CURRENT ESTIMATES OF DISABILITY AND RELATED CONDITIONS

As discussed in Chapter 2, the omission of key groups from national population surveys has important implications for the development of basic estimates of the population with disabilities. Such estimates must be pieced together from various sources, and the figures vary depending on the choice of survey and definition.

Box 3-1 lists several chartbooks and other profiles of disability data in the United States. The U.S. Census Bureau and most other agencies supply public use data sets so that researchers and others can obtain data more recent than those available in such profiles, and a few agency resources also allow some online analysis of the data.¹

BOX 3-1

Selected Recent Chartbooks and Other Profiles of Statistical Data on Disability. Federal Resources Centers for Disease Control and Prevention

One challenge in using information from different surveys is that different surveys rely upon different conceptual notions of disability, which in turn lead to different population estimates. For example, Stein and Silver (2002) found estimates of the rates of disability among children to be in the range of 14 to 17 percent, depending on whether children were identified through chronic conditions, special health care needs, or reports of disability. Moreover, even seemingly minor differences in the phrasing of questions or response options or in the ways of summarizing the data may yield different or even inconsistent pictures of a particular aspect of disability. For example, for older adults, some surveys ask whether they “have difficulty” with an activity, whereas others ask if they “need help” and others ask if they “get help or use special equipment” to perform the activity. Such differences can lead to different estimates of disability levels and trends. Hence, the use of these measures outside of a coherent conceptual framework and the lack of sufficient attention by users to the implications of differences in measures contribute to inconsistency and confusion.

Recognizing these caveats, the committee reviewed recent disability statistics and concluded that the total number of people in the United States with disabilities (defined to include individuals with impairments in body structure or function, activity limitations, or participation restrictions) currently exceeds 40 million. Depending on the survey from which statistics are drawn, the figure could exceed 50 million. No single data source yields estimates for all age groups living in the community and in institutional settings, so the estimate of the population with disabilities must be drawn from several different surveys.

The committee began with data from the U.S. Census Bureau’s 2004 American Community Survey. As summarized in Table 3-1, an estimated 38 million people (4.1 million people between the ages of 5 and 20, 20.2 million people between the ages of 21 and 64, and 13.5 million people ages 65 and older) who live in the community report a disability.² This estimate does not include people living in nursing homes and other institutional settings or children under age 5. The Medicare Current Beneficiary Survey estimates that approximately 2.2 million Medicare beneficiaries live in long-term care facilities and that about 350,000 of this population are adults under age 65 (CMS, 2005a). In addition, other surveys suggest that approximately 16,000 children with intellectual or developmental disabilities are living in out-of-home residential settings, about half which have four or more residents (Prouty et al., 2005). Data from the 2004 National Health Interview Survey indicate that perhaps 700,000 children under age 5 have a limitation in one or more activities because of a chronic condition, which brings the total of children ages 0 to 17 with disabilities to approximately 4.8 million. With these additional groups added to the American Community Survey estimate, the total estimate of people with disabilities exceeds 40 million.

TABLE 3-1

Disability Rates by Sex and Age (Excluding Ages 0 to 4), Civilian Population (Excluding Residents of Nursing Homes, Dormitories, and Other Group Housing), 2004.

Surveys that use a broader conception of disability than that adopted by this committee yield higher estimates. For example, based on the 2002 Survey of Income and Program Participation, which includes numerous questions to identify individuals of all ages with disabilities living in the community, the U.S. Census Bureau estimated that 51 million people have disabilities, including 32 million who have a severe disability (Steinmetz, 2006).³ The Behavioral Risk Factor Surveillance System, which includes two broad items to identify the adult community-based population with disabilities, also places the figure for all states close to 50 million (CDC, 2006a).⁴ Based on its review, the committee concluded that the number of people in the United States with disabilities exceeds 40 million and may exceed 50 million. Although these figures are not directly comparable to the 35 million estimate from the 1991 IOM report (because the survey questions differ), the number of people with disabilities has almost certainly increased since 1988, when most of the data used in that report were collected.

Table 3-2 lists the most common health conditions reported by respondents in the National Health Interview Survey as “causing” or contributing to limitations among people of different ages residing in the community. The survey questions reflect a largely medical model of disability. The committee could identify no national data on the extent to which features of the physical and social environments contribute to disability (see also Chapter 6).

TABLE 3-2

Leading Chronic Health Conditions Reported as Causing Limitations of Activities, by Age, Civilian, Noninstitutional Population, 2002 and 2003.

For children, the primary health conditions contributing to a limitation are cognitive, emotional, or developmental problems, although speech prob lems figure prominently for children under the age of 12 and asthma contributes to activity limitations among children in all age groups. For the 0.7 percent of children who had limitations so severe that they could not attend school, Msall and colleagues (2003), using data from the 1994–1995 disability supplement to the National Health Interview Survey, found that the most common reported reasons for nonattendance were life-threatening or other physical disorders, neurodevelopmental disorders, learning-behavior disorders, and asthma.

Among adults under age 65, musculoskeletal problems (including arthritis) and heart problems become increasingly important as people grow older. Mental illness is the second leading chronic condition mentioned as a cause of activity limitation for individuals ages 18 to 44 and is the fifth most frequently mentioned cause for individuals ages 55 to 64. For some people in this age group who are aging with disabilities, their primary health condition (e.g., cerebral palsy or spinal cord injury) is a risk factor for the development of secondary health conditions that have the potential to contribute to additional impairments, activity limitations, or participation restrictions. In general, survey questions have limited ability to distinguish such secondary disabling conditions from primary disabling conditions. (See Chapter 5 for further discussion.) For people ages 65 and over, musculoskeletal and heart problems continue to be leading contributors to limitations. Among people age 75 and over, senility (the term used by the National Center for Health Statistics but now more commonly referred to as dementia) is a major contributor to limitations.

