Prevalence of Mental Illness in the United
States: Data Sources and Estimates

Erin Bagalman
Analyst in Health Policy
Angela Napili
Information Research Specialist
April 24, 2013
Congressional Research Service
7-5700
www.crs.gov
R43047
CRS Report for Congress
Pr
epared for Members and Committees of Congress

Prevalence of Mental Illness in the United States: Data Sources and Estimates

Summary
Determining how many people have a mental illness can be difficult, and prevalence estimates
vary. While numerous surveys include questions related to mental illness, few provide prevalence
estimates of diagnosable mental illness (e.g., major depressive disorder as opposed to feeling
depressed, or generalized anxiety disorder as opposed to feeling anxious), and fewer still provide
national prevalence estimates of diagnosable mental illness. This report briefly describes the
methodology and results of three large surveys (funded in whole or in part by the U.S.
Department of Health and Human Services) that provide national prevalence estimates of
diagnosable mental illness: the National Comorbidity Survey Replication (NCS-R), the National
Comorbidity Survey Replication Adolescent Supplement (NCS-A), and the National Survey on
Drug Use and Health (NSDUH). The NCS-R and the NCS-A have the advantage of identifying
specific mental illnesses, but they are a decade old. The NSDUH does not identify specific mental
illnesses, but it has the advantage of being conducted annually.
Between February 2001 and April 2003, NCS-R staff interviewed more than 9,000 adults aged 18
or older. Analyses of NCS-R data have yielded different prevalence estimates. One analysis of
NCS-R data estimated that 26.2% of adults had a mental illness within a 12-month period
(hereinafter called 12-month prevalence). Another analysis of NCS-R data estimated the 12-
month prevalence of mental illness to be 32.4% among adults. A third analysis of NCS-R data
estimated the 12-month prevalence of mental illness excluding substance use disorders to be
24.8% among adults. The 12-month prevalence of serious mental illness was estimated to be
5.8% among adults, based on NCS-R data.
Between February 2001 and January 2004, NCS-A staff interviewed more than 10,000
adolescents aged 13 to 17. Using NCS-A data, researchers estimated the 12-month prevalence of
mental illness to be 40.3% among adolescents. Some have suggested that the current approach to
diagnosing mental illness identifies people who should not be considered mentally ill. The 12-
month prevalence of serious mental illness was estimated to be 8.0% among adolescents, based
on NCS-A data.
The National Survey on Drug Use and Health (NSDUH) is an annual survey of approximately
70,000 adults and adolescents aged 12 years or older in the United States. According to the 2011
NSDUH, the estimated 12-month prevalence of mental illness excluding substance use disorders
was 19.6% among adults aged 18 or older; this estimate was stable between 2008 and 2011. The
estimated 12-month prevalence of serious mental illness (excluding substance use disorders) was
5.0% among adults. Although the NSDUH collects information related to mental illness (e.g.,
symptoms of depression) from adolescents aged 12 to 17, it does not produce estimates of mental
illness for that population.
The prevalence estimates discussed in this report may raise questions for Congress. Should
federal mental health policy focus on adults or adolescents with any mental illness (including
some whose mental illnesses may be mild and even transient) or on those with serious mental
illness? Should substance use disorders be addressed through the same policies as other mental
illnesses? Members of Congress may approach mental health policy differently depending in part
on how they answer such questions.

Congressional Research Service

Prevalence of Mental Illness in the United States: Data Sources and Estimates

Contents
Introduction ...................................................................................................................................... 1
Estimating Prevalence of Mental Illness.......................................................................................... 2
National Comorbidity Survey Replication (NCS-R) ....................................................................... 3
Prevalence of Any Mental Illness Among Adults ...................................................................... 4
Prevalence of Serious Mental Illness Among Adults ................................................................ 4
National Comorbidity Survey Replication Adolescent Supplement (NCS-A) ................................ 5
Prevalence of Any Mental Illness Among Adolescents ............................................................. 5
Prevalence of Serious Mental Illness Among Adolescents ........................................................ 6
National Survey on Drug Use and Health (NSDUH) ...................................................................... 6
Prevalence of Any Mental Illness Among Adults ...................................................................... 7
Prevalence of Serious Mental Illness Among Adults ................................................................ 7
Concluding Comments .................................................................................................................... 7

