National Trends in Mental Health Disability, 1997–2009

Abstract

Objectives. I assessed recent trends in mental health disability in the US nonelderly adult population in the context of trends in physical disabilities and psychological distress.

Methods. Using data for 312 364 adults aged 18 to 64 years from the US National Health Interview Survey, 1997 to 2009, I examined time trends in self-reported disability attributed to mental health conditions, disability attributed to other chronic problems, and significant psychological distress (measured by using the K6 instrument).

Results. The prevalence of self-reported mental health disability increased from 2.0% of the nonelderly adult population in the first 3 years (1997 to 1999) to 2.7% in the last 3 years (2007 to 2009), corresponding to an increase of almost 2 million disabled adults. Disability attributed to other chronic conditions decreased and significant psychological distress did not change appreciably. Change in self-reported mental health disability was more pronounced in adults who also reported disability attributed to other chronic conditions or significant psychological distress but who had no mental health contacts in the past year.

Conclusions. These findings highlight the need for improved access to mental health services in the community and for better integration of these services with primary care.

Mental illness is often associated with significant impairment in role functioning and in relationships.^1–5 Much of the social and economic burden of mental illness is attributed to this impairment in role functioning.⁶ Although the number of individuals with mental health conditions who receive treatment in the community has increased dramatically over the past few decades,^7–9 there is no consensus as to whether these changes represent a genuine increase in the prevalence of mental illness.^10,11

Studies of trends in the community prevalence of mental disorders have produced conflicting results.^7,12–14 One study found no change in the prevalence of common mental disorders in the United States between the early 1990s and the early 2000s,⁷ whereas another study found an increase in the prevalence of depression during this time span.¹³ Other studies have reported increased prevalence of depression in specific subgroups of the population, such as young women¹⁴ or middle-aged women.¹² An analysis of 3 general population surveys conducted in 1957, 1976, and 1996 also revealed an increase in the proportion of respondents who reported an “impending nervous breakdown” during this time span, suggesting an increase in the prevalence of psychological distress over time.¹⁵ However, less information is available on national trends in mental health disability, partly because of the paucity of useful data sources.

Reports based on US Supplemental Security Income and Social Security Disability Insurance data have recorded increases in the number of disabled as a result of mental illness over the past few decades.^16,17 A recent report revealed a more than 2-fold increase in the number of individuals who received disability payments because of mental health problems between 1987 and 2007.¹⁶ Sickness benefits data from Great Britain have also recorded dramatic increases in disability attributed to “depression and neurotic conditions” between the mid 1980s and mid 1990s.¹⁸ However, disability benefits data are likely impacted by various extraneous factors, including economic cycles and changes in disability eligibility criteria.

I used recent data from a multiwave, nationally representative survey of the US general population to assess trends in self-reported mental health disability and significant psychological distress and to explore parallel trends in disability attributed to other chronic health problems. More specifically, I used data from 13 years of the US National Health Interview Survey (NHIS)¹⁹ from 1997 to 2009 to assess trends in self-reported disability and distress in the adult population. An earlier report based on the NHIS recorded an increase in self-reported mental illness disability in working-aged adults from 1988 to 1996.²⁰ By contrast, a report based on older (≥ 65 years) participants of the NHIS in 1997 to 2004 found a decline in the prevalence of both severe mental distress and self-reported disability among participants with such distress.²¹ However, mood and anxiety disorders are less common in the older populations,^22,23 and time trends of mental health disability in these populations may not be representative of trends in younger adults. In the present report, I focused on adults aged 18 to 64 years and included more recent survey waves. Furthermore, I assessed trends in disability and psychological distress within specific population subgroups.

METHODS

The NHIS is an annual, cross-sectional survey of the noninstitutionalized US population that is conducted by the National Center for Health Statistics.¹⁹ The survey consists of a family interview with a knowledgeable adult informant regarding each family member. In addition, a randomly selected adult member is interviewed regarding chronic conditions and disability. The data for the present report came from these adult interviews. The final response rates, which were calculated by multiplying the response rates for the family interview by those for the sample adult interview, ranged from 63% to 80% across survey years, with a median value of 72%. A total of 386 887 adults participated in the NHIS adult interviews from 1997 to 2009. Of these, 312 364 were in the age range of 18 to 64 years and made up the sample for this study.