MONITORING TRENDS IN DISABILITY

Monitoring trends in disability is important for several reasons. First, trend data provide a barometer of the nation’s achievements in terms of disability prevention. Second, when such data include measures of social, medical, and environmental risk factors, they can point policy makers to effective strategies for future interventions that will prevent or limit disabilities. Third, by including individuals of all ages, trend data can provide important insights into the future and serve as a basis for making assumptions that can be incorporated into projections.

Studies that track and attempt to explain changes over time in the population require a high degree of consistency in survey design. Changes in the wording of questions, the type and coverage of the sample frame, the use of proxy respondents, and the frequency and timing of interviews, among other factors, can all influence the conclusions drawn from such studies (Freedman et al., 2002). Unfortunately, despite the growing number of data sources and the repetition of certain major surveys, the available data—on the whole—permit few direct comparisons with data from the 1980s and earlier. Notably, although several surveys that focused on older adults have allowed analysis of late-life trends since the 1980s, discontinuities in the surveys make comparisons of present survey data with data from previous surveys difficult for those in early and middle life. This difficulty will be particularly evident in discussions of data from the National Health Interview Survey before and after the major revisions made to that survey in 1997 (see Chapter 2 for further discussion of the revisions). The 1991 IOM report used data from the 1988 National Health Interview Survey.

With these limitations noted, the next three sections of this chapter review trends in disability in early, middle, and late life. These divisions of the human life span are necessarily artificial to some degree. They are based on a mix of social conventions, statistical convenience, and public policy considerations; but they reflect important distinctions. Thus age 18, when people become legal adults in nearly all states, is used as the endpoint for early life (childhood and adolescence), even though important physical and psychological development continues past that age and well into the third decade of life (see the discussion of the transition into adult life in Chapter 4). Midlife, ages 18 to 64, encompasses a particularly broad period of life, almost five decades. As discussed in Chapter 5, many people in this group have conditions that in years past commonly led to early death, and many are experiencing premature or atypical aging that neither they nor their physicians have anticipated. The late-life period is defined as beginning at age 65, although some studies focus on an older group (age 70 and older or age 85 and older) that is at a considerably higher risk of disability than younger groups within the older population.

TRENDS IN EARLY LIFE

In a background paper prepared for an IOM workshop held in August 2005, Stein observed that “over a 40-year period, the proportion of children reported to have major limitations in their activities related to play and school has gone from less than 2 percent to close to 7 percent” (Stein, 2006, p. 146). The reasons for this trend are complex and varied. In part, the trend reflects changes in the epidemiology of childhood illness and functioning. For example, data show disturbing increases in recent years in the number of children who are reported to have potentially disabling chronic conditions, such as asthma and autism, as well as increases in the prevalence of preterm births (Stein, 2006). In addition, longer-term trends likely reflect increases in the recognition and treatment of learning-related disabilities and other conditions. The next section reviews these trends in more detail, and the subsequent section describes two public health successes: declines in the rates of spina bifida and lead poisoning.

Activity Limitations

Examining disability trends among children presents special challenges. Especially in the first few years of life, children’s developmental changes make it difficult or inappropriate to identify certain kinds of behaviors as impairments or activity limitations. Furthermore, “the functioning of children is always a moving target, as children mature at different rates, live in different cultures with different expectations of independence and self-sufficiency, and grow up in environments that vary markedly in the demands that they place on the performance of activities by children” (Stein, 2006, p. 145). Estimates of disability also vary depending on whether children are identified as having disabilities because of chronic conditions, special health care needs, or reports of activity limitations (Stein and Silver, 2002).⁵ In addition, as it is true for all age groups, revisions in national surveys complicate comparisons of disability trends for children. Despite these challenges and complexities, the data suggest increases in the proportion of children with activity limitations and conditions that put them at risk of disabilities.

The earliest data from the National Health Interview Survey (Newacheck et al., 1984, 1986) reported activity limitations for only 1.8 percent of children under age 18 in 1960. The rate group increased to 3.8 percent for 1979 to 1981. For the period from 1984 to 1996, reports of the rates of activity limitations for this group increased from 5.1 to 6.1 percent (Table 3-3).

TABLE 3-3

Percentage of Children (Under Age 18) with Activity Limitations, by Type of Limitation and Age, 1984 to 1996.

More recent data, based on new National Health Interview Survey questions that asked about the receipt of special education services, the need for assistance with personal care, and limitations in walking and cognition, suggest that increases in the numbers of children with activity limitations also occurred between 1997 and 2004 (Table 3-4). In this more recent period, the increase in activity limitations was particularly affected by the increasing receipt of special education services, especially for boys. Boys are almost twice as likely as girls to be reported to be receiving such services. In addition, pooled data from the 2000 to 2002 National Health Interview Survey show that boys have higher rates of mental retardation, learning disabilities (including attention deficit or hyperactivity disorder), asthma, and vision or hearing problems (Xiang et al., 2005). Although special education services may be aimed, in particular, at children with conditions that primarily affect cognition (e.g., mental retardation), they also serve children with medical conditions that may secondarily affect the ability to learn. (For the activity limitations and major activity limitations reported before 1997, boys also showed a higher prevalence of disability.)⁶

TABLE 3-4

Percentage of Children (Under Age 18) with Activity Limitations, by Type of Limitation, Age Group, and Gender, 1997 to 2004.

Potential Explanations of Trends

As suggested above, the tripling of activity limitations among children over four decades likely reflects a confluence of forces. In part, there have been real changes in the epidemiology of illnesses and related disabilities among children. In addition, the trend may be capturing in part the increasing awareness by parents, health professionals, and other agencies (especially schools) of conditions that merit attention and intervention, particularly through educational programs.⁷

Shifts in Conditions Potentially Contributing to Disability

Over the last few decades, the rise in the rates of potentially disabling childhood conditions—in particular, asthma—along with increases in the rates of preterm births, deserves special consideration in the analysis of activity limitation trends in children. During the early 1970s, when the rates of severe limitations grew from 2.7 to 3.7 percent, Newacheck and colleagues (1984, 1986) found increasing rates of several health conditions, especially mental health conditions, asthma, orthopedic conditions, and hearing loss. Unfortunately, changes in the questions as part of the redesign of the National Health Interview Survey in 1997 make comparisons over the entire time period inappropriate.⁸ Instead, the committee focused its review on the role of changes in the prevalence of a few potentially disabling health conditions.

