Page 1

2 The burden of mental disorders

worldwide

Results from the World Mental

Health surveys

Ronald C. Kessler, Sergio Aguilar- Gaxiola,

Jordi Alonso, Somnath Chatterji, Sing Lee,

Daphna Levinson, Johan Ormel,

T. Bedirhan Üstün, and Philip S. Wang

Introduction

This chapter reviews epidemiologic evidence on the burden of mental dis-

orders worldwide and focuses on estimates of the disability of commonly

occurring mental disorders. Many studies in high- income countries have

estimated the effects of specific disorders on disability (Berto, D’Ilario,

Ruffo, Di Virgilio, & Rizzo, 2000; Maetzel & Li, 2002; Reed, Lee, &

McCrory, 2004). In particular, a considerable amount of research has

been carried out in the U.S. to quantify the magnitude of the short- term

societal costs of mental disorders in terms of healthcare expenditures,

impaired functioning, and reduced longevity (Greenberg & Birnbaum,

2005; Greenberg et al., 1999). The magnitude of the cost estimates in

these studies is staggering. For example, Greenberg and colleagues (1999)

estimated that over the decade of the 1990s, the annual societal costs of

anxiety disorders in the U.S. exceeded $42 billion. Further, this estimate is

likely conservative, as it excludes the indirect costs of early- onset anxiety

disorders due to adverse life course outcomes (e.g., the effects of child-

adolescent anxiety disorders on subsequent low educational attainment

and consequent long- term effects on income) and to increased risk of

other disorders (e.g., the effects of anxiety disorders on subsequent cardio-

vascular disorder).

Comparative studies, however, are rare (Druss et al., 2008; Merikangas

et al., 2007). However, data on comparative illness burden are critical for

making resources allocation decisions, as these decisions inevitably require

comparative assessments (Lopez & Mathers, 2007; Murray & Lopez,

1996; Murray, Lopez, Mathers, & Stein, 2001). Recognizing the impor-

tance of this information, one of the main aims of the World Health

Organization’s (WHO) World Mental Health (WMH) survey initiative is

to produce comparative data on the prevalence and severity of mental

10 R. C. Kessler et al.

disorders in participating WMH countries throughout the world. Although

this is still a work in progress, enough useful information has been pro-

duced to warrant a review.

The World Mental Health (WMH) survey initiative

The WMH survey initiative is designed to help countries throughout the

world carry out and analyze epidemiologic surveys on the prevalence and

correlates of mental disorders. A key aim is to help countries that would

not otherwise have the expertise or infrastructure to implement high-

quality, community epidemiologic surveys, by providing centralized instru-

ment development, training, and data analysis (www.hcp.med.harvard.

edu/wmh). Twenty- four countries so far have completed WMH surveys

(see Table 2.1). The vast majority of these surveys are nationally represent-

ative, although a few report data from only a single region (e.g., the São

Paolo Metropolitan Area in Brazil and the Beijing, Shanghai, and Shenzhen

Metropolitan Areas in the People’s Republic of China) or regions (e.g., six

metropolitan areas in Japan). Detailed descriptions of the field procedures

(Pennell et al., 2008) and sample characteristics (Heeringa et al., 2008) of

the WMH surveys are presented elsewhere.

All WMH surveys use the same diagnostic interview, the WHO Com-

posite International Diagnostic Interview (CIDI) (Kessler & Üstün, 2004).

The CIDI is a state- of-the- art, fully structured research diagnostic interview

designed for use by trained lay interviewers who do not have any clinical

experience. Consistent training materials, training programs, and quality

control monitoring procedures are used in all WMH surveys to guarantee

comparability across surveys. The same WHO translation, back-

translation, and harmonization procedures for the survey and the training

materials are also used across countries (Harkness et al., 2008; Pennell et

al., 2008). Blinded clinician re- interviews with a probability subsample of

WMH respondents confirm that the diagnoses generated by the CIDI are

consistent with independent clinical diagnoses generated by culturally com-

petent clinicians (Haro et al., 2006).

Due to our interest in disease burden, the CIDI was designed to go well

beyond the mere assessment of mental disorders to include a wide range of

measures about a number of correlates. Five of these are of special impor-

tance for the current report. First, the CIDI assesses disorder severity,

which is important in light of the finding in previous epidemiologic surveys

that quite a high proportion of the general population in many countries

meets criteria set out in the Diagnostic and Statistical Manual of Mental

Disorders (DSM) or the International Classification of Disorders (ICD) for

a mental disorder (Somers, Goldner, Waraich, & Hsu, 2006; Waraich,

Goldner, Somers, & Hsu, 2004; Wittchen & Jacobi, 2005). Faced with

this high prevalence, mental health policy planning efforts need to consider

disorder severity for treatment planning purposes, as the simple presence

Table 2.1

World Mental Health sample characteristics by World Bank Income Categories

a

Survey

b

Sample Characteristics

c

Field Dates

Age Range

Sample Size

Response Rate

d

Part I

Part II

I. Low/lower-middle income countries Colombia

NSMH

Stratified multistage clustered area probability sample of household residents in all urban areas of the country (approximately 73% of the total national population)

2003

18–65

4,426

2,381

87.7

India

WMHI

Stratified multistage clustered area probability sample of household residents in Pondicherry region. NR

2003–2005

18+

2,992

1,373

98.8

Iraq

IMHS

Stratified multistage clustered area probability sample of household residents. NR

2006–2007

18+

4,332

4,332

95.2

Nigeria

NSMHW

Stratified multistage clustered area probability sample of households in 21 of the 36 states in the country, representing 57 percent of the national population. The surveys were conducted in Yoruba, Igbo, Hausa, and Efik languages.

2002–2003

18+

6,752

2,143

79.3

PRC

B-WMH S-WMH

Stratified multistage clustered area probability sample of household residents in the Beijing and Shanghai metropolitan areas.

2002–2003

18+

5,201

1,628

74.7

continued

PRC

Shenzhen

Stratified multistage clustered area probability sample of household residents and temporary residents in the Shenzhen area.

2006–2007

18+

7,134

2,476

80.0

Ukraine

CMDPSD

Stratified multistage clustered area probability sample of household residents. NR

2002

18+

4,725

1,720

78.3

Total

35,562

16,053

82.6

II. Upper-middle income countries Brazil

São Paulo Megacity

Stratified multistage clustered area probability sample of household residents in the São Paulo metropolitan area.