Tables
Table 1. Examples of Survey Instruments Assessing Mental Illness ............................................... 3

Contacts
Author Contact Information............................................................................................................. 8
Acknowledgments ........................................................................................................................... 8

Congressional Research Service

Prevalence of Mental Illness in the United States: Data Sources and Estimates

Introduction
Congress has demonstrated an interest in mental health and mental illness,1 and knowing how
many people have a mental illness may be helpful in addressing related policy issues.
Determining how many people have a mental illness can be difficult, and prevalence2 estimates
vary. While numerous surveys include questions related to mental illness, few provide prevalence
estimates of diagnosable mental illness, and fewer still provide national prevalence estimates of
diagnosable mental illness.3
This report briefly describes the methodology and selected findings of three large federally
funded surveys that provide national prevalence estimates of diagnosable mental illness: the
National Comorbidity Survey Replication (NCS-R), the National Comorbidity Survey
Replication Adolescent Supplement (NCS-A), and the National Survey on Drug Use and Health
(NSDUH). This report presents prevalence estimates of any mental illness and serious mental
illness4 based on each survey and ends with a brief discussion of how these prevalence estimates
might inform policy discussions.
One data source may be preferred over another in specific situations. For example, the NCS-R
and the NCS-A are a decade old, so the NSDUH (which is conducted annually) may be preferred
for more recent prevalence estimates. On the other hand, the NCS-R and the NCS-A provide
prevalence estimates for specific disorders,5 which the NSDUH does not provide.6

1 See, for example, U.S. Congress, Senate Committee on Health, Education, Labor, & Pensions, Assessing the State of
America’s Mental Health System
, 113th Cong., 1st sess., January 24, 2013, http://www.help.senate.gov/hearings/
hearing/?id=b2048a10-5056-a032-529c-340d7ae5f237; and U.S. Congress, House Committee on Energy and
Commerce, Subcommittee on Oversight and Investigations, After Newtown: A National Conversation on Violence and
Severe Mental Illness
, 113th Cong., 1st sess., March 5, 2013, http://energycommerce.house.gov/event/after-newtown-
national-conversation-violence-and-severe-mental-illness.
2 Prevalence is the share of a population affected by a given condition during a specified period of time. The term “12-
month prevalence” refers to the share of study participants with symptoms that could be identified as mental illness in
the 12 months before the interview. Another way to express prevalence is “lifetime prevalence,” which is based on the
share of study participants that had ever had a mental illness as of the time of the interview. This report focuses on 12-
month prevalence because it more closely reflects the number of people with mental illness at a given time, which
might translate into need for services at a given time.
3 See U.S. Department of Health and Human Services (HHS), Centers for Disease Control and Prevention (CDC),
“Mental Illness Surveillance in the United States,” Morbidity and Mortality Weekly Report, vol. 60, Supplement
(September 2, 2011), http://www.cdc.gov/mmwr/preview/mmwrhtml/su6003a1.htm; and HHS, Substance Abuse and
Mental Health Services Administration (SAMHSA), Center for Behavioral Health Statistics and Quality (CBHSQ),
Comparison of NSDUH Mental Health Data and Methods with Other Data Sources, March 2012,
http://www.samhsa.gov/data/NSDUH/2k12Findings/CBHSQDataReviewC2MentalHealth2012.pdf.
4 Severity of mental illness is generally defined by the number of symptoms (i.e., whether the individual has just
enough symptoms to meet diagnostic criteria or has excess symptoms) and/or the degree of functional impairment (e.g.,
whether an individual with mental illness is unable to work as a consequence of the illness). Each survey discussed in
this report uses its own definition of serious mental illness.
5 Estimates of specific disorders are not presented in this report; they are available in the cited documents.
6 The NSDUH provides more detailed information about substance use disorders, but not other mental illness.
Congressional Research Service
1