Approximately 12 percent of children have at some time been diagnosed with asthma—the single most prevalent condition associated with childhood disability—and 8 percent currently have asthma (Federal Interagency Forum on Child and Family Statistics, 2005). In 2000, an analysis by the Centers for Disease Control and Prevention (CDC) reported that the prevalence of asthma among children increased by approximately 5 percent per year during the period from 1980 to 1995 but dropped by 17 percent in 1996 (CDC, 2000a). For the period from 1969–1970 to 1994–1995, Newacheck and Halfon (2000) documented the contribution of asthma to the rise in the overall prevalence of disabilities among children. They reported that the prevalence of disability related to asthma (based on data from the National Health Interview Survey) increased 232 percent over this time span, whereas the prevalence of disability related to all other chronic childhood conditions increased by 113 percent over the same period.

Changes in the questions about asthma in the National Health Interview Survey in 1997 preclude a comparison of recent data with earlier data. The prevalence of asthma did not, however, change much between 1997 and 2004. During this more recent time period, the percentage of children with asthma or breathing problems that caused limitations declined from 0.9 percent in 1997 to 0.6 percent in 2004 (unpublished tabulations of the National Health Interview Survey prepared by this committee).

Why the rate of asthma has increased and whether the rates have truly leveled off remain unclear. A recent review of evidence to explain the increase in the prevalence and severity of asthma in the United States and a number of other countries since the 1960s suggested that “the combination of the control of infectious diseases, prolonged indoor exposure, and a sedentary lifestyle … is the key to the asthma epidemic and, in particular, the key to the rise in severity” (Platts-Mills, 2005, p. 1026).

Another condition potentially contributing to disability is preterm birth. For a variety of reasons, including the increased rates of survival of high-risk infants and the growth in the number of multiple births associated with certain fertility treatments, the numbers of infants born prematurely and with low birth weights have increased (IOM, 2006a). For example, the rate of premature births (as a percentage of all births) increased from 10.6 percent in 1990 to 12.5 percent in 2004. Prematurity (birth before 37 weeks of completed gestation) and low birth weight (birth weight less than 2,500 grams) are risk factors for a number of short-term and long-term neurodevelopmental and other health problems and disabilities related to cerebral palsy, mental retardation, and sensory impairments, as well as more subtle disorders, such as attention deficit or hyperactivity disorder.

The reported prevalence of autism and related disorders has also grown significantly (Newschaffer et al., 2005;CDC, 2007). These conditions are characterized by impairments in social interactions and communication patterns and restricted, stereotyped, repetitive sets of activities and interests (WHO, 2006). CDC recently reported data from 14 sites showing an average rate of autism of 6.5 per 1,000 children aged 8 years in 2002 or approximately one child in 150 (CDC, 2007), a figure that is considerably higher than those suggested in earlier studies and reviews (see, e.g., Gillberg and Wing [1999] and Fombonne [2002]). The range reported in the CDC study was 3.3 to 10.6 per 1,000 children. An IOM study that examined the relationship between autism and childhood vaccines concluded that the published literature was generally uninformative in helping assess trends in the condition (IOM, 2004c). The growth in the reported rates of autism undoubtedly reflects a variety of factors, including the broader range of conditions currently encompassed in the spectrum of autism disorders (e.g., Asperger’s syndrome, other childhood disintegrative disorder, and pervasive developmental disorder, unspecified) and the addition of a substantial number of young people on the milder edge of the autism spectrum (APA, 2000; Shattuck, 2006). A substantial growth in the levels of awareness of autism and related disorders by both parents and clinicians has also led to the better identification of children who have autism spectrum disorders. In years past, these children might have been given another diagnosis (e.g., unspecified mental retardation) (Gurney et al., 2003; Barbaresi et al., 2005; Newschaffer, 2006). Some of the growth in autism rates likely reflects a true growth in the prevalence of autism.

Although the connection of obesity to disability warrants more investigation, another IOM committee has described the increasing prevalence of childhood obesity as a “startling setback” for child health (IOM, 2005a. 21). Obesity is a risk factor for a number of serious health conditions, such as diabetes, that are, in turn, risk factors for disabilities. Figure 3-1 shows trend data for children 6 to 11 and 12 to 19 years of age. Based in part on concerns about the stigmatization of children and in part on concerns about the reliability of the body mass index (BMI) as a measure of fatness for children, CDC has used the term “overweight” for children who would count as overweight or obese on the basis of BMI criteria (Wechsler, 2004). The earlier IOM committee, however, concluded that the term “obesity” was appropriate for children 2 years of age and older who have a BMI at or above the 95th percentile for their age and sex groups (IOM, 2005a).

FIGURE 3-1

Trends in overweight (obesity) for children ages 6 to 11 and adolescents ages 12 to 19, civilian noninstitional population, 1963 to 2002. Overweight (the term used by CDC) is defined as a BMI value at or above the 95th percentile cutoff points in the (more...)

Explanations for the increases in the rates of childhood obesity, which occurred primarily in the 1980s and 1990s, were cited in a 2001 report from the U.S. Surgeon General (U.S. Public Health Service, 2001). Factors include inactive lifestyles (e.g., watching television and playing computer and video games) and unhealthy eating patterns exacerbated by media exposure.