2005–2007

18+

5,037

2,942

81.3

Bulgaria

NSHS

Stratified multistage clustered area probability sample of household residents. NR

2003–2007

18+

5,318

2,233

72.0

Lebanon

LEBANON

Stratified multistage clustered area probability sample of household residents. NR

2002–2003

18+

2,857

1,031

70.0

Table 2.1

continued

Survey

b

Sample Characteristics

c

Field Dates

Age Range

Sample Size

Response Rate

d

Part I

Part II

Mexico

M-NCS

Stratified multistage clustered area probability sample of household residents in all urban areas of the country (approximately 75% of the total national population).

2001–2002

18–65

5,782

2,362

76.6

Romania

RMHS

Stratified multistage clustered area probability sample of household residents. NR

2005–2006

18+

2,357

2,357

70.9

South Africa

SASH

Stratified multistage clustered area probability sample of household residents. NR

2003–2004

18+

4,315

4,315

87.1

Total

25,666

15,240

76.6

III. High income countries Belgium

ESEMeD

Stratified multistage clustered probability sample of individuals residing in households from the national register of Belgium residents. NR

2001–2002

18+

2,419

1,043

50.6

France

ESEMeD

Stratified multistage clustered sample of working telephone numbers merged with a reverse directory (for listed numbers). Initial recruitment was by telephone, with supplemental in-person recruitment in households with listed numbers. NR

2001–2002

18+

2,894

1,436

45.9

continued

Germany

ESEMeD

Stratified multistage clustered probability sample of individuals from community resident registries. NR

2002–2003

18+

3,555

1,323

57.8

Israel

NHS

Stratified multistage clustered area probability sample of individuals from a national resident register. NR

2002–2004

21+

4,859

4,859

72.6

Italy

ESEMeD

Stratified multistage clustered probability sample of individuals from municipality resident registries. NR

2001–2002

18+

4,712

1,779

71.3

Japan

WMHJ2002– 2006

Un-clustered two-stage probability sample of individuals residing in households in eleven metropolitan areas

2002–2006

20+

4,129

1,682

55.1

Netherlands

ESEMeD

Stratified multistage clustered probability sample of individuals residing in households that are listed in municipal postal registries. NR

2002–2003

18+

2,372

1,094

56.4

New Zealand

e

NZMHS

Stratified multistage clustered area probability sample of household residents. NR

2003–2004

18+

12,790

7,312

73.3

N Ireland

NISHS

Stratified multistage clustered area probability sample of household residents. NR

2004–2007

18+

4,340

1,986

68.4

Table 2.1

continued

Survey

b

Sample Characteristics

c

Field Dates

Age Range

Sample Size

Response Rate

d

Part I

Part II

Portugal

NMHS

Stratified multistage clustered area probability sample of household residents. NR

2008–2009

18+

3,849

2,060

57.3

Spain

ESEMeD

Stratified multistage clustered area probability sample of household residents. NR

2001–2002

18+

5,473

2,121

78.6

United States

NCS-R

Stratified multistage clustered area probability sample of household residents. NR

2002–2003

18+

9,282

5,692

70.9

Total

60,674

32,387

65.4

IV. Total

121,902

63,680

72.0

Notes a

The World Bank. (2008). Data and Statistics. Accessed May 12, 2009 at: http://go.worldbank.org/D7SN0B8YU0

b NSMH (The Colombian National Study of Mental Health); WMHI (World Mental Health India); IMHS (Iraq Mental Health Survey); NSMHW

(The

Nigerian Survey of Mental Health and Wellbeing); B-WMH (The Beijing World Mental Health Survey); S-WMH (The Shanghai World Mental Health Survey); CMDPSD (Comorbid Mental Disorders during Periods of Social Disruption); NSHS (Bulgaria National Survey of Health and Stress); LEBANON (Lebanese Evaluation of the Burden of Ailments and Needs of the Nation); M-NCS (The Mexico National Comorbidity Survey); RMHS (Romania Mental Health Survey); SASH (South Africa Health Survey); ESEMeD (The European Study Of The Epidemiology Of Mental Disorders); NHS (Israel National Health Survey); WMHJ2002–2006 (World Mental Health Japan Survey); NZMHS (New Zealand Mental Health Survey); NISHS (Northern Ireland Study of Health and Stress); NMHS (Portugal National Mental Health Survey); NCS-R (The US National Comorbidity Survey Replication).

c Most WMH surveys are based on stratified multistage clustered area probability household samples in which samples of areas equivalent to counties or

municipalities in the US were selected in the first stage followed by one or more subsequent stages of geographic sampling (e.g., towns within counties, blocks within towns, households within blocks) to arrive at a sample of households, in each of which a listing of household mem

bers was created and one

or two people were selected from this listing to be interviewed. No substitution was allowed when the originally sampled househ

old resident could not be

interviewed. These household samples were selected from Census area data in all countries other than France (where telephone directories were used to select households) and the Netherlands (where postal registries were used to select households). Several WMH surveys (Belgium, Germany, Italy) used municipal resident registries to select respondents without listing households. The Japanese sample is the only totally un-clustered sample, with households randomly selected in each of the four sample areas and one random respondent selected in each sample household. 18 of the 24 su

rveys are based on

nationally representative (NR) household samples.

d The response rate is calculated as the ratio of the number of households in which an interview was completed to the number of h

ouseholds originally

sampled, excluding from the denominator households known not to be eligible either because of being vacant at the time of initial contact or because the residents were unable to speak the designated languages of the survey.

e

New Zealand interviewed respondents 16+ but for the purposes of cross-national comparisons we limit the sample to those 18+.

16 R. C. Kessler et al.

of a diagnosis may not indicate a level of need sufficient to require

treatment. Consequently, all 12-month mental disorders in the WMH

surveys are classified as serious, moderate, or mild. Serious disorders are

defined as non- affective psychosis, bipolar I disorder, or substance depend-

ence with a physiological dependence syndrome; making a suicide attempt

in conjunction with any other disorder; reporting severe role impairment

due to a mental disorder in at least two areas of functioning measured by

the Sheehan Disability Scale (SDS) (Leon, Olfson, Portera, Farber, &

Sheehan, 1997); or having overall functional impairment from any dis-

order consistent with a Global Assessment of Functioning (GAF ) (Endicott,

Spitzer, Fleiss, & Cohen, 1976) score of 50 or less. Disorders not classified

as serious were classified as moderate if the respondent had substance

dependence without a physiological dependence syndrome or at least mod-

erate interference in the disorder- specific scale of role impairment. All

other disorders were classified as mild.

