, based on the 1988 National Health Interview Survey. In another recent study, Newachek and Haflon (1998) found a prevalence of 16.96 per thousand using the 1992 to 1994 National Health Interview Core Surveys to estimate the prevalence of impairment of speech, senses, or intelligence as the main cause of disability.
In summary, there are many systematic approaches to studying the prevalence of mental retardation and developmental disabilities. Each method has notable limitations, and together they show the complexities of establishing a meaningful range of estimates for the prevalence of developmental disabilities and mental retardation. Among the many goals for the 2-year Disability Supplement to the National Health Interview Survey was to be able to address these complexities more comprehensively. In this paper we describe one approach to using the NHIS-D to do just that.
The National Health Interview Survey
The National Health Interview Survey (NHIS) is a household survey conducted by the U.S. Bureau of the Census annually since 1957. The NHIS is focused on the civilian, noninstitutionalized population in the United States. Each year this survey is used to randomly sample approximately 46,000 households with 116,000 members from 201 primary sampling units nationally. In 1994 and 1995, a special 2-year Disability Supplement (NHIS-D) was added to the NHIS to enable the gathering of nationally representative data on the characteristics, service use, needs, circumstances, and experiences of noninstitutionalized people with disabilities in the United States. The NHIS-D was administered in two stages. In an initial visit to each sampled household, the core NHIS interview and the Phase I Disability Supplement was completed for all members in the selected households. During this initial interview, questions were answered by any available adult in the household who was knowledgeable about the health of other household members. In the second phase, interviewers returned to households that included members with disabilities. This phase, called the Disability Supplement–Phase 2, was conducted several months after the Core and Phase 1 interviews. Interviews were usually face-to-face but sometimes were conducted over the telephone. When possible, the individual with the disability completed the Phase 2 interviews; but proxy interviewers were used for approximately 20% of adults and most of the children. The 1994/1995 NHIS and Disability Supplements covered a wide range of topics clustered into categories as shown.
NHIS Core Survey Topics
Housing arrangements and household composition
Demographics (race, ethnicity, family income, education)
Health and medical information (doctor visits, hospital visits, health indicators, conditions)
Abilities and limitations in activities of daily living (ADLs): getting around inside the home, bathing, dressing, eating, and toileting)
Abilities and limitations in instrumental activities of daily living (IADLs): going outside the home, keeping track of money or bills, preparing meals, doing light housework, and using the telephone; limitations or specific conditions among household members; service needs and access; and a wide range of related information
Disability Supplement Phase I Topics
Disability (sensory, communication and mobility limitations; specific conditions such as Down syndrome and mental retardation; activities of daily living and instrumental activities of daily living; functional limitations; mental health; services and benefits; special health needs of children; early child development; education; relationship to respondent; perceived disability; health care for specific conditions)
Family resources (access to care, health care coverage, private plan and coverage, income and assets)
Year 2000 objectives (environmental health, tobacco, occupational health and safety, heart disease and stroke, clinical preventative services, family, firearm safety)
Acquired Immunity Deficiency Syndrome (AIDS) knowledge and attitudes
Disability Follow-Back Survey (NHIS Phase II) Topics
Work, vocational rehabilitation, or school experiences
Services used (home care services, transportation, work, child care, medical services, assistive devices, educational services, other services, coordination of services)
Assistance with key activities
Participation in social activities
Mental health services and needs
Physical activity limitations
Personal adjustment role skills
Family structure and relationships
Impact on the family
Housing and long-term care services
Defining Mental Retardation and Developmental Disabilities
The NHIS-D permits examination of the implications of various definitions of mental retardation and/or developmental disabilities on prevalence estimates, service eligibility, and other topics requiring large and flexible data sets. Although the terms mental retardation and developmental disabilities are often used interchangeably, the terms have different operational definitions and, therefore, different meanings in law and regulation as well as in practice. Mental retardation is determined through established professional protocols for determining that an individual has a significantly subaverage IQ and concomitant limitations in social and behavioral skills. Developmental disability is defined by the presence of functional limitations in three or more areas of major life activity, but the definition lacks the standardized practices and instruments associated with determination of mental retardation.