Assessment

Disability was assessed by asking whether the participant had difficulty in performing a wide range of physical activities (e.g., walking a certain distance), performing instrumental activities of daily life (e.g., going shopping), participating in social activities (e.g., going to parties), and doing things to relax at home or for leisure (Appendix A, available as a supplement to the online version of this article at http://www.ajph.org). Each difficulty area was scored on a scale from “not at all difficult” (0) to “can't do it” (4). The participant was next asked to select from a list those health conditions or problems that caused difficulty with these activities.

The list of health conditions assessed in 1997–2003 NHIS was somewhat different from the list in 2004–2009 NHIS. Therefore, the analyses were limited to 19 conditions that were assessed consistently across the survey years. These conditions included some of the most common chronic medical conditions (e.g., problems with vision or hearing, arthritis or rheumatism, back or neck problems, fractures or other bone or joint injuries, heart disease, stroke, hypertension, diabetes, lung disease such as asthma and emphysema, cancer, birth defects, developmental problems such as mental retardation and cerebral palsy, obesity, and missing limbs) as well as a category of mental health problems described as “depression, anxiety, or emotional problems.” For the purpose of this study, the disability attributed to mental health problems—whether other conditions were reported or not—was contrasted with disability attributed to all other chronic conditions combined. In this article, I used the term mental health disability to identify this NHIS rating of self-reported disability attributed to depression, anxiety, or emotional problems.

As would be expected, participants with mental health disability reported a different set of difficulties than did those who reported disability attributed to other, mostly physical, chronic conditions. Participants with mental health disability reported a greater level of difficulty in going shopping, to the movies, or to sporting events; participating in social activities; and doing things to relax at home or for leisure. By contrast, participants with non-mental-health chronic disabilities reported a greater level of difficulty with physical activities (Appendix A, available as a supplement to the online version of this article at http://www.ajph.org).

The validity of the mental health disability question was assessed against the 12-month diagnoses of major depressive episodes and generalized anxiety disorder according to the American Psychiatric Association's Diagnostic and Statistical Manual of Mental Disorders, fourth edition (DSM-IV)²⁴ and by using the Composite International Diagnostic Interview-Short Form (CIDI-SF)²⁵ administered in 1999. The psychometric properties of the CIDI-SF and its use in the NHIS 1999 have been described elsewhere.^25–27 The mental health disability rating was strongly associated with the 2 DSM-IV/CIDI-SF disorders assessed (unadjusted odds ratio [OR] = 19.24; 99% confidence interval [CI] = 14.96, 24.74; P < .001): 70.3% of the participants with mental health disability also met the criteria for the disorders compared with 11.0% of the other participants. Duration of the mental health disability was assessed by asking the participants, “How long have you had depression, anxiety, or an emotional problem?”

Psychological distress was assessed with K6, a standardized screening instrument developed specifically for use in the general population surveys.^28,29 The K6 rating is the summary score of 6 items, assessing frequency of nonspecific psychological distress over the past 30 days. Items are rated on a Likert scale from “none” (0) to “all the time” (4) with a summary score range of 0 to 24. In a clinical reappraisal study, the K6 was shown to have acceptable psychometric properties to detect DSM-IV²⁴ mental disorders as ascertained by the Structured Clinical Interview for DSM-IV (SCID)³⁰ and a score of 60 or less on the Global Assessment of Functioning scale (GAF).²⁹ On the basis of ROC curve analysis,³¹ a score of 13 or higher on the K6 was selected as the optimal cutoff that equalized false-positive and false-negative results.²⁹ This cutoff is associated with a sensitivity of 0.36, a specificity of 0.92, and a total classification accuracy of 0.92. On the basis of this cutoff, 3 categories of psychological distress were identified for this study: “no psychological distress” (K6 = 0), “some psychological distress” (K6 = 1–12), and “significant psychological distress” (K6 ≥ 13).

Contact with mental health professionals in the past year was assessed by asking the participants whether during the past 12 months they had seen or talked to a mental health professional such as a psychiatrist, psychologist, psychiatric nurse, or clinical social worker. In addition, the participants were asked whether there was any time in the past year when they needed mental health care or counseling but could not afford it.

Health insurance was assessed by a series of questions regarding current health insurance, the last time the participant was covered by any health insurance (for those without current insurance), and whether there was a gap in insurance coverage in the past year (for those with current insurance). On the basis of these variables, 3 categories of insurance coverage for the past year were defined: (1) covered all of the past year, (2) covered for some part of the past year but not all of it, and (3) no health insurance throughout the past year.

In addition, the variables of gender, age, and racial/ethnic group (non-Hispanic White, non-Hispanic Black, Hispanic, and other) and the ratio of family income over the Federal Poverty Level³² were included in the analyses.