Second, the CIDI includes a disorder- specific measure of role impair-

ment administered in exactly the same way for each mental disorder

assessed in the surveys, as well as for each of a number of physical dis-

orders that are assessed for comparison purposes. This measure, the SDS,

is a widely used self- report measure of condition- specific role impairment.

The WMH version of the SDS consists of four questions, each asking the

respondent to rate on a 0 to 10 scale the extent to which a particular dis-

order “interfered with” activities in one of four role domains during the

month in the past year when the disorder was most severe. The four

domains include

1 “your home management, like cleaning, shopping, and taking care of

the (house/apartment)” (home);

2 “your ability to work” (work);

3 “your social life” (social); and

4 “your ability to form and maintain close relationships with other

people” (close relationships).

The 0 to 10 response options are presented in a visual analogue format

with labels for the response options (none (0), mild (1 to 3), moderate (4

to 6), severe (7 to 9), and very severe (10)). A global SDS disability score

was created for each disorder assessed in the WMH surveys by assigning

each respondent the highest SDS domain score reported across the four

domains. We found good internal consistency reliability (Cronbach’s

alpha) across the SDS domains, in the range of 0.82 to 0.92 over countries

and equivalent in both high- income countries (median 0.86; inter- quartile

range 0.84 to 0.88) and lower- middle-income countries (median 0.90;

inter- quartile range 0.88 to 0.90) (Ormel et al., 2008). Previous methodo-

logical studies have also documented good discrimination between the role

functioning of cases and controls based on SDS scores in studies of a

The burden of mental disorders worldwide 17

number of disorders (Connor & Davidson, 2001; Hambrick, Turk,

Heimberg, Schneier, & Liebowitz, 2004; Leon et al., 1997; Pallanti,

Bernardi, & Quercioli, 2006).

Third, the CIDI assesses days out of role in the 30 days before interview,

making possible statistical analysis to determine which of the many mental

and physical disorders assessed in the surveys are most strongly related to

this important measure of role functioning. Fourth, WMH respondents

were asked to describe their own overall physical and mental health during

the past 30 days using a 0 to 100 visual analogue scale (VAS), where 0 rep-

resents “the worst possible health a person can have” and 100 represents

“perfect health.” Respondents were asked to make these global health valu-

ations near the end of their interview, taking into consideration all the phys-

ical and mental conditions reviewed in the survey.

Days out of role and health valuations are important outcomes not only

in substantive terms but also because, unlike the previous two measures

(i.e., disorder severity and disorder- specific SDS scores), they are not

disorder- specific measures. Instead, they are general measures of overall

functioning that allow us to make even- handed comparisons of the extent

to which specific disorders are independent predictors of these outcomes.

This also makes possible the study of the effects of comorbidity, which is

quite an important issue in light of the fact that many chronic- recurrent

physical and mental disorders are highly comorbid (Merikangas et al.,

2007). Importantly, comorbidity leads to overestimation of the burden of

individual disorders in analyses that fail to adjust for it (Alonso et al.,

2010), resulting in differential overestimation of the effects of disorders

based on differences in their patterns of comorbidity.

A fifth measure, earnings in the 12 months before the interview, was also

used as an outcome to evaluate the effects of mental disorders. Previous

studies in the U.S. have documented strong associations of mental disorders

with decrements in earnings (Harwood et al., 2000; Kessler et al., 2008;

Rice, Kelman, Miller, & Dunmeyer, 1990; Rice & Miller, 1998), but we are

aware of no previous cross- national study that examined this association.

Disorder prevalence estimates in the WMH surveys

The WMH surveys show clearly that mental disorders are common in all

the countries studied. The proportion of respondents estimated to have any

DSM- IV/CIDI disorder in the 12 months before interview averages (mean)

16.7% across surveys, with a median of 13.6% (see Table 2.2). The

highest prevalence is 29.6% in São Paulo and the lowest is 6.0% in

Nigeria. The inter- quartile range (IQR, 25th to 75th %iles) of prevalence

estimates across surveys was 10 to 20.7%. Relative prevalence estimates

are quite consistent across surveys, with anxiety disorders the most

common disorders in 22 of 24 countries. The two exceptions are Israel and

Ukraine, where mood disorders are estimated to be the most common

Table 2.2

Twelve-month prevalence of DSM-IV/CIDI disorders

a

in the World Mental Health surveys

a

Any disorder

b

Anxiety disorders

b

Mood disorders

b

Impulse-control disorders

b,c

Substance disorders

b

%

(se)

%

(se)

%

(se)

%

(se)

%

(se)

I. Low/lower-middle income countries Colombia

21.0

(1.0)

14.4

(1.0)

6.9

(0.4)

4.4

(0.4)

2.8

(0.4)

India – Pondicherry

20.0

(1.1)

10.5

(0.8)

5.5

(0.5)

4.3

(0.7)

5.3

(0.6)

Iraq

13.6

(0.8)

10.4

(0.7)

4.1

(0.4)

1.7

(0.3)

0.3

(0.1)

Nigeria

6.0

(0.6)

4.2

(0.5)

1.2

(0.2)

0.1

(0.0)

0.9

(0.2)

PRC – Beijing, Shanghai

7.1

(0.9)

3.0

(0.5)

2.2

(0.4)

2.7

(0.6)

1.6

(0.4)

PRC – Shenzhen

16.0

(0.9)

11.4

(0.9)

4.8

(0.4)

2.9

(0.3)

0.0

(0.0)

Ukraine

21.4

(1.3)

6.8

(0.7)

10.0

(0.8)

5.1

(0.8)

6.4

(0.8)

Total

14.8

(0.4)

9.2

0.3

4.8

(0.2)

2.7

(0.2)

1.9

(0.1)

II. Upper-middle income countries Brazil – São Paulo

29.6

(1.0)

19.9

(0.8)

11.8

(0.7)

5.3

(0.7)

3.8

(0.4)

Bulgaria

11.2

(0.8)

7.6

(0.7)

3.2

(0.3)

0.8

(0.3)

1.2

(0.3)

Lebanon

17.9

(1.6)

12.1

(1.2)

7.0

(0.8)

2.6

(0.7)

1.3

(0.8)

Mexico

13.4

(0.9)

8.4

(0.6)