To examine the nature and congruence of prevalence estimates for mental retardation and for developmental disabilities as defined in the Developmental Disabilities Assistance and Bill of Rights Act Amendments, we developed a multistage process for operationalizing both mental retardation and developmental disabilities using data elements contained in the 1994/1995 NHIS-D. This process included reviewing the definitions of these constructs used in planning and developing the NHIS-D and its instruments; preparing and revising initial operational definitions using the NHIS-D data elements based on reviews solicited from a panel of national experts in mental retardation and developmental disabilities diagnosis and research; and reviews and revisions of specific codes based on internal validation analyses and internal and external reviews of results. The final definitions are described in the following paragraphs.
Identifying People With Mental Retardation
People with mental retardation were identified in one or more of the following ways.
People were identified as having mental retardation if the household respondent answered affirmatively to a question asking whether anyone in the household had mental retardation.
People were also identified as having mental retardation when mental retardation was indicated as the cause of age-specific general activity limitations, which included limitations in (a) play for children ages 5 and younger, (b) school activities for children ages 5 to 17, (c) work for adults ages 18 to 69, and (d) overall limitations in activities for persons of all ages.
People were identified as having mental retardation if mental retardation was identified as the primary cause of limitations in communication, getting along with others, activities of daily living, instrumental activities of daily living, and other functional limitations; or if mental retardation was the International Classification of Diseases (ICD) code listed as the reason the person had a doctor's visit, a physician consultation regarding communication, or as the reason for receiving occupational therapy.
People were also identified as having mental retardation through a two-step process involving individuals who reported having a condition frequently associated with mental retardation. The first step identified people if the household respondent answered affirmatively to a question asking whether anyone in the household had autism, cerebral palsy, Down syndrome, spina bifida, or hydrocephalus, or reported having one or more of the following conditions as the cause of either age-specific general activity limitations in the Core Survey, as the cause of specific activity limitations (e.g., communicating, getting along with others) or as the reason for receiving various services (e.g., occupational or physical therapy) in the Phase 1 Disability Supplement: acromegaly, amino acid transport disease, autism, branched chain amino acid disturbance, cerebral palsy, congenital anomaly, congenital birth defect, congenital cytomegalovirus (CMV), congenital hypothyroidism, congenital syphilis, copper metabolic disorder, deformity of the skull, encephalopathy, epilepsy, hydrocephalus, Klinefelter syndrome, leukodystrophy, lipodoses, noxious substance affecting newborn, other brain or CNS conditions or deformities, other developmental delays, phenylketonuria (PKU), reduction deformity of the brain, sex chromosome anomaly, spina bifida, thalassemias, tuberous sclerosis, or unknown congenital or birth injury. Individuals meeting these criteria were identified as having “related conditions.”
In the second step, individuals with a related condition were screened to identify people who also reported having both a learning disability (as reported by the household respondent when asked whether anyone in the household had a “learning disability”) and a significant functional limitation in learning. A significant functional learning limitation was ascribed to children ages 1 to 17 years who had “problems or delays in understanding things, that is, delays in cognitive or mental development” with the problem having been “mentioned by a physician or other health care professional.” A significant functional learning limitation was also ascribed to individuals ages 5 or older who had “serious difficulty learning how to do things that most people their age are able to learn.” Individuals who had one or more related conditions who also reported having a learning disability and a significant functional learning limitation were considered to have mental retardation for this analysis. However, children ages 5 to 17 with related conditions were not included in the mental retardation group based on their related condition if their learning difficulty was not serious enough to require special education or to limit or prevent school attendance. Adults with related conditions were not included in the mental retardation group based on their related condition if they completed more than one year of postsecondary education. Furthermore, adults with related conditions were not included in the mental retardation group based on their related condition if they had Alzheimer's disease because that condition may have been the cause of their learning limitations. Although these criteria may have excluded a few individuals who would have been clinically assessed to have mental retardation, they were used to reduce the likelihood that we would mistakenly include large numbers of individuals who would not have been so assessed.