The Role of Increased Identification

Although the data of Newacheck and colleagues (1984, 1986) did not allow an estimate of how much of the growth in the activity limitations among children in the 1970s might reflect an increased awareness and identification of these limitations, their data raised the possibility that these factors might have played a role. For example, after noting that the marginal increase in activity limitations in the late 1970s (from 3.7 to 3.8 percent) was primarily accounted for by an increase in educationally related impairments, the authors proposed that the increase might have been associated with the implementation of the Education for All Handicapped Children Act of 1975 (which was renamed Individuals with Disabilities Education Act in 1990) and associated efforts to identify and provide mainstream education to children with disabilities (see additional discussion in Chapter 4).

In addition, other policies that have promoted screening of children for health and learning problems may have contributed to increases in the measured prevalence of disabilities. For example, the Social Security Amendments of 1967 created the Early and Periodic Screening, Diagnostic, and Treatment (EPSDT) benefit for children eligible for Medicaid. The objective was the early identification and management of problems that can impede children’s normal development. (Screening for the EPSDT benefit includes a comprehensive health and developmental history and a physical examination.) The implementation of the EPDST benefit has been far from universal, but millions of children in low-income families have been evaluated for disabling or potentially disabling conditions and were identified to have conditions that might otherwise have gone undiagnosed (Herz et al., 1998; GAO, 2001a).

Changes in eligibility criteria for the Supplemental Security Income (SSI) program in the early 1990s could also have influenced the rates of diagnosis or identification of functional limitations in children, especially among children with more severe disabilities (Ettner et al., 2000; Perrin et al., 1999). One change (resulting from the 1990 U.S. Supreme Court decision in Sullivan v. Zebley [493 U.S. 521]) required that disability de terminations for children include individual functional assessments, and the implementing regulations allowed the enrollment of children who had multiple health problems that, taken together (if not singly), were disabling.⁹ Also in 1990, the Social Security Administration expanded from 4 to 11 the number of qualifying mental impairments for children, one of which was attention deficit or hyperactivity disorder. Moreover, both the U.S. Congress (in 1989) and the U.S. Supreme Court provided for increased efforts to inform families of the SSI program and the eligibility criteria for the program, with an emphasis on active outreach to hard-to-find populations. Between 1989 and 1996, the number of children receiving SSI benefits grew from approximately 265,000 to 955,000 (or from 5.8 to 14.4 percent of all SSI beneficiaries) (SSA, 2006a). A study by Perrin and colleagues (1999) suggests, however, that the SSI changes had some impact but that the impact on the rates of reported potentially disabling conditions was fairly minimal. That study examined Medicaid claims for the period from 1989 to 1992 and found similar increases in the rates of asthma and mental health conditions among children insured by Medicaid, whether they received SSI benefits or not. A tightening of eligibility criteria in 1996 cut the number of children’s receiving SSI benefits to 847,000 by 2000, after which the number grew again to almost 960,000 at the end of 2003 and to more than 1,035,000 by the end of 2005 (Schmidt, 2004; SSA, 2006a).

TRENDS IN DISABILITY IN MIDDLE LIFE

Trends in impairments, activity limitations, and disability in middle life are important not only in themselves but also for what they may suggest about the future of disability in America when people now in the middle decades of life enter late life. Analyses of trends involving these adults have mostly focused on work limitations and employment issues, as discussed below, rather than on other kinds of activity limitations. Few systematic analyses have tracked trends in the health and environmental factors that contribute to disability in this age group. On balance, such studies that do exist suggest that the rates of disability are rising among America’s non-elderly adults, at least in part because of increases in the rates of obesity.

Personal Care or Routine Care Limitations

Figure 3-2 presents National Health Interview Survey data on personal care limitations (referred to as activities of daily living [ADLs]) and routine care limitations (referred to as instrumental activities of daily living [IADL]) over the period from 1984 to 2004 for individuals ages 18 to 44 and individuals ages 45 to 64. Although the numbers are generally low, Figure 3-2 shows that the prevalence of ADL limitations for both age groups rose through the early 1990s, followed by a slight decrease. The ADL trend line was essentially flat both before and after the 1997 survey redesign. Between 1997 and 2004, the percentage of those in the group ages 45 to 64 needing assistance with ADLs rose but only from 1.09 to 1.37 percent. For IADL limitations, Figure 3-2 shows considerable fluctuation, with a rise in prevalence during the early 1990s, followed by some decline. Since 1997, the rates have been flat.

FIGURE 3-2

Trends in the proportion of the civilian, noninstitutional population ages 18 to 44 and 45 to 64 needing help with ADLs or IADLs only, 1984 to 2004. ADL = needs help with personal care activities, such as eating, bathing, dressing, or getting around the (more...)

In a study that used National Health Interview Survey data for the population ages 18 to 69, Lakdawalla and colleagues analyzed trends in personal care (ADL) and routine care (IADL) needs among the working-age population between 1984 and 1996 and between 1997 and 2000 (Lakdawalla et al., 2005). During the period from 1984 to 1996, when the percentages of both kinds of disabilities increased among those ages 18 to 59 overall, the largest percent change occurred among those ages 30 to 39. In contrast, the prevalence of both kinds of disability decreased for people 60 to 69 years of age.

Limitations in Work

The last decade has seen considerable analysis of work limitations. It has also seen considerable controversy about the concept of work disability, the data used for analyses of disability and work, the analytic techniques used, and the interpretation of the data, particularly with respect to the effects of public policies on the employment of people with disabilities (Acemoglu and Angrist, 2001; Burkhauser et al., 2002; Autor and Duggan, 2003; Stapleton and Burkhauser, 2003; Burkhauser and Houtenville, 2004; Kaye, 2005). Several of these analyses are reviewed in Appendix E.

In contrast to the questions about personal care and routine care needs, the National Health Interview Survey does not ask people whether they need assistance to work. Rather, it asks whether they have a health problem or disability that (1) prevents them from working or (2) limits the amount or the kind of work that they can do. The Current Population Survey asks the same question, although the context is different, in that that survey does not focus primarily on health status. The American Community Survey asks whether a person has any difficulty working at a job or business because of a physical, mental, or emotional condition that has lasted for 6 months or more, although that question would be eliminated in proposed revisions to the survey (see the discussion in Chapter 2).