Analysis

I conducted the analyses in 2 stages. First, I assessed time trends in disability attributed to mental health problems and significant psychological distress across survey years by using bivariate and multivariate logistic regression models. The independent variable of interest in these analyses was the transformed survey year computed by subtracting 1997 from the survey year and dividing the results by 12. Thus, the transformed variable took a value of 0 for 1997 and 1 for 2009. The exponentiated logistic regression coefficient associated with this variable represented a change in odds of outcome across the whole 1997 to 2009 period. Multivariate analyses were adjusted for gender, age, race/ethnicity, and family income, thus adjusting for variations in these characteristics across survey waves. To provide a comparison, I conducted similar regression analyses for disability attributed to other chronic conditions.

Second, to assess whether the observed trends were general or were limited to specific groups of individuals, I repeated the logistic regression analyses within specific sociodemographic strata. In addition, I conducted stratified analyses within subgroups of participants defined according to the presence of disability attributed to other chronic conditions, the level of psychological distress, and contact with mental health professionals. Further stratified analyses focused on participants categorized according to (1) both disability caused by other chronic conditions and mental health contacts, and (2) both psychological distress and mental health contacts. Variations in regression coefficients across strata were examined by using interaction terms of stratification variable with the variable of survey year in a logistic regression analysis of the full sample.

Any change in the prevalence of a condition can be attributed either to a change in its incidence or to a change in its duration. In further analyses, I assessed variations in duration of mental health problems across survey years by using a linear regression model. In further analyses, I also assessed the association of mental health disability with nonreceipt of needed mental health care because of unaffordability.

The NHIS uses a complex sampling design. Sampling weights, strata, and primary sampling units provided by the National Center for Health Statistics were used to make the results representative of the US general population by using the complex survey analysis modules of the STATA 11.0 software (StataCorp, College Station, TX). All percentages reported were weighted by sampling weights (inverse of the probability that the observation is included because of the sampling design and nonresponse) to make the results representative of the US population. Because of the very large sample size, a value of P < .01 was used to judge the statistical significance of the tests.

RESULTS

Overall, 66 885 (20.7%) of the 312 364 NHIS participants aged 18 to 64 years reported a disability. In bivariate analyses, the percentage of participants reporting any such disability did not change appreciably from 1997–1999 (20.7%) to 2007–2009 (20.9%; OR = 0.97; 99% CI = 0.94, 1.01; P = .056). However, the percentage with mental health disability increased from 2.0% in 1997–1999 to 2.7% in 2007–2009 (OR = 1.50; 99% CI = 1.31, 1.71; P < .001). Based on population estimates for US adults aged 18 to 64 years,^33,34 this change translates into an increase of approximately 2 million adults with mental health disability in this time period (from approximately 3.2 million to approximately 5.2 million). Disability attributed to other chronic conditions did not change appreciably from 1997–1999 (19.7%) to 2007–2009 (19.8%; OR = 0.99; 99% CI = 0.94, 1.05; P = .722). Furthermore, the odds of significant psychological distress as defined by a K6 score of 13 or higher only modestly increased from 2.9% in 1997–1999 to 3.2% in 2007–2009, and this change did not reach the P < .01 significance level (OR = 1.10; 99% CI = 0.99, 1.23; P = .018).

In multivariate analyses adjusted for sociodemographic characteristics of the NHIS participants across survey waves, the association of survey year with mental health disability persisted, whereas a decreasing trend for disability attributed to other chronic conditions was found (Table 1), and no meaningful changes were observed in significant psychological distress (adjusted OR = 1.06; 99% CI = 0.94, 1.18; P = .227). Furthermore, gender, age, race/ethnicity, and family income were independently related to disability attributed to both mental health problems and other chronic conditions (Table 1).

TABLE 1.

Association of Survey Year With Mental Health and Other Chronic Condition Disability in Multivariate Logistic Regression Analyses of Nonelderly Adults: US National Health Interview Survey, 1997–2009