Overall, 871 people were identified as having mental retardation from the 1994 NHIS-D sample, and 740 people were identified as having mental retardation in the 1995 NHIS-D sample. Of these people, 116 in 1994 and 106 in 1995 were identified as having mental retardation based solely on having a related condition along with a learning disability and a significant functional learning limitation.
To understand more about the characteristics of people we identified as having mental retardation based on their “related conditions,” we analyzed each related condition to determine (a) the total number of people identified in the NHIS-D as having the condition, (b) the number of people with the condition who met one or more of our definitions for mental retardation, (c) and the number of people with each condition we identified as having mental retardation based solely on their related condition and learning limitations. We did not include Down syndrome in this table because the ICD codes for Down syndrome were not used in the NHIS public release files. Table 2 shows the results of this analysis. The most common related conditions in this sample were cerebral palsy, epilepsy, spina bifida, and autism. The proportion of people in this sample who met any criteria for mental retardation ranged from 11% for spina bifida, and 17% for cerebral palsy to 100% for sex chromosome anomaly, PKU, and reduction deformity of the brain. No individuals with thalassemias, encephalopathy, lipidoses, acromegaly, branched chain amino acid disturbance, congenital CMV, copper metabolic disorder, or Klinefelter syndrome met our criteria for mental retardation. Almost all of the people who met our definition for having mental retardation solely based on their related condition plus learning limitations had cerebral palsy, epilepsy, or autism. For those conditions, approximately one half of all people whom we classified as having mental retardation were so classified based on solely on their related condition plus learning limitations.
Definition of Developmental Disabilities
According to the Developmental Disabilities Assistance and Bill of Rights Act (DD Act) Amendments of 1996, a developmental disability is a severe, chronic disability of an individual 5 years of age or older that is attributable to a mental or physical impairment or combination of mental and physical impairments; is manifested before the individual attains age 22; is likely to continue indefinitely; results in substantial functional limitations in three or more of the following areas of major life activity: self-care, receptive and expressive language, learning, mobility, self-direction, capacity for independent living, and economic self-sufficiency; and reflects the individual's need for a combination and sequence of special, interdisciplinary, or generic services, supports, or other assistance that are of lifelong or extended duration and are individually planned and coordinated, except
that such a term, when applied to infants and young children means individuals from birth to age 5, inclusive, who have substantial developmental delay or specific congenital or acquired conditions with a high probability of resulting in developmental disabilities if services are not provided (P.L. 103–120).
A multistage process was used to identify individuals with developmental disabilities. The process involved using NHIS-D variables to construct operational definitions for each of the seven areas of functional limitation in the federal developmental disabilities definition: (a) self-care, (b) expressive or receptive language, (c) learning, (d) mobility, (e) self-direction, (f) capacity for independent living, and (g) economic self-sufficiency. The initial operational definitions were revised following review by a panel of experts in disability research. As conveyed in the DD Act, separate definitional approaches were required for persons 5 years and older and for children from birth to 5 years. The final operational definitions of functional limitations are summarized next. Actual NHIS-D item language is enclosed in quotation marks. In each of the seven “areas of major life activity,” to meet our criteria, limitations must have been first experienced before age 22 and must have been expected to endure for “at least 12 months.” This latter expectation was the best available proxy for the DD Act criterion of “is likely to continue indefinitely.” In general, the age limitations noted here reflect the ages of the people who were asked to respond to particular questions on the NHIS.
Final Operational Definitions of Functional Limitations Used to Define Developmental Disabilities
A person 5 years or older “has a lot of difficulty” or “is unable to” dress, eat, bathe, get in and out of bed or chairs, use the toilet, or get around the house.