Prevalence of Mental Illness in the United States: Data Sources and Estimates

Estimating Prevalence of Mental Illness
In clinical practice, mental health professionals diagnose mental illnesses based on criteria in the
American Psychiatric Association’s Diagnostic and Statistical Manual of Mental Disorders
(DSM)7 and exercise clinical judgment in doing so.8 Large surveys, however, may not allow for
lengthy interviews and use of clinical judgment. Considerations in generating national prevalence
estimates of diagnosable mental illness9 through large surveys include (1) what survey instrument
(i.e., set of questions) is used, (2) who administers it, and (3) how generalizable the findings are.
Each of these considerations is discussed briefly below.
When designing surveys, researchers must weigh the value of detailed information against the
time required to collect that information. Longer survey instruments may be better able to identify
mental illness accurately by asking all or most of the questions necessary to assess DSM
diagnostic criteria. Shorter survey instruments, while more practical to include in a survey, may
not identify mental illness as accurately as longer instruments. Table 1 provides examples of
survey instruments assessing mental illness.
Survey instruments may be administered in different ways, which may affect both the accuracy of
their assessment of mental illness and the feasibility of their inclusion in large surveys. Some,
such as the Structured Clinical Interview for DSM Disorders (described in Table 1), require
trained mental health professionals. Others require trained interviewers who need not be mental
health professionals. Still others may be self-administered or administered by interviewers
without extensive training.
The prevalence estimates described in this report are weighted to reflect the general U.S.
population as closely as possible.10 This is accomplished by assigning a weight to each survey
respondent based on information that is available for both the survey respondents and the general
U.S. population (e.g., age and gender, among others). Weighting can correct for known
differences between the survey respondents and the general U.S. population (e.g., if people in a
certain age range are overrepresented in the survey). Weighting cannot, however, correct for
subpopulations that are excluded from the survey altogether (e.g., the homeless). Prevalence
estimates are not generalizable to excluded subpopulations.11

7 American Psychiatric Association, Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Text
Revision (DSM-IV-TR),
(Washington, DC: American Psychiatric Association, 2000). Some of the interviews designed
for research are based on the fourth edition prior to the text revision (DSM-IV), published in 1994. A fifth edition
(DSM-5) is scheduled for release in 2013.
8 A diagnosis of mental illness may be based on what an individual reports about himself or herself; what other people
report about the individual (if, for example, a family member is available); and what the mental health professional
observes about the individual. People may not be forthcoming with information about certain behaviors (e.g., drug use)
or symptoms (e.g., hallucinations); this is a challenge in both clinical settings and research settings.
9 This report uses the term “diagnosable” as distinct from “diagnosed” in acknowledgement of the potential for surveys
to identify people suffering from mental illnesses that have not been diagnosed by a mental health professional. Also
important is the distinction between symptoms of mental illness (e.g., feelings of depression or anxiety) and a
combination of symptoms meeting DSM criteria for a diagnosis of mental illness (e.g., major depressive disorder or
generalized anxiety disorder). While all DSM disorders may be considered mental illnesses, surveys generally do not
assess all of them. In this report, some of the prevalence estimates include substance use disorders and others do not.
10 The weighting was done by the researchers whose published results CRS cites in this report, not by CRS.
11 Excluded subpopulations may have higher or lower risk of mental illness than the general population. Thus if these
subpopulations had been included in the surveys, the overall prevalence estimates would have been slightly higher or
(continued...)
Congressional Research Service
2

