U.S. Public Health Service Seal Contents Search Order Press Resources Links Home

Chapter 2
The Fundamentals of Mental Health
and Mental Illness

The Neuroscience of Mental Health

Overview of Mental Illness

Overview of Etiology

Overview of Development, Temperament, and Risk Factors

Overview of Prevention

Overview of Treatment

Overview of Mental Health Services

Overview of Cultural Diversity and Mental Health Services

Overview of Consumer and Family Movements

Overview of Recovery



Epidemiology of Mental Illness

Few families in the United States are untouched by mental illness. Determining just how many people have mental illness is one of the many purposes of the field of epidemiology. Epidemiology is the study of patterns of disease in the population. Among the key terms of this discipline, encountered throughout this report, are incidence, which refers to new cases of a condition which occur during a specified period of time, and prevalence, which refers to cases (i.e., new and existing) of a condition observed at a point in time or during a period of time. According to current epidemiological estimates, at least one in five people has a diagnosable mental disorder during the course of a year (i.e., 1-year prevalence).

Epidemiological estimates have shifted over time because of changes in the definitions and diagnosis of mental health and mental illness. In the early 1950s, the rates of mental illness estimated by epidemiologists were far higher than those of today. One study, for example, found 81.5 percent of the population of Manhattan, New York, to have had signs and symptoms of mental distress (Srole, 1962). This led the authors of the study to conclude that mental illness was widespread. However, other studies began to find lower rates when they used more restrictive definitions that reflected more contemporary views about mental illness. Instead of classifying anyone with signs and symptoms as being mentally ill, this more recent line of epidemiological research only identified people as mentally ill if they had a cluster of signs and symptoms that, when taken together, impaired people’s ability to function (Pasamanick, 1959; Weissman et al., 1978). By 1978, the President’s Commission on Mental Health (1978) concluded conservatively that the annual prevalence of specific mental disorders in the United States was about 15 percent. This figure comports with recent estimates of the extent of mental illness in the population. Even as this figure has become more sharply delineated, the older and larger estimates underscore the magnitude of mental distress in the population, which this report refers to as “mental health problems.”


The current prevalence estimate is that about 20 percent of the U.S. population are affected by mental disorders during a given year. This estimate comes from two epidemiologic surveys: the Epidemiologic Catchment Area (ECA) study of the early 1980s and the National Comorbidity Survey (NCS) of the early 1990s. Those surveys defined mental illness according to the prevailing editions of the Diagnostic and Statistical Manual of Mental Disorders (i.e., DSM-III and DSM-IIIR). The surveys estimate that during a 1-year period, 22 to 23 percent of the U.S. adult population—or 44 million people—have diagnosable mental disorders, according to reliable, established criteria. In general, 19 percent of the adult U.S. population have a mental disorder alone (in 1 year); 3 percent have both mental and addictive disorders; and 6 percent have addictive disorders alone.3 Consequently, about 28 to 30 percent of the population have either a mental or addictive disorder (Regier et al., 1993b; Kessler et al., 1994). Table 2-6 summarizes the results synthesized from these two large national surveys.

Individuals with co-occurring disorders (about 3 percent of the population in 1 year) are more likely to experience a chronic course and to utilize services than are those with either type of disorder alone. Clinicians, program developers, and policy makers need to be aware of these high rates of comorbidity—about 15 percent of those with a mental disorder in 1 year (Regier et al., 1993a; Kessler et al., 1996).

Based on data on functional impairment, it is estimated that 9 percent of all U.S. adults have the mental disorders listed in Table 2-6 and experience some significant functional impairment (National Advisory Mental Health Council [NAMHC], 1993). Most (7 percent of adults) have disorders that persist for at least 1 year (Regier et al., 1993b; Regier et al., in press). A subpopulation of 5.4 percent of adults is considered to have a “serious” mental illness (SMI) (Kessler et al., 1996). Serious mental illness is a term defined by Federal regulations that generally applies to mental disorders that interfere with some area of social functioning. About half of those with SMI (or 2.6 percent of all adults) were identified as being even more seriously affected, that is, by having “severe and persistent” mental illness (SPMI) (NAMHC, 1993; Kessler et al., 1996). This category includes schizophrenia, bipolar disorder, other severe forms of depression, panic disorder, and obsessive-compulsive disorder. These disorders and the problems faced by these special populations with SMI and SPMI are described further in subsequent chapters. Among those most severely disabled are the approximately 0.5 percent of the population who receive disability benefits for mental health-related reasons from the Social Security Administration (NAMHC, 1993).