Expressive or Receptive Language
A person 5 years or older “has serious difficulty communicating so the family can understand” or “has serious difficulty understanding others when they talk or ask questions.”
A person 18 years or older “has serious difficulty” or “cannot use” the telephone.
A person from 5 to 17 years old “has a problem or delay in speech or language development” mentioned by a doctor and these limitations were serious enough to require special education or to prevent or limit school attendance; or “has significant problems at school communicating with teachers and other students” or “cannot communicate with teachers and other students because of limitations” and these limitations were serious enough to require special education or to prevent or limit school attendance.
A person 5 years or older has a diagnosis of mental retardation, or autism, and “has serious difficulty learning how to do things most people their age can learn.”
A person 5 years or older has “serious difficulty learning how to do things most people their age can learn” and the person does not have Alzheimer's or another senility disorder.
A person from 5 to 17 years old “has problems or delays in understanding things, that is, delays in cognitive or mental development” that was serious enough to be mentioned by a doctor.
A person from 5 to 17 years old “has significant problems at school understanding instructional materials” or has been diagnosed as having learning disability and the person was receiving special education, had an IEP, or had a limitation in school attendance.
An adult age 18 or older has been diagnosed as having a learning disability that was significant enough that the person's overall educational background included less than 2 years of postsecondary education.
A person 5 years or older “has difficulty” or “is unable” to walk up stairs, walk three city blocks, or transfer to or from a bed or chair.
A person 5 to 17 years old “has difficulty participating in strenuous activity compared to other children his/her age” (and it is associated with arthritis, rheumatism, disorders of the bone or cartilage, absence or paralysis of an extremity or orthopedic impairment or deformity) or “has significant problems in physical development.”
A person 5 years or older, because of a physical, mental, or emotional problem, “needs to be reminded or have someone close by” for dressing, eating, bathing, toileting, or transferring.
A person 18 years or older has or needs “a case manager to coordinate personal care, social or medical services” or “has a court appointed guardian” during the last 12 months.
A person 5 to 17 years old has “significant problems with” or “cannot pay attention in class” or “control behavior in class” due to a limitation, and these problems were severe enough that the child was receiving special education, had an IEP, or was limited in school attendance.
Capacity for Independent Living
A person 18 years or older “requires help or supervision” or “has a lot of difficulty with” or “is unable” to prepare meals, shop for personal items or medicine, manage his or her money, do light work around the house like doing dishes, straightening up, light cleaning or taking out the trash, or do heavy work around the house.
A person 18 years or older “has never been able to work” or is “currently unable to work because of a mental or emotional problem” or “is limited in kind or amount of work” due to a limitation; or “has trouble finding or keeping a job or doing job tasks because of mental/emotional problems.”
A person 18 years or older has “participated in” or is “on the waiting list” for a sheltered workshop, transitional work training, supported employment or a day activity center; or is unable to work because of a mental or emotional problem.
Determination of Developmental Disability
People 5 years and older were considered to have a developmental disability if they had substantial limitations as defined earlier in three or more of the seven “areas of major life activity” expected to endure at least 12 months, with a limitation in at least one of these areas occurring before age 22. Based on the definition in the DD Act, children from birth to 71 months old were considered to have a developmental disability if they had a specific congenital or acquired condition that was consistently associated with developmental disabilities among older children and adults in the NHIS sample, or a substantial developmental delay in physical development, chewing, eating or toileting, receptive or expressive communication, understanding instructional materials, cognitive or mental development, participation in strenuous activity, or mobility. Children ages 5 and younger who had only a hearing, vision or speech impairment were not considered to have a developmental disability.