5.0

(0.4)

1.6

(0.3)

2.5

(0.4)

Romania

8.2

(0.7)

4.9

(0.5)

2.5

(0.3)

1.9

(0.7)

1.0

(0.2)

South Africa

16.9

(0.9)

8.4

(0.6)

4.9

(0.4)

1.9

(0.3)

5.7

(0.6)

Total

16.7

(0.4)

10.2

(0.3)

5.8

(0.2)

2.5

(0.2)

3.2

(0.2)

III. High income countries Belgium

13.2

(1.5)

8.4

(1.4)

6.1

(0.8)

1.7

(1.0)

1.3

(0.4)

France

18.9

(1.4)

13.7

(1.1)

6.8

(0.7)

2.4

(0.6)

0.8

(0.3)

Germany

11.0

(1.3)

8.3

(1.1)

3.4

(0.3)

0.6

(0.3)

1.2

(0.4)

Israel

10.0

(0.5)

3.6

(0.3)

6.4

(0.4)

0.0

(0.0)

1.3

(0.2)

Italy

8.8

(0.7)

6.5

(0.6)

3.6

(0.3)

0.4

(0.2)

0.1

(0.1)

Japan

8.0

(0.7)

4.8

(0.6)

2.8

(0.4)

0.2

(0.1)

1.0

(0.3)

Netherlands

13.6

(1.0)

8.9

(1.0)

5.5

(0.7)

1.9

(0.7)

1.7

(0.5)

New Zealand

20.7

(0.6)

15.0

(0.5)

8.0

(0.4)

0.0

(0.0)

3.4

(0.3)

Northern Ireland

23.1

(1.4)

14.6

(1.0)

10.6

(0.9)

4.5

(1.0)

3.5

(0.5)

Portugal

22.9

(1.0)

16.5

(1.0)

8.3

(0.6)

3.5

(0.4)

1.6

(0.3)

Spain

9.7

(0.8)

6.6

(0.9)

4.4

(0.4)

0.5

(0.2)

0.3

(0.2)

United States

27.0

(0.9)

19.0

(0.7)

9.8

(0.4)

10.5

(0.7)

3.8

(0.4)

Total

17.7

(0.3)

11.9

(0.2)

7.2

(0.2)

2.7

(0.2)

2.2

(0.1)

IV. Total

16.7

(0.2)

10.8

(0.2)

6.2

(0.1)

2.6

(0.1)

2.4

(0.1)

Notes a The disorders included anxiety disorders (generalized anxiety disorder, panic disorder, agoraphobia, specific phobia, social pho

bia, post-traumatic stress

disorder, and separation anxiety disorder), mood disorders (major depressive disorder, dysthymic disorder, bipolar disorder), impulse-control disorders (attention-deficit/hyperactivity disorder, oppositional-defiant disorder, conduct disorder, intermittent explosive disorder), and

substance disorders

(alcohol and drug abuse with or without dependence).

b Between-country differences in prevalence are significant both for any disorder (χ

2

24

= 1401.2,

p

< 0.001) and for each class of disorder

(χ

2

24

= 715.4–1099.9,

p

< 0.001).

c Prevalence of impulse-control disorders was estimated in the sub-sample of respondents who were 44 years of age or younger at the time of the

interview.

20 R. C. Kessler et al.

disorders. Mood disorders are the next most common class of disorders in

all but two other countries. The exceptions are South Africa, where sub-

stance use disorders are more common than mood disorders, and Beijing-

Shanghai and the U.S., where behavior disorders are more common than

mood disorders. The 12-month prevalence of having any disorder varied

significantly across countries (χ2

24 = 1,401.2, p < 0.001).

A number of recent literature reviews have presented detailed compara-

tive data on the estimated prevalence of individual mental disorders and

classes of disorder based on published community epidemiologic surveys

(Somers et al., 2006; Waraich et al., 2004; Wittchen & Jacobi, 2005).

Several consistent patterns that emerge in these reviews are replicated in

the WMH data. First, anxiety disorders are consistently found to be the

most prevalent class of mental disorders in the general population, and

mood disorders are generally the next most prevalent class. Third, specific

phobia is the most prevalent individual mental disorder in the general

population (Silverman & Moreno, 2005). Major depression and social

phobia are generally found to be the next most prevalent disorders, again

consistent with WMH results. It is important to note that these relatively

high prevalence estimates are, if anything, conservative, as critics argue

that the diagnostic criteria in the DSM and ICD systems are overly con-

servative (e.g., Mylle & Maes, 2004; Ruscio et al., 2007). A related issue is

that considerable evidence exists for clinically significant subthreshold

manifestations of many mental disorders that are in some cases more pre-

valent than the disorders themselves (e.g., Brown & Barlow, 2005; Matsu-

naga & Seedat, 2007; Skeppar & Adolfsson, 2006).

Severity of mental disorders

While many previous epidemiologic surveys estimated disorder prevalence,

the WMH surveys are the first to generate systematic estimates of disorder

severity. Roughly one- quarter (24.5%) of all disorders are classified as a

serious mental illness (SMI) in the WMH surveys, using the definition of

SMI described in Table 2.3. The median proportion of cases with SMI

across surveys is 22.3%. The range is 6.2 to 36.9%, and the IQR is 18.6

to 25.8%. A higher proportion of all disorders is classified moderate (mean

37.8%, median 38.7%), with a range of 12.5 to 50.6% and an IQR of

32.7 to 42.9%. A roughly similar proportion of all disorders was classified

mild (mean and median both 37.7%), with a range of 28.3 to 74.8% and

IQR of 34.9 to 42.1%.

The severity distribution among cases varies significantly across coun-

tries (χ2

48 = 352.9, p < 0.001), although these differences are modest in sub-

stantive terms, with a Pearson contingency coefficient of 0.06 for the

association between country income level and disorder severity. There are

much more substantial positive associations (Pearson correlations) of

overall disorder prevalence with both the proportion of cases classified as

The burden of mental disorders worldwide 21

serious (0.30) and the proportion of cases classified as either serious or

moderate (0.40). The positive association found between estimated preva-

lence and severity across countries is potentially important because it

speaks to an issue raised in the methodologic literature regarding the possi-

bility of biased prevalence estimates.

Two separate research groups found an opposite sort of pattern than

the one found in the WMH surveys. The first was a study comparing

results from the Korean Epidemiologic Catchment Area (KECA) Study

(Chang et al., 2008) with results from parallel surveys in other countries.