Children and Adolescents

The annual prevalence of mental disorders in children and adolescents is not as well documented as that for adults. About 20 percent of children are estimated to have mental disorders with at least mild functional impairment (see Table 2-7). Federal regulations also define a sub-population of children and adolescents with more severe functional limitations, known as “serious emotional disturbance” (SED).4 Children and adolescents with SED number approximately 5 to 9 percent of children ages 9 to 17 (Friedman et al., 1996b).

Table 2-6. Best estimate 1-year prevalence rates based on ECA and NCS, ages 18—54

ECA Prevalence (%)
NCS Prevalence (%)
Best Estimate ** (%)
Any Anxiety Disorder
    Simple Phobia
    Social Phobia
    Panic Disorder
Any Mood Disorder
    MD Episode
    Unipolar MD
    Bipolar I
    Bipolar II
1.3 1.3
    Nonaffective Psychosis
0.2 0.2
0.2 0.2
2.1 2.1
    Anorexia Nervosa
0.1 0.1
    Severe Cognitive Impairment
1.2 1.2
Any Disorder 19.5 23.4 21.0

*Numbers in parentheses indicate the prevalence of the disorder without any comorbidity. These rates were calculated using the NCS data for GAD and PTSD, and the ECA data for OCD. The rates were not used in calculating the any anxiety disorder and any disorder totals for the ECA and NCS columns. The unduplicated GAD and PTSD rates were added to the best estimate total for any anxiety disorder (3.3%) and any disorder (1.5%).

**In developing best-estimate 1-year prevalence rates from the two studies, a conservative procedure was followed that had previously been used in an independent scientific analysis comparing these two data sets (Andrews, 1995). For any mood disorder and any anxiety disorder, the lower estimate of the two surveys was selected, which for these data was the ECA. The best estimate rates for the individual mood and anxiety disorders were then chosen from the ECA only, in order to maintain the relationships between the individual disorders. For other disorders that were not covered in both surveys, the available estimate was used.

Key to abbreviations: ECA, Epidemiologic Catchment Area; NCS, National Comorbidity Study; GAD, generalized anxiety disorder; OCD, obsessive-compulsive disorder; PTSD, post-traumatic stress disorder; MD, major depression; ASP, antisocial personality disorder.

Source: D. Regier, W. Narrow, & D. Rae, personal communication, 1999

Table 2-7. Children and adolescents ages 9 to 17 with mental or addictive disorders,* combined MECA sample

Prevalence (%)
Anxiety disorders
Mood disorders
Disruptive disorders
Substance use disorders
Any disorder

* Disorders include diagnosis-specific impairment and Child Global Assessment Scale <or=70 (mild global impairment).

Source: Shaffer et al., 1996

Not all mental disorders identified in childhood and adolescence persist into adulthood, even though the prevalence of mental disorders in children and adolescents is about the same as that for adults (i.e., about 20 percent of each age population). While some disorders do continue into adulthood, a substantial fraction of children and adolescents recover or “grow out of” a disorder, whereas, a substantial fraction of adults develops mental disorders in adulthood. In short, the nature and distribution of mental disorders in young people are somewhat different from those of adults.

Older Adults

The annual prevalence of mental disorders among older adults (ages 55 years and older) is also not as well documented as that for younger adults. Estimates generated from the ECA survey indicate that 19.8 percent of the older adult population have a diagnosable mental disorder during a 1-year period (Table 2-8). Almost 4 percent of older adults have SMI, and just under 1 percent has SPMI (Kessler et al., 1996); these figures do not include individuals with severe cognitive impairments such as Alzheimer’s disease.