Overall, 1,207 people of all ages in 1994 and 1,118 people in 1995 met the criteria for having a developmental disability. Of those people, 464 in 1994 and 396 in 1995 also met the criteria for having mental retardation (see Table 3). A total of 1,614 people in 1994 and 1,462 people in 1995 met the criteria for either mental retardation, developmental disabilities, or both conditions. An additional 770 people in 1994 and 840 people in 1995 were identified as having cerebral palsy, autism, spina bifida, epilepsy, or one of the other related conditions but did not meet the criteria for mental retardation or developmental disabilities as defined for these analyses.
Prevalence of Mental Retardation and/or Developmental Disabilities
Once we finalized the definitions for mental retardation and developmental disabilities and identified sample members meeting those definitions, we used the population weights provided by the National Center for Health Statistics to estimate the prevalence of those conditions in the general noninstitutionalized population of the United States. We used the SUDAAN software crosstabs procedure to make population estimates of the prevalence of these conditions overall and in various age categories and to identify the standard error of estimate (Shah, Barnwell, & Bieler, 1997). Based on our definitions of mental retardation and developmental disabilities, we estimate that there are 14.9 people with one or the other or both of those conditions for every 1,000 people in the noninstitutionalized population of the United States (see Table 4).
Combining the 1994 and 1995 samples using the final population weights divided by two, this suggests that an estimated 3,887,158 noninstitutionalized Americans (plus or minus 1.9% or 75,440) have either mental retardation or developmental disabilities or both. The estimated prevalence varied by age. In children ages birth to 5 years, the prevalence of mental retardation and/or developmental disabilities is 38.4 per thousand. In children and youth ages 6 to 17 years, the prevalence of mental retardation and/or developmental disabilities is 31.7 per thousand. Finally, in adults ages 18 years and older, the prevalence is 7.9 people per thousand.
There is some overlap between people identified with developmental disabilities and those identified with mental retardation, but they are not the same groups (see Figure 1). Across all ages, 28.0% of people meeting the criteria for either mental retardation or developmental disabilities met both sets of criteria; 24.3% met the criteria for mental retardation but not for developmental disabilities; and 47.7% met the criteria for having developmental disabilities but not for mental retardation. Because the definition of developmental disabilities varied for persons of different ages, the distribution of population estimates was examined for three age groups (see Figure 2). Among noninstitutionalized children 5 years and younger, 88.3% met the criteria for having developmental disabilities but not for mental retardation, and 11.7% met both the criteria for having mental retardation and for developmental disabilities. (By definition, children 5 years and younger meeting the criteria for mental retardation also met the criteria for developmental disability.) Among noninstitutionalized children 6 to 17 years old, 35.9% met the criteria for developmental disabilities but not for mental retardation, 38.5% met the criteria for mental retardation but not for developmental disabilities, and 25.7% met both criteria. Finally, among adults, 33.6% met the criteria for developmental disabilities but not for mental retardation: 25.8%, for mental retardation but not for developmental disabilities, and 40.5% met both criteria.
The NHIS-D sample did not include persons in institutional or military settings. In 1999, state agencies reported 245,720 individuals with mental retardation and developmental disabilities living in nursing homes, psychiatric facilities, or congregate care (institutional) settings of 4 or more residents (Prouty & Lakin, 1999). Including these persons with the estimated 3,887,158 people with MR/DD in the noninstitutionalized U.S. population yields a combined estimated total number of people with mental retardation or developmental disabilities in the United States of 4,132,878 (a prevalence of 15.8 people per 1,000). This estimate does not include people with mental retardation or developmental disabilities who are in institutions that do not report their disability status to the state MR/DD agencies (e.g., jails, prisons, rehabilitation facilities, and certain nursing facilities) or in military settings.