Figure 3-3 shows data for 1984 to 2003 from the National Health Interview Survey for people ages 18 to 44 and ages 45 to 64 who were reported to be unable to work. Among those ages 18 to 44, the percentage of people unable to work rose from the 1980s through the 1990s, before the survey redesign. The prevalence of people ages 45 to 64 reporting an inability to work (which was much higher overall than that for the younger age group) decreased in the 1980s but rose in the early 1990s. After 1997, the figures for both groups showed slight decreases, before increasing after 2000, especially for the older age group. The trends in reported work limitations (as opposed to an inability to work) were substantially similar, although the levels of work limitation are lower than the levels of inability to work among those ages 45 to 64.¹⁰

FIGURE 3-3

Trends in the proportion of the civilian, noninstitutional population who are unable to work, by gender, ages 18 to 44 and 45 to 64, 1984 to 2004. SOURCE: H. Stephen Kaye, Disability Statistics Center, University of California at San Francisco, unpublished (more...)

Employment Trends Among Adults with Disabilities

Most studies of disability and employment have found some reductions in the rates of employment among adults with disabilities over time (DeLeire, 2000; Acemoglu and Angrist, 2001; Bound and Waidmann, 2002; Houtenville and Burkhauser, 2004; Jolls and Prescott, 2004; Moon and Shin, 2006), but some analyses have reported increases in the rates of employment among individuals with severe functional limitations who report the ability and desire to work (Kruse and Schur, 2003). Consistent with the controversies over measurement, the interpretation of data on employment trends among adults with disabilities has generated considerable disagreement (see Appendix E). For example, some suggest that declines in employment reflect employer anxiety about the requirements of Title I of the Americans with Disabilities Act, which was passed in 1990 (DeLeire, 2000; Acemoglu and Angrist, 2001). Others attribute the declines to changes in the Social Security Disability Insurance (SSDI) and SSI programs during the 1980s that liberalized eligibility criteria (Houtenville and Burkhauser, 2004).

Figure 3-4 shows the trends in employment rates for people with and people without activity limitations. Overall, the graph suggests an increas ing gap between the level of employment for people without activity limitations and the level for people with activity limitations. The recessions of the early 1990s and 2000s appear to have affected all groups, but people with activity limitations have not experienced the more recent recovery in employment levels shown for people without such limitations. In addition, as employment rates among those with limitations have fallen faster than the rates among those without such limitations, the gap in employment between the two groups has grown.

FIGURE 3-4

Trends in employment rates for people with any type of activity limitation and for people without such limitations, ages 18 to 44 and 45 to 64, civilian, noninstitutional population, 1984 to 2004. Any type of activity limitation is indicated by a positive (more...)

Conditions Contributing to Disability in Middle Life

Few systematic analyses have investigated the health conditions that contribute to disability among nonelderly adults. In a recent review, Bhattacharya and colleagues (2006) report that adults under age 65 are increasingly likely to have chronic bronchitis (ALA, 2002), congestive heart failure (NHLBI, 1996), and diabetes (Mokdad et al., 2000). At the same time, other data suggest that the rates of heart disease (Reynolds et al., 1999), stroke (Carandang et al., 2006), and arthritis (CDC, 1997) have been declining in this age group. The National Health Interview Survey also allows a glimpse into trends since 1997 in the rates of selected conditions identified by respondents as causing activity limitations. Again, note that the survey questions reflect a largely medical model of disability, without regard to the contribution of environmental factors. For the most common conditions that were listed earlier in this chapter (Table 3-2) as contributors to activity limitations, Table 3-5 suggests very recent declines in fractures and joint injuries, lung disease, heart or other circulatory diseases, and arthritis as “causes” of or contributors to activity limitations. Diabetes, in contrast, increased in importance in 2003 and 2004. Further analysis is needed to sort out whether these changes are statistically significant, whether they are maintained after controlling for demographic shifts during the period, and whether all groups have been equally affected.

TABLE 3-5

Trends in Chronic Health Conditions Causing Limitations of Activity as Reported for Civilian, Noninstitutional Population, Ages 18 to 64, 1997 to 2004.

The Medicare Current Beneficiary Survey offers additional insights into the chronic conditions that exist among a subgroup of people between the ages of 18 and 65: those who have qualified for Medicare benefits not by age but by virtue of having qualified for SSDI. About 4 percent of Medicare beneficiaries are under age 45, and 11 percent are between the ages of 45 and 64. Perhaps the most striking trend revealed in Table 3-6 is the increase in the numbers of beneficiaries with a mental disorder, which is consistent with significant increases in SSDI awards because of mental disorders (Loftis and Salinsky, 2006).

TABLE 3-6

Trends in Self-Reported Health Conditions and Mobility Limitations, Community-Dwelling Medicare Beneficiaries Under Age 65, 1992 to 2002.

Finally, much has been written about the obesity epidemic in the United States (and elsewhere) and its implications. On the basis of data from the National Health and Nutrition Examination Surveys beginning in 1960, Figure 3-5 presents trend data on the rates of obesity among adults ages 20 to 39 and ages 40 to 59. The figure shows substantial increases in the rates of obesity among individuals in both age groups. In a paper prepared for the August 2005 IOM workshop on disability, Bhattacharya and colleagues (2006) linked these increases to increases in ADL and IADL limitations in middle life. Sturm and colleagues (2004) project increases in the prevalence of disability of 1 percent per year in the group ages 50 to 69, if current trends in obesity continue unabated. The link between obesity and disability needs further investigation because not only can obesity be a risk factor for disability but disability can likewise be a risk factor for obesity (see the discussion of secondary conditions in Chapter 5).