The authors argued that the lower estimated prevalence of major depres-

sion in the KECA than in the other surveys was due, at least in part, to a

higher threshold for reporting depression among people in the Korean

population. In support of this assertion, the investigators showed that

Koreans diagnosed as depressed with an earlier version of the CIDI had

considerably higher levels of role impairment than respondents diagnosed

as depressed using the same instrument in the U.S.

The second relevant previous study was carried out as part of the WHO

Collaborative Study on Psychological Problems in General Health Care

(PPG) (Üstün & Sartorius, 1995). Nearly 26,000 primary care patients in

14 countries were assessed using an earlier version of the CIDI that

included an evaluation of current symptoms of depression. As in the WMH

surveys, substantial cross- national variation was found in the prevalence of

major depression. However, the investigators found that the average

amount of impairment associated with depression across countries was

inversely proportional to the estimated prevalence of depression in those

countries (Simon, Goldberg, Von Korff, & Üstün, 2002). This result is

consistent with the possibility that the substantial cross- national variation

in estimated prevalence of depression in the PPG study might be due, at

least in part, to cross- national differences in diagnostic thresholds.

However, we do not find results consistent with these in the WMH

surveys, where the countries with the lowest prevalence estimates of

DSM- IV/CIDI disorders also had the lowest reported levels of impairment

associated with those disorders.

Comparative role disabilities of mental and physical

disorders

As noted above, the WMH surveys assessed disorder- specific disability

with the SDS. Respondents with any of the mental and physical disorders

assessed in the surveys were asked to report the extent to which each such

disorder interfered with their ability to carry out their daily activities in

both productive roles (i.e., job, school, housework) and social roles (i.e.,

social and personal life). Rates of reported disability for the 10 most

common physical disorders and for the 10 most common mental disorders

are compared in each survey (Ormel et al., 2008).

Table 2.3

Twelve-month prevalence of DSM-IV/CIDI disorders by severity in the WMH surveys

a

Unconditional prevalence

b

Conditional prevalence

c

Disorders

Disorders

Serious

Moderate

Mild

Serious

Moderate

Mild

%

(se)

%

(se)

%

(se)

%

(se)

%

(se)

%

(se)

I. Low/lower-middle income countries Colombia

4.9

(0.5)

8.6

(0.7)

7.5

(0.5)

23.3

(2.1)

41.2

(2.6)

35.5

(2.1)

India – Pondicherry

4.3

(0.3)

7.8

(0.7)

7.9

(0.8)

21.7

(1.6)

39.0

(3.1)

39.3

(2.8)

Iraq

3.0

(0.4)

4.9

(0.4)

5.7

(0.6)

21.9

(2.3)

36.0

(2.6)

42.1

(2.9)

Nigeria

0.8

(0.3)

0.8

(0.2)

4.5

(0.5)

12.8

(3.8)

12.5

(2.6)

74.8

(4.2)

PRC – Beijing, Shanghai

1.0

(0.3)

2.3

(0.5)

3.8

(0.6)

13.8

(3.7)

32.2

(4.9)

54.0

(4.6)

PRC – Shenzhen

1.0

(0.3)

5.2

(0.5)

9.8

(0.8)

6.2

(1.6)

32.7

(2.9)

61.2

(3.6)

Ukraine

4.9

(0.4)

8.4

(0.8)

8.1

(1.0)

22.9

(1.8)

39.4

(2.9)

37.7

(3.5)

Total

2.8

(0.2)

5.3

(0.2)

6.7

(0.3)

18.8

(0.9)

35.9

(1.2)

45.3

(1.3)

II. Upper-middle income countries Brazil – São Paulo

10.0

(0.6)

9.8

(0.5)

9.8

(0.6)

33.9

(1.4)

33.0

(1.8)

33.2

(1.4)

Bulgaria

2.3

(0.3)

3.6

(0.5)

5.4

(0.5)

20.3

(2.8)

32.1

(3.6)

47.7

(2.7)

Lebanon

4.0

(0.7)

7.7

(1.0)

6.2

(1.2)

22.3

(3.1)

42.9

(4.9)

34.9

(5.6)

Mexico

3.5

(0.4)

4.7

(0.4)

5.2

(0.5)

26.3

(2.4)

34.8

(2.2)

38.9

(2.5)

Romania

2.3

(0.4)

2.4

(0.3)

3.5

(0.5)

27.9

(3.4)

29.3

(3.7)

42.8

(3.5)

South Africa

4.3

(0.4)

5.3

(0.5)

7.2

(0.6)

25.7

(1.8)

31.4

(2.1)

43.0

(2.1)

Total

4.7

(0.2)

5.5

(0.2)

6.5

(0.3)

28.0

(0.9)

33.1

(1.1)

39.0

(1.0)

III. High income countries Belgium

4.3

(0.8)

5.1

(0.8)

3.8

(0.6)

32.6

(4.2)

38.7

(3.4)

28.8

(4.8)

France

3.5

(0.5)

8.1

(0.8)

7.2

(0.9)

18.6

(2.5)

43.1

(3.0)

38.3

(3.6)

Germany

2.4

(0.4)

4.8

(0.8)

3.9

(0.7)

21.6

(2.5)

43.2

(4.5)

35.2

(4.1)

Israel

3.7

(0.3)

3.5

(0.3)

2.8

(0.2)

36.9

(2.4)

34.8

(2.3)

28.3

(2.1)

Italy

1.4

(0.2)

4.2

(0.5)

3.2

(0.5)

15.9

(2.7)

47.8

(3.9)

36.3

(3.9)

Japan

1.3

(0.4)

3.8

(0.5)

2.9

(0.4)

16.1

(4.5)

47.2

(4.8)

36.7

(3.7)

Netherlands

4.2

(0.6)

4.2

(0.5)

5.2

(0.8)

31.1

(3.5)

31.1

(3.6)

37.8

(4.7)

New Zealand

5.3

(0.3)

8.6

(0.4)

6.7

(0.3)

25.8

(1.0)

41.7

(1.4)

32.5

(1.2)

Northern Ireland

6.7

(0.7)

7.7

(0.7)

8.7

(1.1)

28.8

(3.0)

33.4

(2.6)

37.8

(3.3)

Portugal

4.0

(0.4)

11.6

(0.6)

7.3

(0.5)

17.5

(1.5)

50.6

(2.0)