Functional Limitations of Persons in Various Disability Groups
We identified four groups of persons with disabilities: (a) those who met the criteria for mental retardation who did not meet the criteria for having developmental disabilities; (b) those who met the criteria for developmental disabilities who did not meet the criteria for having mental retardation; (c) those who met the criteria for mental retardation and also met the criteria for developmental disabilities; and (d) a combined group that included people who met the criteria for either mental retardation or developmental disabilities or both. To further clarify the characteristics of people in these four groups, we analyzed the proportion of individuals in all of these groups who had each of the seven identified functional limitations. We included persons ages 6 to 17 years (for whom no measures of economic self-sufficiency or instrumental activities of daily living were included in the 1994/1995 NHIS-D) and adults. Children ages 5 and younger were not included because they were not assessed using the same definitions as were older children and adults. Table 5 shows the results of this analysis.
Among children of school age (i.e., 6 to 17 years), the group that met only the criteria for mental retardation all had functional limitations in learning. However, only one third had limitations in self-direction and only 1 in 10 had limitations in language. Although children in general are not independent in self-direction, for the purpose of this project, the self-direction criteria identified those children who had significant functional limitations with the kinds of self-directed activity that children are typically expected to perform (see Table 4). Almost none of the children in this group had limitations in mobility, and none had limitations in personal care. The group of children with developmental disabilities but not mental retardation was quite different. Almost all of those children had limitations in learning, self-direction, and language. About half had limitations in mobility, and only a few had limitations in personal care. Like children with only developmental disabilities, children who met the criteria for both mental retardation and developmental disabilities almost all had limitations in learning, self-direction, and language. About one third had limitations in mobility. Just over 1 in 10 had limitations in personal care. The major difference between the group with developmental disabilities only and the group with both mental retardation and developmental disabilities was that substantially more of the children who had both disorders had functional limitations in personal care. Among children who met one or more of these criteria, the most common functional limitation was learning, followed by self-direction, language, mobility, and, finally, personal care.
The pattern of functional limitations among adults was somewhat different than for children because seven areas were measured rather than five. Among adults who met the criteria for mental retardation but not for developmental disabilities, all had functional limitations in learning, and almost two thirds had functional limitations in economic self-sufficiency. Fewer than 5% had functional limitations in any of the other areas. Among adults who met the criteria for developmental disabilities but not for mental retardation, almost all had functional limitations in economic self-sufficiency. About two thirds had functional learning limitations and/or limitations in self-direction, and just over half had functional limitations in instrumental activities of daily living (e.g., shopping, cooking). About one quarter had functional limitations in language and/or mobility. Just over 1 in 10 had functional limitations in personal care. Adults who met the criteria for both mental retardation and developmental disabilities almost all had functional limitations in learning and economic self-sufficiency. In addition, two thirds had functional limitations in self-direction and/or instrumental activities of daily living, and half had functional limitations in communication. Fewer than 10% had functional limitations in personal care or mobility.
Among adults who met one or both of these criteria, the most common functional limitations were in economic self-sufficiency and learning, occurring for more than 87% of all identified adults. Limitations in self-direction or instrumental activities of daily living occurred for between 45% and 50% of identified adults. One third of all identified adults had a functional language limitation. Fewer than 15% of identified adults had functional limitations in personal care or mobility.
Prevalence of Selected Conditions
The overall prevalence of developmental disabilities in the noninstitutionalized population was estimated to be 11.28 per 1,000 people in the United States (see Table 6). The overall prevalence of mental retardation was 7.80 people per 1,000. Based on estimates from the 1987 National Medical Expenditures Survey Institutional Population Component (Lakin & Hill, 1990; Lakin, Hill, Chen, & Stephens, 1989), an estimated 94.8% of the people with MR/DD in nursing homes, psychiatric facilities, or congregate care settings of 4 or more residents would be expected to have mental retardation. This would yield a combined institutional and noninstitutional prevalence for mental retardation of 8.73 people per 1,000.
Although our primary focus in this analysis was to identify the combined prevalence of mental retardation and developmental disabilities, we also estimated the prevalence of three conditions commonly associated with MR/DD (cerebral palsy, spina bifida, and autism). For each of these three conditions, the NHIS gathered both ICD codes for conditions causing limitations and specifically asked household respondents in the Phase 1 supplement whether anyone in the household had the condition. Based on responses to these items, we estimate that the prevalence of cerebral palsy in the noninstitutionalized population of the United States is 4.02 people per 1,000, the prevalence of spina bifida is .62 people per 1,000, and the prevalence of autism is .40 people 1,000.