Prevalence of Mental Illness in the United States: Data Sources and Estimates

Table 1. Examples of Survey Instruments Assessing Mental Illness
Who Administers
How Long It
Examples of
Instrument
What It Assesses
It
Takes
Surveys Using It
Structured Clinical
Diagnoses Trained
mental
Approximately 30–
NSDUH (for a
Interview for DSM
health professionals
60 minutes
subsample of
Disorders (SCID)a
respondents)
World Mental
Diagnoses,
Trained lay
Approximately 2
NCS-R; NCS-A uses
Health Survey
functioning,
interviewers
hours, but modular
a modified
Initiative version of
treatment, risk
so that some
Composite
the Composite
factors, socio-
sections can be
International
International
demographic
excluded
Diagnostic Interview
Diagnostic Interview
characteristics, and
(CIDI)
(WMH-CIDI)b
more
World Health
Difficulties (due to
Either self-
Different versions
NSDUH uses a
Organization
health conditions) in
administered or
are available; the
modified version
Disability
the areas of
administered by an
longest (36 items)
Assessment
cognition, mobility,
interviewer (no
takes approximately
Schedule
self-care, getting
training required)
15–20 minutes
(WHODAS)c
along, life activities,
and participation
Kessler-6
Frequency and
Either self-
No more than 2
NSDUH uses a
Psychological
severity of feeling
administered or
minutes
modified version
Distress Scale (K6)d
nervous, hopeless,
administered by an
restless/fidgety,
interviewer (no
sad/depressed, and
training required)
worthless
Note: A survey instrument is a tool for systematically col ecting information from survey respondents; each
instrument listed in this table is a set of questions that may be asked as part of a larger questionnaire.
a. DSM = Diagnostic and Statistical Manual of Mental Disorders. Information about the SCID in this table is
drawn from U.S. Department of Health and Human Services (HHS), Centers for Disease Control and
Prevention (CDC), “Mental Illness Surveillance in the United States,” Morbidity and Mortality Weekly Report,
vol. 60, Supplement (September 2, 2011).
b. Information about the WMH-CIDI in this table is drawn from Ronald C. Kessler and T. Bedirhan Üstün,
“The World Mental Health (WMH) Survey Initiative Version of the World Health Organization (WHO)
Composite International Diagnostic Interview (CIDI),” International Journal of Methods in Psychiatric Research,
vol. 13, no. 2 (2004), pp. 93-121.
c. Information about the WHODAS in this table is drawn from World Health Organization, Classifications,
Classification of Functioning, Disability and Health (ICF), http://www.who.int/classifications/icf/whodasii/en/.
d. Ronald C. Kessler et al., “Screening for Serious Mental Illness in the General Population with the K6
Screening Scale: Results from the WHO World Mental Health (WMH) Survey Initiative,” International Journal
of Methods in Psychiatric Research
, vol. 19 (Supplement 1) (2010), pp. 4-22.
National Comorbidity Survey Replication (NCS-R)
The National Comorbidity Survey Replication (NCS-R) replicated the original National
Comorbidity Survey (conducted between 1990 and 1992), which was the first survey to use fully

(...continued)
lower accordingly. Policy approaches that might be appropriate for excluded subpopulations such as the homeless or
non-English speakers may be different than policy approaches for the general population.
Congressional Research Service
3

Prevalence of Mental Illness in the United States: Data Sources and Estimates

structured research diagnostic interviews to assess a wide range of DSM disorders among a
national sample of adults in the United States.12 The NCS-R was funded primarily by the National
Institute of Mental Health within the National Institutes of Health (NIH) of the Department of
Health and Human Services (HHS), with supplemental support from the National Institute on
Drug Abuse (also within NIH), the Substance Abuse and Mental Health Services Administration
(of HHS), the Robert Wood Johnson Foundation, and the John W. Alden Trust.13 Between
February 2001 and April 2003, NCS-R staff conducted in-person interviews with more than 9,000
adults aged 18 or older, drawing the sample from households in the contiguous United States. The
sample did not include the homeless, individuals in institutions, or non-English speakers; these
exclusions limit the generalizability of findings based on the NCS-R.14
Prevalence of Any Mental Illness Among Adults
The NCS-R determined the presence of any mental illness based on the World Health
Organization’s World Mental Health Survey Initiative version of the Composite International
Diagnostic Interview (WMH-CIDI), which assessed 19 specific DSM diagnoses.15 Analyses of
NCS-R data have yielded different prevalence estimates. One analysis of NCS-R data estimated
that 26.2% of adults had a mental illness within a 12-month period (hereinafter called 12-month
prevalence).16 Another analysis of NCS-R data estimated the 12-month prevalence of mental
illness to be 32.4% among adults; the difference may be attributable to the use of more recent
information about the U.S. population in weighting the NCS-R data.17 Both of these estimates
include substance use disorders as mental illness; an analysis of NCS-R data estimated the 12-
month prevalence of mental illness excluding substance use disorders to be 24.8% among
adults.18
Prevalence of Serious Mental Illness Among Adults
Additional analyses of NSC-R data were conducted to determine the 12-month prevalence of
mental illness at three levels of severity: serious,19 moderate,20 or mild.21 Among the 26.2% of