	Mental Health Disability		Other Chronic Condition Disability
Variables	AOR (95% CI)	P	AOR (95% CI)	P
Survey yeara	1.41 (1.23, 1.62)	< .001	0.89 (0.84, 0.94)	< .001
Gender
Women (Ref)	1.00		1.00
Men	0.71 (0.65, 0.77)	< .001	0.74 (0.71, 0.76)	< .001
Age, y
18–24	1.00		1.00
25–34	1.48 (0.24, 1.76	< .001	1.60 (1.47, 1.73)	< .001
35–44	2.09 (1.76, 2.48)	< .001	2.87 (2.65, 3.11)	< .001
45–54	2.54 (2.14, 3.03)	< .001	5.23 (4.83, 5.65)	< .001
55–64	2.18 (1.81, 2.62)	< .001	8.30 (7.66, 8.99)	< .001
Race/ethnicity
Non-Hispanic White	1.00		1.00
Non-Hispanic Black	0.57 (0.50, 0.64)	< .001	0.81 (0.77, 0.85)	< .001
Hispanic	0.51 (0.45, 0.58)	< .001	0.58 (0.55, 0.61)	< .001
Other	0.79 (0.58, 1.08)	.051	0.80 (0.69, 0.93)	< .001
Family income/federal poverty level ratio
< 1 (Ref)	1.00		1.00
1 to < 2	0.57 (0.51, 0.64)	< .001	0.75 (0.71, 0.79)	< .001
2 to < 4	0.29 (0.26, 0.33)	< .001	0.48 (0.46, 0.51)	< .001
≥ 4	0.17 (0.15, 0.19)	< .001	0.32 (0.30, 0.34)	< .001

Note. AOR = adjusted odds ratio; CI = confidence interval. AORs were obtained from logistic regression analyses with all the variables in the table included in the models.

^aSurvey year was transformed by subtracting 1997 from each year value and dividing the resulting number by 12 so that 1997 had a value of 0 and 2009 a value of 1.

In stratified analyses, the time trend for mental health disability was consistent across gender, age, racial/ethnic groups, and family income levels (Table 2). However, there were statistically significant differences in trends across groups with and without disability attributed to other chronic conditions. The trend was stronger in participants with other chronic condition disability than in those without.

TABLE 2.

Results of Stratified Analyses of the Association of Survey Year With Mental Health Disability in Nonelderly Adults: US National Health Interview Survey, 1997–2009

		Trend Within the Stratum		Statistical Tests for Comparison of Trends Across Strata
	No. (%)	OR (99% CI)	P
Gender				F(1, 6339) = 0.20; P = .653
Women	171 987 (50.9)	1.53 (1.31, 1.78)	< .001
Men	140 377 (49.2)	1.46 (1.17, 1.81)	< .001
Age, y				F(4, 6336) = 2.66; P = .031
18–24	40 230 (15.5)	1.54 (0.98, 2.43)	.014
25–34	73 450 (22.0)	1.71 (1.26, 2.32)	< .001
35–44	78 890 (24.5)	1.41 (1.11, 1.79)	< .001
45–54	68 994 (22.5)	1.22 (0.96, 1.53)	.031
55–64	50 800 (15.6)	1.81 (1.37, 2.38)	< .001
Race/ethnicity				F(3, 6331) = 1.41; P = .239
Non-Hispanic White	191 220 (72.6)	1.56 (1.33, 1.81)	< .001
Non-Hispanic Black	45 997 (12.4)	1.54 (1.08, 2.18)	.002
Hispanic	58 968 (13.6)	1.33 (0.93, 1.90)	.038
Other	4160 (1.4)	2.72 (1.16, 6.42)	.003
Family income/federal poverty level ratio				F(3, 6294) = 2.01; P = .11
< 1	39 667 (11.7)	1.60 (1.27, 2.02)	< .001
1 to < 2	45 662 (16.1)	1.46 (1.11, 1.90)	< .001
2 to < 4	76 242 (30.8)	1.37 (1.06, 1.77)	.001
≥ 4	91 701 (41.5)	1.12 (0.83–1.51)	.318
Other chronic condition disability				F(1, 6339) = 10.76; P = .001
Present	63 621 (19.7)	1.75 (1.49, 2.06)	< .001
Absent	248 743 (80.3)	1.26 (1.03, 1.55)	.003
Insurance coverage				F(2, 5993) = 0.47, P = .626
All of past y	209 066 (77.4)	1.70 (1.44, 2.01)	< .001
Some of past y	26 195 (9.0)	1.50 (1.03, 2.18)	.006
None of past y	42 150 (13.6)	1.53 (1.05, 2.24)	.004
Psychological distress				F(2, 6338) = 3.51; P = .03
No distress (K6 = 0)	152 687 (49.6)	1.05 (0.52, 2.14)	.854
Some distress (K6 = 1–12)	148 524 (47.3)	1.39 (1.17, 1.64)	< .001
Significant distress (K6 ≥ 13)	11 153 (3.1)	1.80 (1.43, 2.26)	< .001
Mental health professional contact in past y				F(1, 6339) = 5.67; P = .017
Any contact	23 539 (7.1)	1.20 (0.99, 1.45)	.017
No contact	285 350 (92.9)	1.50 (1.26, 1.79)	< .001
Mental health professional contact and other chronic condition disability				F(3, 6337) = 8.14; P < .001
Any contact and other disability	8765 (2.5)	1.19 (0.92, 1.53)	.079
Any contact and no other disability	14 774 (4.6)	1.17 (0.86, 1.60)	.197
No contact and other disability	54 429 (17.3)	1.97 (1.58, 2.45)	< .001
No contact and no other disability	230 921 (75.6)	1.16 (0.88, 1.52)	.166
Mental health professional contact and psychological distress				F(5, 6335) = 3.95; P = .001
Any contact and no distress	3990 (1.3)	0.79 (0.23, 2.69)	.615
Any contact and some distress	15 513 (4.7)	1.22 (0.93, 1.60)	.058
Any contact and significant distress	4036 (1.1)	1.35 (0.96, 1.90)	.024
No contact and no distress	146 231 (48.1)	1.00 (0.43, 2.32)	.997
No contact and some distress	132 129 (42.8)	1.33 (1.08, 1.63)	< .001
No contact and significant distress	6990 (2.0)	2.33 (1.67, 3.24)	< .001