MR/DD in Persons With Select Related Conditions
Finally, we examined the proportion of people with the “related conditions” of cerebral palsy, spina bifida, and autism who also had either mental retardation and/or developmental disabilities (see Table 7). People with autism were most likely to have either mental retardation and/or developmental disabilities (87%). Only about 1 in 4 people with cerebral palsy (26%), or spina bifida (23%) also had mental retardation and/or developmental disabilities. Overall, 16% of people with autism, 9% of people with cerebral palsy, and 3% of people with spina bifida met the functional limitations definition of developmental disabilities without meeting the criteria for having mental retardation.
Mental retardation and developmental disabilities are not interchangeable classifications (see Figure 3). Only 28% of the NHIS sample members who met the criteria for either mental retardation or developmental disabilities classifications met the criteria for both. Concordance was lowest among children ages 5 and under (11.7% met both criteria) and highest among adults (40.5% met both criteria). Individuals who met the criteria for mental retardation, but not developmental disabilities, had functional limitations in one or two areas of major life activity, but not the three or more stipulated in the definition of developmental disabilities in the Developmental Disabilities Assistance and Bill of Rights Act Amendments. The differences in the patterns of functional limitations for individuals with mental retardation only, developmental disabilities only, and those who met the criteria for both as shown in Table 5 are striking and are worthy of further investigation. For children ages 5 to 17 years, the NHIS-D did not include variables to assess functional limitations in economic self-sufficiency or in instrumental activities of daily living. Therefore, in that age group, individuals had to have functional limitations in three of five (instead of three of seven areas). That factor probably accounts for much of the variability between children and adults in the concordance between mental retardation and developmental disabilities. The concurrent assessment of mental retardation using the AAMR or APA definitions and relevant clinical practice standards, and developmental disabilities using the DD Act definitions may prove very helpful in further elucidating salient differences between the constructs of mental retardation and developmental disabilities.
In a practical sense, distinctions between mental retardation and developmental disabilities can have important implications in people's lives when they are used as the basis for service eligibility decisions. Where eligibility for services depends on a systematic application of the definitional criteria for mental retardation, mental retardation and related conditions (typically defined as including cerebral palsy, autism, spina bifida, and a few other conditions) or developmental disabilities (as defined by the DD Act), different groups of individuals will be deemed eligible. The overlap between mental retardation and developmental disabilities is much less pronounced among persons who are not living in institutional settings than among persons living in institutions. Most people with MR and/or DD living in institutional settings (with 4 or more residents) have both a functional learning limitation (e.g., a diagnosis of mental retardation; 94% in the NMES) and perceived functional limitations in capacity for independent living as well as functional limitations in other areas (e.g., economic self-sufficiency). Therefore, the proportion of people in institutions meeting both mental retardation and developmental disabilities definitions will be substantially greater than the 40% of adults with either mental retardation or developmental disabilities in noninstitutional settings.
A second important finding in this study is that children were much more likely to have been identified as having mental retardation, developmental disabilities, or both (3.84% of children ages 5 and younger and 3.17% of children ages 6 to 17 years met the criteria for mental retardation, developmental disabilities, or both compared with only .79% of adults). The differences between children and adults may be explained several ways. For example, the results may reflect true differences in prevalence. The finding of differences between children and adults in the estimated prevalence is consistent with assertions that mental retardation is not stable over time (as argued in Luckasson et al., 1992). It is also constant with assertions that once the learning demands of school settings end, some individuals would no longer be considered clinically to have mental retardation in adulthood (as argued by Grossman, 1973, President's Committee on Mental Retardation, 1969, and Susser, 1968). The findings also support assertions that persons who were identified as having mental retardation seek to shed that label in adulthood, whether they still technically qualify for it or not.