12 Harvard Medical School, Department of Health Care Policy, National Comorbidity Survey,
http://www.hcp.med.harvard.edu/ncs/.
13 Ronald C. Kessler et al., “Prevalence, Severity, and Comorbidity of Twelve-month DSM-IV Disorders in the
National Comorbidity Survey Replication (NCS-R),” Archives of General Psychiatry, vol. 62, no. 6 (June 2005), pp.
617-627. Hereinafter, Kessler et al., NCS-R Prevalence, (2005).
14 Kessler et al., NCS-R Prevalence, (2005).
15 The DSM includes more than 19 diagnoses. While all of them may be considered mental illnesses, some are often
excluded from definitions of mental illness (e.g., learning disorders).
16 Kessler et al., NCS-R Prevalence, (2005).
17 Harvard Medical School, Department of Health Care Policy, National Comorbidity Survey, “NSC-R Twelve-Month
Prevalence Estimates” link at http://www.hcp.med.harvard.edu/ncs/. This later analysis was based on updated data as of
July 19, 2007, and “reflect[s] the latest diagnostic, demographic and raw variable information.”
18 Benjamin G. Druss et al. “Impairment in Role Functioning in Mental and Chronic Medical Disorders in the United
States: Results from the National Comorbidity Survey Replication.” Molecular Psychiatry, vol. 14, no. 7 (July 2009),
pp. 728-737.
19 Kessler et al., NCS-R Prevalence, (2005). Any of the following qualified a case of mental illness as serious: “a 12-
month suicide attempt with serious lethality intent; work disability or substantial limitation due to a mental or substance
disorder; positive screen results for non-affective psychosis; bipolar I or II disorder; substance dependence with serious
role impairment (as defined by disorder-specific impairment questions); an impulse control disorder with repeated
serious violence; or any disorder that resulted in 30 or more days out of role in the year” (p. 618).
Congressional Research Service
4

Prevalence of Mental Illness in the United States: Data Sources and Estimates

adults identified with a mental disorder in the analysis, serious disorders (22.3% among adults
with a disorder
) were less common than moderate disorders (37.3%) or mild disorders (40.4%).
The estimated 12-month prevalence of serious mental illness among all adults was 5.8%.22
National Comorbidity Survey Replication
Adolescent Supplement (NCS-A)

The National Comorbidity Survey Replication Adolescent Supplement (NCS-A) was the first
survey to use fully structured research diagnostic interviews to assess a wide range of DSM
disorders among a national sample of adolescents in the United States. Like the NCS-R, the NCS-
A was funded primarily by the National Institute of Mental Health, with supplemental support
from the National Institute on Drug Abuse, the Substance Abuse and Mental Health Services
Administration, the Robert Wood Johnson Foundation, and the John W. Alden Trust.23 Between
February 2001 and January 2004, NCS-A staff interviewed more than 10,000 adolescents aged 13
to 17, drawing the sample from both schools and households in the coterminous United States.
Adolescent participants were interviewed in person, and one parent of each participating
adolescent was asked to complete a self-administered questionnaire about the adolescent’s
developmental history and mental health. The sample did not include the homeless, individuals in
institutions, or non-English speakers; these exclusions limit the generalizability of findings based
on the NCS-A.24
Prevalence of Any Mental Illness Among Adolescents
The NCS-A determined the presence of any mental illness based on a modified Composite
International Diagnostic Interview (CIDI), which assessed 15 specific DSM diagnoses. Using
NCS-A data, researchers estimated the 12-month prevalence of mental illness to be 40.3% among
adolescents.25 Some have suggested that the current version of the DSM identifies people who
should not be considered mentally ill.26

(...continued)
20 Kessler et al., NCS-R Prevalence, (2005). Among cases of mental illness not defined as serious, any of the following
qualified a case as moderate: “suicide gesture, plan, or ideation; substance dependence without serious role impairment;
at least moderate work limitation due to a mental or substance disorder; or any disorder with at least moderate role
impairment in 2 or more domains of the Sheehan Disability Scale [which] assessed disability in work role performance,
household maintenance, social life, and intimate relationships” (p. 618).
21 Kessler et al., NCS-R Prevalence, (2005). All cases of mental illness that were not defined as serious or moderate
were considered mild (p. 618).
22 Kessler et al., NCS-R Prevalence, (2005).
23 Ronald C. Kessler et al., “Prevalence, Persistence, and Sociodemographic Correlates of DSM-IV Disorders in the
National Comorbidity Survey Replication Adolescent Supplement,” Archives of General Psychiatry, vol. 69, no. 4
(April 2012). Hereinafter, Kessler et al., NCS-A Prevalence, (2012).
24 Ronald C. Kessler et al., “Severity of 12-Month DSM-IV Disorders in the National Comorbidity Survey Replication
Adolescent Supplement,” Archives of General Psychiatry, vol. 69, no. 4 (April 2012). Hereinafter, Kessler et al., NCS-
A Severity, (2012).
25 Kessler et al., NCS-A Prevalence, (2012).
26 Kessler et al., NCS-A Severity, (2012).
Congressional Research Service
5