Note. CI = confidence interval; OR = odds ratio. ORs were obtained from logistic regression analyses. Analyses in each stratum assessed the association of the transformed survey year variable (independent variable) with mental health disability (dependent variable). Survey year was transformed by subtracting 1997 from each year value and dividing the resulting number by 12 so that 1997 had a value of 0 and 2009 a value of 1.

The more common chronic physical disabilities that co-occurred in more than 10% of the participants with mental health disability were disabilities caused by back and neck pain (30.1%), arthritis (24.7%), obesity (11.0%), and lung disease (10.1%). Adjustment for these specific conditions in a multivariate regression analysis did not explain away the overall time trend in mental health disability (data not shown).

Time trends across groups defined according to the extent of insurance coverage, distress level, and mental health contact were not different at the predefined level of statistical significance of P < .01 (Table 2). However, trends differed across groups defined by both mental health contact and disability attributed to other chronic conditions and groups defined by both mental health contact and level of distress (Table 2). Mental health disability increased among participants with disability attributed to other chronic conditions who did not have any contacts with mental health professionals in the past year (Table 2 and Figure 1). Mental health disability also increased significantly among participants with psychological distress who did not have any mental health contacts (Table 2 and Figure 2). Of note, the prevalence of any mental health contacts increased across survey years from 6.5% in 1997–1999 to 8.1% in 2007–2009 (OR = 1.31; 99% CI = 1.21, 1.41; P < .001).

FIGURE 1.

Trends in mental health disability in nonelderly adult participants according to other chronic condition disability and whether they (a) had any professional mental health contacts or (b) did not have such contacts in the past 12 months: US National Health Interview Survey, 1997–2009.

FIGURE 2.

Trends in mental health disability in nonelderly adult participants according to level of psychological distress and whether they (a) had any professional mental health contacts or (b) did not have such contacts in the past 12 months: US National Health Interview Survey, 1997–2009.

The median duration of mental health disability in this sample was 10 years. The duration of disability did not change significantly across survey years (linear regression coefficient = 0.402; standard error = 0.635; P = .527).

A total of 3.2% of participants in 2007–2009 reported not receiving mental health care because they could not afford it, up from 2.0% in 1997–1999 (OR = 1.86; 99% CI = 1.65, 2.09; P < .001). The increase was especially pronounced among participants with significant psychological distress (K6 ≥ 13) who had not had any mental health contacts in the past year: from 14.3% in 1997–1999 to 24.8% in 2007–2009 (OR = 2.35; 99% CI = 1.70, 3.25; P < .001). However, adjusting for nonaffordability of mental health care in a multivariate regression model did not fully explain the increasing trend of mental health disability across survey years (data not shown).

DISCUSSION

There were 2 main findings in this study. First, the prevalence of self-reported mental health disability in the US nonelderly population modestly increased over the past decade. By contrast, the prevalence of disability attributed to other chronic conditions decreased and the prevalence of significant psychological distress did not change appreciably. Second, the increase in the prevalence of mental health disability was mainly among adults with co-occurring disability attributed to other chronic conditions and among adults with a greater level of psychological distress who had not had any contacts with mental health professionals over the past year.