FIGURE 3-5

Trends in the prevalence of obesity for men and women ages 20 to 39 and 40 to 59, civilian, noninstitutional population, 1960 to 2004. SOURCE: Compiled from the work of Flegal et al. (2002) and Ogden et al. (2006) on the basis of data from the National (more...)

TRENDS IN DISABILITY IN LATE LIFE

Reflecting the high prevalence of disability in late life and concerns about the implications of an aging population on the economy, the health care system, and other areas of life, trends in disability have best been mapped and analyzed for older age groups. Generally, studies of disability trends in late life have focused on measures of activity limitations: ADLs and IADLs. Variations in the wording of questions across surveys, time periods, and other survey features have led to seemingly conflicting results in some cases. On balance, however, as first noted by Manton and colleagues (1993), trends in the prevalence of disability in late life are largely positive, with the evidence suggesting a substantial decline in the rates of IADL-related disabilities and smaller declines for ADL-related disabilities.

This discussion relies primarily on data from the 1982 to 2004 National Health Interview Surveys. For older individuals, comparison of pre-1997 survey data with data from the revised survey suggested only minor discontinuities in the data on personal care and routine care limitations.¹¹ Therefore, the data are presented without discontinuity by year.

Personal Care and Routine Care Limitations

Figure 3-6 shows the ADL and IADL trend data for 1982 through 2004. These data show a clear overall pattern of decline during the 1980s and 1990s in the percentage of the community-based population who need assistance with limitations in routine care activities, such as shopping or preparing meals, but who do not need help with personal care. The decline has continued in the current decade but at a slower pace.

FIGURE 3-6

Percentage of the civilian, noninstitutional population ages 70 years and older reporting need for help with personal care or help with routine care activities only, 1982 to 2003. ADL = needs help with personal care activities, such as eating, bathing, (more...)

Other data also point to declines in the percentage of the older people with limitations in routine care activities. For example, analyses based on data from the National Long-Term Care Survey suggested particularly large declines in three IADL measures from 1984 to 1999: doing laundry, managing money, and shopping for groceries (Spillman, 2004). A review of 16 relevant studies that used data from eight different surveys found substantial agreement that declines in IADL-related disabilities occurred through the 1980s and early 1990s (Freedman et al., 2002). A few more recent studies (Schoeni et al., 2005; Manton et al., 2006) suggest that such declines have continued into the current decade.

Changes in trends for limitations with personal care activities, such as dressing or bathing, are not as large. As presented in Figure 3-6, data from the National Health Interview Survey showed a small decline in the prevalence of personal care limitations during the entire period, from 8.2 percent in 1982 to 7.8 percent in 2003. The Medicare Current Beneficiary Survey also collects data on ADLs, although it asks about difficulty in performing an activity without help or special equipment. On the basis of data from that survey, Figure 3-7 shows decreases in limitations for community-dwelling beneficiaries in each category for the period from 1992 to 2003. A technical working group explored the inconsistencies in ADL trends across surveys and concluded that during the 1990s there were declines in difficulty with daily activities and the use of help with daily activities and increases in the percentage of individuals who used equipment to assist them with some activities (Freedman et al., 2004a).

FIGURE 3-7

Percentage of community-dwelling Medicare beneficiaries ages 65 and over who have difficulty in performing selected personal care activities without help or special equipment, 1992 to 2003. SOURCE: NCHS, 2007 (data from the Medicare Current Beneficiary (more...)

Despite the percent declines in IADL-related disabilities and the relative steady state of ADL-related disabilities, the total number of people with ADL or IADL limitations increased from approximately 4 million in 1982 to 4.2 million in 2003. (The numbers were calculated by using U.S. Census Bureau data on population by age group, available at http://www.census.gov/cgi-bin/ipc/idbagg.) Without the declines in IADL limitations, the 2003 number would be much larger.

The concern that not all groups may be benefiting equally from the declining rates of disability has led some researchers to investigate disparities in trends by race and socioeconomic status. Findings by race have been inconsistent thus far. On the one hand, data from the National Long-Term Care Survey suggest that racial disparities in chronic ADL- or IADL-related disabilities increased during the 1980s and decreased during the 1990s (Clark, 1997; Manton and Gu, 2001). On the other hand, three studies that used data from the National Health Interview Survey found no statistically significant changes in relative disparities in disabilities between racial minorities and whites, although the absolute size of the gaps may have narrowed (see, for example, the work of Schoeni et al. [2005]).

A review of analyses of disparities in the rates of disabilities by level of education again found inconsistent results (Freedman et al., 2002). One study based on the National Long-Term Care Survey reported no clear pattern for the period from 1982 to 1999 (Manton and Gu, 2001). Another study that was also based on data from the National Health Interview Survey reported a decline in the prevalence of disabilities only among those with 13 or more years of education (Schoeni et al., 2001). More recently, using data from the National Health Interview Survey for the period 1982 through 2002, Schoeni and colleagues (2005) demonstrated widening gaps in the prevalence of disabilities by socioeconomic status.

Disabilities Within Group Residential Care Settings

Because of data limitations, many trend analyses have omitted data for the 2 million older individuals living in nursing homes and other group residential care settings. Data from the 1977 to 2004 National Nursing Home Surveys indicate that the proportion of older individuals living in nursing homes may have declined in recent years (Alecxih, 2006). Other analyses suggest that the short-stay population (those with stays of less than 3 months) has increased sharply, primarily as a result of Medicare’s adoption of hospital prospective payment, which gave hospitals strong incentives to discharge people more quickly, whether it was to home or to a post-acute care facility (Decker, 2005).

Further insight into trends in disability among the population living in long-term care facilities is provided in Figure 3-8, which is based on data from the Medicare Current Beneficiary Survey. Figure 3-8 shows ADL-related data from 1992 to 2003 for beneficiaries age 65 and over living in long-term care facilities. (Note that the y-axis for Figure 3-6 extends from 0 to 16 percent, whereas it extends from 0 to 100 percent for Figure 3-8). Figure 3-8 shows increases in ADL disabilities for this group and, as ex pected, very high levels of limitations. Data from the National Nursing Home Survey cited above show that the percentage of nursing home residents ages 85 and over increased from 35 to 47 percent between 1977 and 1999 (Decker, 2005).