Prevalence of Mental Illness in the United States: Data Sources and Estimates

Prevalence of Serious Mental Illness Among Adolescents
Additional analyses of NCS-A data were conducted to determine the presence of serious mental
illness (called serious emotional disturbance in NCS-A), which was defined as at least one
diagnosis accompanied by an estimated score of 50 or less on the Children’s Global Assessment
Scale (CGAS), indicating either severe functional impairment in one area of living or moderate
functional impairment in most areas of living. The CGAS score was not measured directly as part
of the NCS-A. Instead the NCS-A used a sophisticated approach that involved (1) selection of a
subsample of 347 adolescent-parent pairs from the main study; (2) follow-up telephone
interviews conducted by mental health professionals who assigned CGAS scores; (3)
development of a statistical model linking CGAS scores to responses to questions in the main
study; and (4) application of the statistical model to the full sample to impute CGAS scores based
on responses to questions in the main study.27
Using the imputed CGAS scores, researchers assessed serious mental illness among 6,483
adolescent NCS-A participants with complete data (including parent questionnaires). This
analysis yielded a 12-month prevalence of any mental illness (42.6%) that was slightly higher
than the previous estimate and found most cases to be mild (58.2% among adolescents with a
disorder
) or moderate (22.9%), rather than serious (18.8%). The estimated 12-month prevalence
of serious mental illness among all adolescents was 8.0%. Some people have suggested that most
adolescents with mental disorders do not need treatment because their disorders are mild and will
resolve on their own; however, some research has shown that mild disorders during adolescence
may predict serious disorders during adulthood.28
National Survey on Drug Use and Health (NSDUH)
The National Survey on Drug Use and Health (NSDUH) focuses primarily on the use of illegal
drugs, alcohol, and tobacco and also includes several modules that focus on mental health
issues.29 The NSDUH is funded by the Substance Abuse and Mental Health Services
Administration. Each year, the NSDUH surveys approximately 70,000 non-institutionalized
civilians aged 12 years or older in the United States, divided roughly between 45,000 adults (aged
18 or older) and 25,000 adolescents (aged 12 to 17). The NSDUH is conducted in both English
and Spanish. Participants are interviewed in their homes using a combination of personal
interviewing and audio computer-assisted self-interviewing, which offers more privacy in order to
encourage honest reporting of sensitive topics such as illicit drug use. The sample does not
include the homeless, individuals in institutions, those who speak a language other than English

27 Kessler et al., NCS-A Prevalence, (2012); Kessler et al., NCS-A Severity, (2012); and Ronald C. Kessler et al., “The
National Comorbidity Survey Adolescent Supplement (NCSA): III. Concordance of DSM-IV/CIDI Diagnoses with
Clinical Reassessments,” Journal of the American Academy of Child and Adolescent Psychiatry, vol. 48, no. 4 (April
2009), pp. 386-399.
28 Kessler et al., NCS-A Severity, (2012).
29 U.S. Department of Health and Human Services (HHS), Substance Abuse and Mental Health Services
Administration (SAMHSA), Center for Behavioral Health Statistics and Quality (CBHSQ), Results from the 2011
National Survey on Drug Use and Health: Mental Health Findings
, (SMA) 12-4725, Rockville, MD, November 2012,
http://www.samhsa.gov/data/NSDUH/2k11MH_FindingsandDetTables/Index.aspx. Hereinafter, 2011 NSDUH: Mental
Health Findings.