Future Directions for Epidemiology

The epidemiology of mental disorders is somewhat handicapped by the difficulty of identifying a “case” of a mental disorder. “Case” is an epidemiological term for someone who meets the criteria for a disease or disorder. It is not always easy to establish a threshold for a mental disorder, particularly in light of how common symptoms of mental distress are and the lack of objective, physical symptoms. It is sometimes difficult to determine when a set of symptoms rises to the level of a mental disorder, a problem that affects other areas of health (e.g., criteria for certain pain syndromes). In many cases, symptoms are not of sufficient intensity or duration to meet the criteria for a disorder and the threshold may vary from culture to culture.

Table 2-8. Best estimate prevalence rates based on Epidemiological Catchment Area, age 55+

Prevalence (%)
Any Anxiety Disorder 11.4
Simple Phobia 7.3
Social Phobia 1.0
Agoraphobia 4.1
Panic Disorder 0.5
Obsessive-Compulsive Disorder 1.5
Any Mood Disorder 4.4
Major Depressive Episode 3.8
Unipolar Major Depression 3.7
Dysthymia 1.6
Bipolar I 0.2
Bipolar II 0.1
Schizophrenia 0.6
Somatization 0.3
Antisocial Personality Disorder 0.0
Anorexia Nervosa 0.0
Severe Cognitive Impairment 6.6
Any Disorder 19.8

Source: D. Regier, W. Narrow, & D. Rae, personal communication, 1999

Diagnosis of mental disorders is made on the basis of a multidimensional assessment that takes into account observable signs and symptoms of illness, the course and duration of illness, response to treatment, and degree of functional impairment. One problem has been that there is no clearly measurable threshold for functional impairments. Efforts are currently under way in the epidemiology of mental disorders to create a threshold, or agreed-upon minimum level of functional limitation, that should be required to establish a “case” (i.e., a clinically significant condition). Epidemiology reflecting the state of psychiatric nosology during the past two decades has focused primarily on symptom clusters and has not uniformly applied—or, at times, even measured—the level of dysfunction. Ongoing reanalyses of existing epidemiological data are expected to yield better understanding of the rates of mental disorder and dysfunction in the population.

Another limitation of contemporary mental health knowledge is the lack of standard measures of “need for treatment,” particularly those which are culturally appropriate. Such measures are at the heart of the public health approach to mental health. Current epidemiological estimates therefore cannot definitively identify those who are in need of treatment. Other estimates presented in Chapter 6 indicate that some individuals with mental disorders are in treatment and others are not; some are seen in primary care settings and others in specialty care. In the absence of valid measures of need, rates of disorder estimated in epidemiological surveys serve as an imperfect proxy for the need for care and treatment (Regier et al., in press).

Subsequent sections of this report reveal the population basis of our understanding of mental health. Where appropriate, the report discusses mental health and illness across the entire population. At other times, the focus is on care in specialized mental health settings, primary health care, schools, the criminal justice system, and even the streets. A mainstream public health and population-based perspective demands such a broad view of mental health and mental illness.

Costs of Mental Illness

The costs of mental illness are exceedingly high. Although the question of cost is discussed more fully in Chapter 6, a few of the central findings are presented here. The direct costs of mental health services in the United States in 1996 totaled $69.0 billion. This figure represents 7.3 percent of total health spending. An additional $17.7 billion was spent on Alzheimer’s disease and $12.6 billion on substance abuse treatment. Direct costs correspond to spending for treatment and rehabilitation nationwide.

When economists calculate the costs of an illness, they also strive to identify indirect costs. Indirect costs can be defined in different ways, but here they refer to lost productivity at the workplace, school, and home due to premature death or disability. The indirect costs of mental illness were estimated in 1990 at $78.6 billion (Rice & Miller, 1996). More than 80 percent of these costs stemmed from disability rather than death because mortality from mental disorders is relatively low.

3 Although addictive disorders are included as mental disorders in the DSM classification system, the ECA and NCS distinguish between addictive disorders and (all other) mental disorders. Epidemiologic data in this report follow that convention.

4 The term “serious emotional disturbance” is used in a variety of Federal statutes in reference to children under the age of 18 with a diagnosable mental health problem that severely disrupts their ability to function socially, academically, and emotionally. The term does not signify any particular diagnosis; rather, it is a legal term that triggers a host of mandated services to meet the needs of these children.

Back to Top

Home | Contents | Previous | Next