FIGURE 3-8

Percentage of Medicare beneficiaries ages 65 and over living in nursing homes and other facilities who have any difficulty in performing selected personal care activities because of a health or physical problem, 1992 to 2003. SOURCE: NCHS, 2007 (data (more...)

Other evidence also suggests that older adults living in nursing homes are frailer and have higher levels of disability now than older adults did a decade ago, perhaps because alternative care arrangements are more available for people with lesser but still significant levels of impairment (Spillman and Black, 2005b). For example, assisted living facilities, which offer supported living arrangements for those who need assistance but who do not require 24-hour skilled nursing care, grew from being almost nonexistent in the 1970s to serving about 750,000 older adults in 2000 (Spillman and Black, 2006).¹² As described in Chapter 9, state Medicaid programs are paying for an increasing proportion of long-term care in home rather than institutional settings.

Health Conditions Associated with Disability in Late Life

The extent to which some chronic conditions develop into disability may have been ameliorated in recent decades, in particular for arthritis (Freedman and Martin, 2000) and cardiovascular diseases (Cutler, 2003), even as the prevalence of many of the conditions in the older population has increased (Crimmins and Saito, 2000; Freedman and Martin, 2000; Crimmins, 2004). With respect to sensory functioning, the findings are mixed. Data from the Survey of Income and Program Participation suggest substantial declines in the percentage of older Americans with difficulty seeing from 1984 to 1993 (Freedman and Martin, 1998) and in the percentage with difficulty seeing or hearing from 1984 through 1999 (Cutler, 2001). Data from the Supplements on Aging to the National Health Interview Survey, however, show that the rates of blindness, deafness, and hearing impairment remained constant between 1984 and 1995 (Desai et al., 2001). Like the review for the nonelderly adult population, this summary of trends in health conditions and certain kinds of disabilities draws upon many data sources, data from different time periods, and different measures.¹³

The committee also reviewed trends in conditions that contribute to disability using data from the 1997 to 2004 National Health Interview Survey. The most common conditions linked to activity limitations among the population age 65 and older (Table 3-7) include arthritis and other musculoskeletal conditions, heart and other circulatory conditions, hearing and vision problems, diabetes, lung disease, and senility (dementia). The prevalence of these conditions as contributors to activity limitations declined for all conditions except diabetes and senility.

TABLE 3-7

Trends in Chronic Health Conditions Related to Activity Limitations, Civilian, Noninstitutional Population, Ages 65 and Over, 1997 to 2004.

PROJECTING THE FUTURE OF DISABILITY IN AMERICA

The number of people with impairments in body structures or functions is likely to grow substantially in the coming years. Unless substantial progress is made in reducing the chances of such impairments developing into activity limitations and participation restrictions, particularly at older ages, the number of people in the United States facing barriers to work, health care, and independent living will no doubt escalate. The number of individuals needing personal assistance, whether it is from family members or from paid caregivers, will also increase. Designing and implementing strategies and policies to promote the health and well-being of people with existing disabilities and to prevent the development or progression of potentially disabling conditions should, therefore, be national priorities.

In preparing this report, the committee reviewed the few existing studies that project the size of the population with disabilities. The review included, for example, projections related to the SSDI and SSI programs, the demand for long-term care services, and the rates of disability among the elderly population (see Bhattacharya et al. [2004] and Waidmann and Liu [2000]).¹⁴ Although a comprehensive review of different methodologies is beyond the scope of this report, a few key issues can be cited. In projections of the future size of the older population, assumptions about life expectancy (that is, how long people can be expected to live) are key, whereas projections of the size of younger populations rely more heavily on assumptions about future birth rates. Also, projection methodologies generally make some assumptions about whether disability rates will increase, decline, or remain unchanged in the future. Some projection models have built elaborate assumptions about such rates (allowing variation, for example, in the timing of disease onset and progression), whereas others simply project forward the existing rates onto the numbers of people in each age group. Methods also vary in whether changes in other factors linked to disability (e.g., years of completed education) are incorporated into projections or are held constant. Finally, projections of program enrollment (in, e.g., the SSDI program) rather than disability per se may involve assumptions about additional variables, such as labor force participation rates.

No source for projections of the number of children with disabilities could be identified. However, in the last half century and more, advances in biomedicine have saved many children who once would have died from conditions such as prematurity, traumatic injury, and a variety of genetic conditions or fetal developmental problems. Some of these children have gone onto experience slight to substantial physical or cognitive impairments. Other children have experienced impairments related to exposures to toxic substances. Furthermore, because disabling or potentially disabling conditions that are present at birth or that arise during childhood can have long-term consequences (Forrest and Riley, 2004; IOM, 2003a, 2005a, 2006a), the increases in the numbers of children with conditions such as childhood obesity and diabetes could eventually add to the numbers of adults and older adults with disabilities. The lack of projections based on these developments represents an important gap in present knowledge.

The number of working-age adults eligible for SSDI (meaning that they have a qualifying disability and have worked long enough in Social Security-covered jobs to qualify for benefits) is projected to increase. For example, the Congressional Budget Office projects that caseloads will increase from 6.7 million in 2000 to 10.4 million in 2015 (CBO, 2005b). The number of beneficiaries per covered worker is also projected to increase, as described in Chapter 1. However, the consequences of premature aging or atypical aging for people with disabilities—including individuals with conditions that once were incompatible with long-term survival—have been little discussed with respect to their potential impact on the SSDI program. More generally, analysts have predicted that ADL disabilities and work limitations among adults will increase (irrespective of program eligibility), in part as a consequence of the growing prevalence of obesity and related disorders (Sturm et al., 2004).