On the other hand, it could be theorized that adults who identified themselves as having mental retardation but had functional limitations only in learning and economic self-sufficiency may, if reassessed, no longer meet the clinical definition for mental retardation. Because identification of persons with mental retardation relied on reports of the person or a proxy respondent indicating that the individual had been identified as having mental retardation or that mental retardation was the primary cause of an assessed limitation rather than on direct clinical assessment, however, it was not possible to test such hypotheses using the NHIS-D data.
Another factor that may explain differences between adults and children is definitional. Because the DD Act definition for young children is much less stringent than for older children and adults, it is not surprising that proportionally more young children were identified than were older children or adults. In addition, the NHIS-D interviewers asked different questions about children than about adults. With the exception of personal care and mobility, children ages 5 to 17 were assessed using different (although functionally similar) items than adults.
A third factor that may account for some of the differences in prevalence estimates for children versus adults is that there is increased mortality with age. In the NHIS-D, for example, whereas .9% of individuals ages 25 to 34 were identified as having mental retardation and/or developmental disabilities, only .4% of individuals 65 years or older were so identified (despite having identical criteria for identification). Of course, this analysis is somewhat confounded by the fact that institutionalization is more likely with increasing age, and the NHIS-D only included individuals in noninstitutional settings. However, even if all 245,720 individuals in institutional settings were adults, increases in the prevalence estimates for adults would still be much lower than the estimates for children. Further investigation of the possible explanations for and implications of differences in prevalence estimates for children and adults may be needed to completely understand these findings.
This analysis has demonstrated the importance of a comprehensive approach to identifying sample members as meeting the criteria for mental retardation, developmental disabilities, and related conditions. The comprehensive approach used in these analyses identified many more sample members with mental retardation and/or developmental disabilities than would have been identified with single-item screens used to identify people with mental retardation or developmental disabilities in most national data sets (including earlier years of the NHIS). For example, of the sample members who met one or more of the four criteria indicating mental retardation, only 38.4% said yes to the following question on the Phase 1 follow-up survey “I am going to read a list of medical conditions. Tell me if anyone in the family has any of these conditions. . . Mental retardation.” In all, the people who met our criteria for mental retardation reported more than 100 different ICD codes for the primary cause of one or more functional limitations on the core survey. Similarly, 21.5% (341 cases in 1994, 346 cases in 1995) of the people we identified as having mental retardation or developmental disabilities did not report activity limitations in the activities included in the Core NHIS survey, with children being particularly likely to be missed (419 were 0 to 5 years old, 169 were 6 to 17 years, and 99 were 18 years or older). Finally, 28,266 sample members who had one or more activity limitations on the core survey did not meet the criteria for either mental retardation or developmental disabilities. Clearly, the nature and comprehensiveness of inclusion criteria for establishing the presence of various disabilities are of major significance in prevalence estimates derived from the NHIS.
Project funding was provided by the National Institute on Disability and Rehabilitation Research (NIDRR), U.S. Department of Education, through a Cooperative Agreement (No. H133A60051) with the Center on Emergent Disability, University of Illinois at Chicago; through a NIDRR Field-Initiated Grant (No. H133G80082) to the Research and Training Center on Community Living at the University of Minnesota; through support of the Residential Information System Project provided by the Administration on Developmental Disabilities (Grant No. 90DN0028/01), and through the Rehabilitation Research and Training Center on Aging With Mental Retardation funded by the National Institute on Disability and Rehabilitation Research of the U.S. Department of Education under Grant No. H133B980046. The opinions expressed are those of the authors and do not necessarily reflect the views of the Institute, Center, University, or their funding sources. Requests for reprints should be sent to Sheryl A. Larson, Research and Training Center on Community Living, University of Minnesota, 214 Pattee Hall, 150 Pillsbury Drive SE, Minneapolis, MN 55455. (email@example.com)