31.9

(1.9)

Spain

1.9

(0.2)

4.2

(0.5)

3.6

(0.6)

19.9

(2.4)

43.5

(4.1)

36.6

(4.8)

United States

6.9

(0.4)

10.7

(0.5)

9.4

(0.6)

25.5

(1.4)

39.7

(1.2)

34.8

(1.4)

Total

4.5

(0.1)

7.2

(0.2)

6.0

(0.2)

25.4

(0.6)

40.7

(0.7)

33.9

(0.7)

IV. Total

4.1

(0.1)

6.3

(0.1)

6.3

(0.1)

24.5

(0.5)

37.8

(0.5)

37.7

(0.6)

Notes a See Table 3 footnote 1 for a list of the disorders. A respondent was defined as having a serious disorder if he either met criteria for bipolar I disorder or

substance dependence with a physiological dependence syndrome; made a suicide attempt in conjunction with any DSM-IV/CIDI disorder; reported at least two areas of role functioning with severe impairment due to a mental disorder on the Sheehan Disability Scales (SDS; Leon

et al., 1997); or

reported an overall level of functional impairment at a level consistent with a Global Assessment of Functioning (Endicott et al., 1976) score of 50 or less in conjunction with any DSM-IV/CIDI disorder. Respondents not classified as having a serious disorder were classified modera

te if they had a SDS

score rated at least moderate or met criteria for substance dependence without a physiological dependence syndrome. All other respondents with DSM-IV/CIDI disorders were classified mild.

b Unconditional prevalence is prevalence in the total sample. Conditional prevalence is prevalence among cases. For example, the 4.9 percent of respond-

ents in the Colombia survey with a 12-month serious disorder represent 23.3 percent of the 21.0 percent of respondents in the C

olombia who had any

12-month DSM-IV/CIDI disorder. (The 21.0 percent total prevalence is reported in Table 2.)

c Between-country differences in prevalence are significant both for unconditional prevalence for each class of disorders (χ

2

24

=

377.9 – 741.6,

p

< 0.001)

and for conditional prevalence for each class of disorders (χ

2

24

=

146.8 – 187.2,

p

< 0.001).

24 R. C. Kessler et al.

Of the 100 logically possible pair- wise, disorder- specific mental- physical

comparisons, the proportion of disability ratings in the severe range is

higher for the mental than physical disorder in 76 comparisons in high-

income countries and 84 comparisons in lower- middle-income countries

(Table 2.4). Nearly all of these higher mental than physical disability

ratings are statistically significant at the 0.05 level and hold in within-

person comparisons (i.e., comparing the reported disabilities associated

with a particular mental- physical disorder pair in the subsample of

respondents who had both disorders). A similar pattern is found when

treated physical disorders are compared with all (i.e., treated or not)

mental disorders to address the concern that the more superficial

Table 2.4 Disorder-specific global Sheehan Disability Scale ratings for commonly

occurring mental and chronic physical disorders in high-income and

lower-middle-income World Mental Health Countriesa

Proportion rated severely disabling

High-income

Lower-middle income

%

(se)

%

(se)

I. Physical disorders

Arthritis

23.3

(1.5)

22.8

(3.0)

Asthma

8.2*

(1.4)

26.9

(5.4)

Back/neck

34.6*

(1.5)

22.7

(1.8)

Cancer

16.6

(3.2)

23.9

(10.3)

Chronic pain

40.9*

(3.6)

24.8

(3.8)

Diabetes

13.6

(3.4)

23.7

(6.1)

Headaches

42.1*

(1.9)

28.1

(2.1)

Heart disease

26.5

(3.9)

27.8

(5.2)

High blood

pressure

5.3*

(0.9)

23.8

(2.6)

Ulcer

15.3

(3.9)

18.3

(3.6)

II. Mental disorders

ADHD

37.6

(3.6)

24.3

(7.4)

Bipolar

68.3*

(2.6)

52.1

(4.9)

Depression

65.8*

(1.6)

52.0

(1.8)

GAD

56.3*

(1.9)

42.0

(4.2)

IED

36.3

(2.8)

27.8

(3.6)

ODD

34.2

(6.0)

41.3

(10.3)

Panic disorde

48.4*

(2.6)

38.8

(4.7)

PTSD

54.8*

(2.8)

41.2

(7.3)

Social phobia

35.1

(1.4)

41.4

(3.6)

Specific phobia 18.6

(1.1)

16.2

(1.6)

Notes

* Significant difference between high income and lower-middle income countries based on

0.05-level two-sided tests

a See the text for a description of the Sheehan Disability Scale and the definition of severe

disability.

The burden of mental disorders worldwide 25

assessment of physical than mental disorders might have led to the inclu-

sion of subthreshold cases of physical disorders with low disability.

Days out of role associated with mental disorders

It is noteworthy that the disorder severity classification used in the

WMH surveys is validated by a consistently monotonic association

between reported disorder severity and mean number of days out of role

associated with the disorders. This association is statistically significant

in all but four surveys (Table 2.5). Respondents with serious disorders in

most surveys reported at least 40 days in the past year when they were

totally unable to carry out usual activities because of these disorders

(IQR: 56.7 to 135.9 days). The mean days out of role for mild disorders,

in comparison, is in the range 11.7 to 68.9 days, while the mean for

moderate disorders is intermediate between these extremes (21.1 to

109.4 days; IQR: 39.3 to 65.3 days). When we compared between-

country differences in these means with between- country differences in

prevalence, using the same logic as in the previous section, we once again

found a positive association between prevalence and the indicator of

severity. For example, in the three WMH countries with the highest

estimated overall 12-month prevalence of DSM- IV/CIDI disorders (US,

Ukraine, New Zealand), the mean number of days out of role associated

with disorders classified “severe” is in the range 98.1 to 142.5, compared

to means in the range 48.7 to 56.7 in the three countries with the lowest

12-month prevalence estimates (Japan, Nigeria, People’s Republic of

China).

Another possibility is that we underestimated prevalence in some coun-

tries because the DSM- IV categories are less relevant to symptom expres-

sion in those countries. We did not investigate this in the WMH surveys,

but a sophisticated analysis of this possibility was carried out as part of the

PPG (Üstün & Sartorius, 1995). In that study, an analysis of cross- national

variation in the structure of depressive symptoms was carried out using

item response theory (IRT) methods (Simon et al., 2002). The results

showed clearly that both the latent structure of depressive symptoms and

the associations between specific depressive symptoms and this latent struc-

ture were very similar across the countries studied. These results do not

support the suggestion that the large cross- national variation in estimated

prevalence of depression is due to cross- national differences in the nature

of depression. Comparable psychometric analyses have not yet been com-

pleted for other disorders, however, so it remains possible that cross-

national differences exist in latent structure that might play a part in

explaining the substantial differences in 12-month prevalence documented

in the WMH surveys.