Focusing on older adults, if the rate of activity and other limitations for those ages 65 or over were to remain what it is today (roughly 40 percent from Table 3-1), the number of older people with impairments or limitations would increase from approximately 14 million today to more than 28 million in 2030. Using different definitions and a more complex methodology, Waidmann and Liu (2000) have projected that the numbers of older adults with activity limitations would grow from 22 million in 2005 to 38 million by 2030. They also projected that the number of people with limitations in ADLs or IADLs would increase from 12 million to 22 million over this period, if the rates of people with limitations in ADLs and IADLs remained constant but the numbers of people classified by age, sex, and education change according to U.S. Census Bureau projections. Importantly, Waidmann and Liu also project that if rates of activity limitations did not hold constant in the future but, instead, declined at the same rate observed during the 1990s, the number of people with such limitations would still grow to about 38 million by 2030. In contrast, the number of older people with limitations in ADLs or IADLs would stay steady at about 12 million. Another analysis, which assumes that disability rates remain constant, estimates that the number of Americans of all ages with limitations in ADLs only will increase from 5.2 million in 2000 to 9.3 million in 2030 (Friedland, 2004).

Taken together, these projections suggest that the number of people with disabilities is likely to increase in the coming years, unless steps are taken to reduce the environmental barriers that contribute to avoidable activity limitations and participation restrictions. The good news is that the rates of limitations for some activities have already been declining for older individuals. As discussed earlier in this chapter, it is not completely clear why this is the case. As discussed in later chapters of this report and as recommended in Chapter 10, further research is needed to identify and disseminate practices and programs that minimize activity limitations and participation restrictions.

Footnotes

1

As indicated in the source citations for most of the figures in this chapter, the committee contracted with H. Stephen Kaye of the University of California at San Francisco to supply information from the public use data sets for the National Health Interview Survey.

2

In the American Community Survey, “disability” is defined as a long-lasting sensory, physical, mental, or emotional condition that can make it difficult for a person to walk, climb stairs, dress, bathe, learn, remember, go outside the home alone, or work at a job or business. The term also covers vision or hearing impairments.

3

For adults, questions in the Survey of Income and Program Participation ask about mobility-related assistive technology use; activity limitations; learning disabilities; the presence of a mental or emotional condition, or both; mental retardation; developmental disabilities; Alzheimer’s disease; and conditions limiting employment or work around the house. For children, the questions involve specified conditions (autism, cerebral palsy, mental retardation, developmental disabilities); activity limitations (seeing, hearing, speaking, walking, running, taking part in sports); developmental delays; difficulty walking, running, or playing; or difficulty moving the arms or legs (Steinmetz, 2006). Not all the questions in the survey involve disability as defined in this report.

4

Widely cited estimates from the 2000 decennial census also put the estimate of the civilian, noninstitutional population with disabilities near 50 million. However, U.S. Census Bureau analysts attributed this estimate to a formatting problem with the census questionnaire that may have incorrectly increased positive responses to questions about disabilities with going-outside-the-home and work limitations (Stern and Brault, 2005). Subsequent estimates from the American Community Survey of the U.S. Census Bureau put the figure closer to 40 million.

5

As defined by the U.S. Maternal and Child Health Bureau, children with special health care needs are “those who have or are at increased risk for a chronic physical, developmental, behavioral, or emotional condition and who also require health and related services of a type or amount beyond that required by children generally” (McPherson et al., 1998, p. 137). This broad definition was developed to help in implementing amendments to the Social Security Act that provided for the development of community-based services for children with special health care needs and their families.

6

Other surveys besides the National Health Interview Survey suggest significant increases in the rates of certain potentially disabling chronic conditions among children and youth, especially in recent years. For example, the National Longitudinal Survey of Labor Market Experience, Youth Cohort, provides information on the health-related conditions of children born to women who were ages 14 to 21 in 1979. These data show an increase in the prevalence of chronic conditions from 11 percent in 1994 to 24 percent in 2000 among children who were ages 8 to14 in those years (unpublished tabulations from James Perrin, committee member).

7

Increased awareness may affect the estimation of disability rates for other age groups besides children; see, for example, the work of Waidmann and Liu (2000) for the influence of changes in Social Security Disability Insurance policies and criteria on self-reports of disability by adults.

8

In addition, because of further changes in questions in the National Health Interview Survey after 1997, of the six leading conditions (speech problems; asthma or breathing problems; mental retardation or other development problems; other mental, emotional, or behavioral problems; attention deficit or hyperactivity disorder; and learning disabilities), only the survey questions related to speech problems and asthma are comparable over the 1997 to 2004 time period. For example, before 2000, parents were asked only if they had been told that their child had attention deficit disorder; after that they were also asked about attention deficit or hyperactivity disorder (Pastor and Reuben, 2005).

9

The Zebley decision was based on a notion of equity. Until that decision, the Social Security Administration provided functional assessments for adults who might not otherwise qualify for SSI benefits on the basis of their clinical symptoms alone.

10

Burkhauser and colleagues (2002) point out that many people with impairments that would seem to be work limiting are not actually reporting such work limitations in the National Health Interview Survey. One explanation is that many people who report no work limitations make such a report on the basis of their experience of having found a job that they can do without ongoing difficulty either because of the nature of the job or because of the accommodations in work conditions that employers or workers have made, or both.

11

This discussion draws on a background paper prepared for the August 2005 IOM workshop organized during the first phase of this study (Freedman, 2006).

12

At a minimum, these facilities offer 24-hour supervision and assistance, as well as meals in a common dining area. Other services may include housekeeping and laundry services, medication reminders or help with medications, help with personal care activities, transportation, security, health monitoring, care management, and activities.

13

See IOM workshop paper prepared by Freedman (2006) for more details.

14

Projections of SSDI caseloads are available from CBO (2005b) and the work of Toder et al. (2002). Long-term care projections are available from CBO (1999). Social Security projections of SSI and SSDI are available from SSA (2006c).