The finding of an increasing trend in self-reported mental health disability in this study is consistent with findings from past studies that were based on disability benefits data from the United States and Great Britain.^{16–18,20,35} Whitaker attributed the increase in the number of disability benefits awarded because of mental health problems in the United States in the 1987 to 2007 period to the growing use of psychiatric medications.¹⁶ By contrast, the present study found evidence that the greatest increase in mental health disability in recent years was among participants who had not seen any mental health professionals in the past year.

The difference in the time trends in psychological distress and mental health disability is puzzling. Although the participants did not rate themselves as more distressed in more recent years, they rated themselves as more disabled by their psychological symptoms. The selective increase in self-reported mental health disability might represent a greater appreciation of the impact of mental health problems on the person's life associated with increased mental health literacy in recent years.^36,37 This same growing trend in recognition of the impact of mental illness and the greater attribution of disability to mental health problems may have contributed to increased mental health treatment-seeking over the past 2 decades.^7,38 More speculatively, other socioeconomic developments in recent decades may have contributed to trends in self-evaluation of mental health disability; these include the decline in social capital,³⁹ the growing economic inequality,^40,41 and changes in job market from manufacturing to services jobs, which may have introduced new demands.⁴² The potential effects of these factors and the recent economic downturn⁴³ and declining employment rates on mental health disability need to be assessed in future research.

The increasing trend in mental health disability was more pronounced among participants who also reported disability attributed to other chronic conditions, particularly those who did not have any contacts with mental health professionals. This finding is intriguing and difficult to explain. The time trend could not be explained away by trends in the prevalence of specific conditions that have increased in recent years (such as obesity and back pain)^44,45 and that are common among participants with mental health disability. Interestingly, this trend coincides with a greater professional recognition of mental illness comorbidity and its consequences in patients with physical health conditions^46–48 that may have resulted in greater detection and diagnosis of these problems in primary care settings.

The finding of the common comorbidity between mental health and chronic physical disability also highlights an opportunity to address comorbid mental health issues in general medical settings where patients with chronic physical conditions often seek treatment. Over the years, several mental health interventions specifically designed for these settings have been introduced.^49–52 Nevertheless, many cases of common and disabling mental disorders in primary care settings go undetected and untreated,^53–55 and the implementation of comprehensive management programs for mental health conditions in these settings often lags behind the implementation of similar programs for other chronic conditions.⁵⁶ The growing rate of mental health disability among individuals with disability attributed to other chronic conditions calls for renewed efforts at dissemination and implementation of evidence-based mental health interventions in general medical settings.⁵⁷

The findings of this study should be interpreted in the context of the limitations of the study and of the NHIS survey. First, mental health disability in this study was assessed by a self-report measure. No objective measures of disability were used. Second, mental illness is associated with impairment in a number of domains (including cognitive functioning and relationships) that are not captured by the NHIS. Third, the K6 is not a diagnostic instrument and mental illnesses were assessed only in NHIS 1999. Therefore, time trends in the prevalence of mental disorders could not be assessed. Fourth, the NHIS measure of mental health contacts is a crude measure and captures no information regarding the extent and quality of mental health treatments among those who did have a mental health contact. However, the measure fully captures the nonreceipt of treatment from mental health professionals. Most of the change in the prevalence of mental health disability occurred in this latter group of participants. Fifth, the impact of deinstituationalization⁵⁸ and recent legislation, such as the Olmstead Decision of 1999,⁵⁹ on the number of individuals with mental illness disability who live in the community could not be examined in this study. The potential impact of these factors needs to be assessed in future research. Finally, a growing number of individuals receive mental health care in primary care and other general medical settings.⁷ The NHIS does not assess mental health treatments received in these settings.

In the context of these limitations, the data presented here provide a first glimpse of the recent trends in mental health disability in the US population. The growing trend in mental health disability, especially among individuals with physical disability or significant psychological distress and limited access to mental health professionals, points to the growing need for mental health care in this population. Despite advances in mental health care,⁵⁸ large segments of the population with common mental disorders remain out of care or receive substandard care as a result of rising financial and structural barriers.⁶⁰

One of the aims of the mental health parity legislation of 2008 was to reduce financial barriers to mental health services.⁶¹ The Affordable Care Act of 2010 also sought to increase access by expanding insurance benefits to uninsured Americans.⁶² Although these legislative initiatives may not fully eliminate the mental health treatment gap,⁶³ they seek to remove some of the more prominent financial barriers. It will be important to continue tracking trends in service use and mental health disability in the US population as these initiatives unfold in the coming years.