At the same time, it is noteworthy that the countries with the lowest

disorder prevalence estimates in the WMH series also have the highest

26 R. C. Kessler et al.

proportions of treated cases classified as “subthreshold”; i.e., not meeting

criteria for any of the DSM- IV/CIDI disorders assessed in the WMH

interview. This finding at least indirectly raises the possibility that the

assessments in the CIDI are less adequate in capturing the

psychopathological syndromes that are common in all the WMH coun-

tries. In particular, the syndromes associated with treatment in low-

prevalence countries are not well characterized by the CIDI. Additional

WMH clinical reappraisal studies using flexible and culturally sensitive

assessments of psychopathology are currently underway in both high-

income and lower- middle-income countries.

These results involve individual- level effects; however, also instructive

are examinations of societal- level effects, by which we mean effects that

Table 2.5 Association between severity of 12-month DSM-IV/CIDI disorders and

days out of role in the World Mental Health surveys

Serious

Moderate

Mild

Wald Fa

Mean (se)

█

Mean (se)

█

Mean (se)

█

I. WHO Region: Pan American Health Organization (PAHO)

Colombia

53.0

(8.9)

33.7

(6.7) 15.6

(3.0)

10.8*

Mexico

42.8

(6.9)

26.3

(5.3) 11.7

(2.7)

11.7*

United States

135.9

(6.9)

65.3

(4.6) 35.7

(2.7) 126.1*

II. WHO Region: African Regional Office (AFRO)

Nigeria

56.7 (22.3)

51.5 (18.8) 25.9

(7.4)

1.6

South Africa

73.1

(9.7)

49.3

(6.5) 32.5

(4.8)

9.1*

III. WHO Region: Eastern Mediterranean Regional Office (EMRO)

Lebanon

81.4 (10.6)

42.0

(9.5) 13.6

(5.4)

14.4*

IV. WHO Region: European Regional Office (EURO)

Belgium

96.1 (26.0)

59.9 (11.6) 42.5

(9.6)

3.7*

France

105.7 (14.3)

71.8 (16.5) 67.6

(17.3)

2.7

Germany

77.8 (18.1)

33.2

(8.2) 45.7

(12.1)

2.2

Israel

184.6 (12.5) 109.4 (10.1) 44.6

(9.1)

41.8*

Italy

178.5 (25.6)

55.6 (10.9) 41.7

(11.2)

11.7*

Netherlands

140.7 (19.9)

87.1 (17.1) 68.9

(22.7)

4.0*

Spain

131.5 (15.8)

56.6 (10.0) 57.4

(22.0)

8.1*

Ukraine

142.5 (14.5) 103.2

(9.2) 51.6

(9.9)

13.9*

V. WHO Region: Western Pacific Regional Office (WPRO)

People’s Republic of

China

48.7 (18.4)

21.1

(5.2) 21.3

(7.2)

1.5

Japan

51.0 (17.3)

39.3 (10.6) 22.5

(6.4)

3.7*

New Zealand

98.1

(5.9)

54.6

(3.4) 36.4

(3.6)

40.7*

Notes

* Significant association between severity and days out of role at the 0.05 level.

a Results are based on simple mean comparisons. No control variables were used in the

analysis.

The burden of mental disorders worldwide 27

take into consideration not only relative impairment but also relative prev-

alence of different disorders. We are only beginning to do this in the cross-

national WMH data, but results of this sort have been generated for the

U.S. WMH survey (Merikangas et al., 2007). That analysis estimated that

fully one- third of all the days out of role associated with chronic- recurrent

health problems in the US population are due to mental disorders. This

amounts to literally billions of days of lost functioning per year in the U.S.

population. We do not yet know if comparable results will be obtained in

parallel analyses of WMH surveys in other countries, but preliminary

results suggest that this is likely to be the case.

Comparative health valuations of mental and physical disorders

The WMH analysis of health valuations was similar to the analysis of SDS

in that both compared mental disorders to physical disorders. However, a

somewhat different set of disorders was included in the two analyses – the

most commonly occurring disorders in the SDS analysis and the disorders

most strongly related to VAS scores in the health valuation analysis. The

health valuation analysis was also more textured than the SDS analysis in

that it allowed us to investigate the effects of comorbidity by examining

regression equations that predicted VAS scores from information about the

presence of individual disorders either considered alone or controlling for

comorbid disorders. This was not possible in the SDS analysis because SDS

scores were disorder- specific, whereas VAS scores were not.

Bivariate linear regression models (i.e., using only one disorder at a time

to predict VAS scores) were used to study predictive associations of disor-

ders with VAS scores. Every one of the disorders had a negative associ-

ation with VAS scores (i.e., associated with reduced perceived health) (see

Table 2.6). Coefficients based on the multivariate model that controlled for

comorbid disorders are consistently lower than those based on bivariate

models. The condition- specific ratio of the former to the latter is in the

range 0.24 to 0.70 with a median IQR of 0.42 (0.31 to 0.51). Very similar

results were found in lower- middle-income (0.53; 0.35 to 0.62) and high-

income (0.41; 0.27 to 0.51) countries (Alonso et al., 2010). The influence

of comorbidity can also be seen in the fact that the correlation across dis-

orders between mean number of comorbid disorders and the ratio of the

coefficient based on the bivariate model to the coefficient based on the

multivariate model is statistically significant (–0.46). More detailed ana-

lysis showed that coefficients are generally very similar in high- income and

lower- middle income- countries, with a Spearman rank- order correlation of

0.54. This is why results are presented here for high- income and lower-

middle-income countries combined.