Congressional Research Service
6

Prevalence of Mental Illness in the United States: Data Sources and Estimates

or Spanish, or military personnel on active duty; these exclusions limit the generalizability of
findings based on the NSDUH.30
NSDUH-based prevalence estimates of any mental illness and serious mental illness among
adults aged 18 or older are described below. Although the NSDUH collects information related to
mental illness (e.g., symptoms of depression) from adolescents aged 12 to 17, it does not produce
prevalence estimates of mental illness for that population.
Prevalence of Any Mental Illness Among Adults
The NSDUH determines the presence of any mental illness based on a combination of relatively
short modules of questions (including modified versions of the K6 and WHODAS described in
Table 1) in the main survey and information from an additional follow-up interview conducted
with a subsample of adults from the main survey.31 Unlike the NCS-R and the NCS-A, the core
component of NSDUH does not include questions designed to identify specific DSM diagnoses.
Instead, the NSDUH uses an approach similar to the one used by NCS-A to impute the CGAS,
involving (1) selection of a subsample of adults from the main study; (2) follow-up telephone
interviews conducted by mental health professionals who conduct the Structured Clinical
Interview for DSM Disorders (SCID); (3) development of a statistical model linking the SCID-
based diagnosis of mental illness to responses to questions in the main study; and (4) application
of the statistical model to the full sample to impute the presence of mental illness based on
responses to questions in the main study.32 According to the 2011 NSDUH, the estimated 12-
month prevalence of mental illness excluding substance use disorders is 19.6% of adults (aged 18
or older); this estimate was stable between 2008 and 2011.33
Prevalence of Serious Mental Illness Among Adults
The NSDUH identifies adults (aged 18 or older) as having a serious mental illness if (1) they
have a mental illness (excluding substance use disorders and developmental disorders) and (2) the
illness substantially interferes with or limits at least one major life activity. The same approach
used to impute any mental illness is applied to impute serious mental illness. According to the
2011 NSDUH, the estimated 12-month prevalence of serious mental illness excluding substance
use disorders
is 5.0% among adults (aged 18 or older).34
Concluding Comments
Knowing how many people are likely to be affected by policies related to mental illness may help
policy makers identify specific problems as well as their scope; however, the national prevalence
estimates discussed in this report may raise as many questions as they answer. For example, given

30 2011 NSDUH: Mental Health Findings. The NSDUH has been ongoing since 1971; prior to 2002, it was called the
National Household Survey on Drug Abuse.
31 2011 NSDUH: Mental Health Findings.
32 This approach, called the Mental Health Surveillance Study, began with the 2008 NSDUH.
33 2011 NSDUH: Mental Health Findings.
34 2011 NSDUH: Mental Health Findings.
Congressional Research Service
7

Prevalence of Mental Illness in the United States: Data Sources and Estimates

the difference in prevalence estimates between any mental illness and serious mental illness
among adolescents, might policy makers choose to focus on a large group of adolescents that
includes many whose mental illnesses may be mild and even transient, or might they choose to
focus more narrowly on adolescents with serious mental illness? As clinical practice is moving
toward more integrated care,35 should substance use disorders be included in the definition of
mental illness (as in the NCS-R and the NCS-A) and addressed through the same policies, or
should they be identified separately (as in the NSDUH) and addressed through different policies?
How might policy makers address mental illness among populations that are excluded from the
prevalence estimates (e.g., the homeless)? Policy makers may come to different conclusions
about the best policy approach depending in part on how they answer such questions.

Author Contact Information

Erin Bagalman
Angela Napili
Analyst in Health Policy
Information Research Specialist
ebagalman@crs.loc.gov, 7-5345
anapili@crs.loc.gov, 7-0135

Acknowledgments
The authors gratefully acknowledge the significant contribution made to this report by Mariel C. Garcia
during her internship at CRS.

35 Like clinical practice, financing is moving toward including substance use disorders with the rest of mental illness.
The Mental Health Parity Act of 1996 (MHPA, P.L. 104-204) did not apply to coverage of substance use disorders. The
more recent Paul Wellstone and Pete Domenici Mental Health Parity and Addiction Equity Act of 2008 (MHPAEA,
P.L. 110-343) applies to both substance use disorders and the rest of mental illness. See CRS Report R41768, Mental
Health Parity and Mandated Coverage of Mental Health and Substance Use Disorder Services After the ACA
.
Congressional Research Service
8