At the bivariate level, the coefficients associated with most mental dis-

orders, in the range 7.3 to 17.8 in predicting scores on the 0 to 100 VAS

health valuation scale, are larger than those associated with all but two of

Table 2.6 Individual-level condition-specific estimates of predictive associations

between individual disorders and Visual Analogue Scale (VAS) health

evaluations based on bivariate and the best-fitting multivariate model in

the total samplea

Bivariate (Bi)b

Multivariate (Mul) Mul/Bic

Mean

comorbidityd

Est

(se)

█

Est

(se)

█

Est

█

I. Chronic physical conditions

Arthritis

–9.5* (0.5)

–4.9*

(0.4)

0.51

2.0

Cancer

–2.6* (1.1)

–0.8

(0.9)

0.31

2.1

Cardiovascular

disorders

–8.4* (0.4)

–4.9*

(0.4)

0.59

1.8

Chronic pain

conditions

–10.9* (0.4)

–6.8*

(0.4)

0.63

1.8

Diabetes

–8.8* (1.0)

–6.1*

(0.8)

0.70

2.0

Digestive disorders

–9.9* (0.9)

–4.1*

(0.8)

0.41

2.3

Headaches or

migraines

–9.9* (0.4)

–4.5*

(0.4)

0.45

2.0

Insomnia

–16.0* (0.7)

–7.9*

(0.7)

0.50

2.9

Neurological

disorders

–17.8* (1.7) –12.0*

(1.4)

0.67

2.6

Respiratory

disorders

–4.3* (0.4)

–1.4*

(0.4)

0.31

1.6

II. Mental conditions

Alcohol abuse

–7.3* (1.1)

–3.2*

(1.1)

0.44

1.8

Bipolar disorder

–17.8* (1.4)

–5.3*

(1.5)

0.30

3.9

Drug abuse

–12.4* (1.8)

–5.2*

(1.7)

0.42

2.6

Generalized anxiety

disorder

–13.4* (1.1)

–4.5*

(1.1)

0.34

3.0

Major depressive

episode

–14.8* (0.5)

–7.6*

(0.5)

0.52

2.5

Panic disorder

–16.6* (1.0)

–6.7*

(1.0)

0.40

3.4

Post–traumatic

stress disorder

–15.3* (1.1)

–4.7*

(0.9)

0.31

3.5

Social phobia

–11.2* (0.8)

–2.6*

(0.9)

0.24

2.9

Notes

* Significant at the 0.05 level, two-sided test.

a These estimates have a similar interpretation to metric regression coefficients, but differ

from the latter in that they are based on simulations generated from the coefficients in the

nonlinear/non-additive best-fitting multivariate mode. The latter model included controls

for both number and types of comorbid disorders. The model and the estimation procedure

are both described in detail elsewhere (Alonso et al., 2010).

b A separate model for each condition adjusted by socio-demographic controls.

c The ratio of the estimate based on the best-fitting multivariate model to the estimate based

on the bivariate model.

d Mean number of other conditions reported by respondents with the condition in the row.

The burden of mental disorders worldwide 29

the physical disorders (2.6 to 10.9). The two exceptions are insomnia

(16.0) and neurological disorders (17.8), which have coefficients as large

as those of the most severe mental disorders (panic disorder and bipolar

disorder, with coefficients of 16.6 to 17.8). It should be noted, though,

that insomnia and neurological disorders also have the highest overlap

with mental disorders of all the physical disorders considered. Indeed,

insomnia is a common symptom of numerous mental disorders and could

arguably have been classified as a mental disorder rather than as a physical

disorder. (The DSM- IV, in fact, includes a diagnosis of insomnia.) Some

neurological disorders, in comparison, are so closely related and difficult

to distinguish from certain mental disorders that they are often referred to

as neuropsychiatric disorders. Thus, it is noteworthy that with the excep-

tions of insomnia and neurological disorders, the bivariate coefficients

associated with all the physical disorders considered are consistently lower

than those associated with the bulk of the mental disorders. These results

are broadly consistent with those found in our analysis of the SDS data.

The situation is quite different, though, in the multivariate model, where

the coefficients associated with mental disorders (median 4.7; IQR 3.2 to

5.3) are quite comparable in magnitude to those associated with physical

disorders (median 4.9; IQR 4.1 to 6.1). This difference is due to the coeffi-

cients associated with mental disorders shrinking much more than those

associated with the physical disorders in the multivariate model compared to

the bivariate models. This greater shrinkage, in turn, is due to the higher

comorbidity of mental disorders (median number of comorbid disorders 2.9;

IQR 2.5 to 3.4) than physical disorders (median 2.0, IQR 1.8 to 2.3). When

we estimated only aggregate coefficients for having any (i.e., one or more)

mental disorder and any physical disorder, the coefficients (standard errors)

were significantly lower (t = 2.8, p = 0.004) for any mental disorder (7.4 (0.3))

than for any physical disorder (8.6 (0.3)). The fact that the effect of mental

disorders is lower than the effect of physical disorders in the aggregate mul-

tivariate analysis, whereas the effects of individual mental disorders are gen-

erally higher than those of individual physical disorders in bivariate analyses

means that the indirect effects of mental disorders due to comorbidity are to

some meaningful extent mediated by physical comorbidities (e.g., comorbid-

ity of mental disorders with insomnia and neurological disorders).

The long- term adverse effects of mental disorders

All of the associations described above dealt with short- term effects of current

mental disorders on various aspects of current functioning or on current per-

ceived health. Mental disorders are also known to have long- term effects.

Commonly occurring mental disorders have much earlier age- of-onset (AOO)

distributions than most chronic physical disorders (Kessler et al., 2007).

WMH respondents with a lifetime history of each disorder were asked to

report retrospectively how old they were when the disorder first began. AOO

30 R. C. Kessler et al.

distributions were generated from these reports and are very consistent across

countries (Kessler et al., 2007). Some anxiety disorders, most notably the

phobias and separation anxiety disorder (SAD), had very early AOO distribu-

tions, with median AOO in the range of seven to 14 and the vast majority of

lifetime cases occurring within five to 10 years of these medians. Similarly,

early onsets were typical for the externalizing disorders considered in the

WMH surveys. In comparison, the other common anxiety disorders (panic

disorder, generalized anxiety disorder, and posttraumatic stress disorder) and

mood disorders have later AOO distributions, with median AOO in the age

range 25 to 50 and a wide IQR (15 to 75). Substance use disorders have

intermediate median AOO (20 to 35), with the vast majority of cases having

onsets within 10 years of these medians.

WMH analyses show that early- onset mental disorders are significant

predictors of the subsequent onset and persistence of a wide range of phys-

ical disorders (He et al., 2008; Ormel et al., 2007). This is part of a larger

pattern of associations between early- onset mental disorders and a wide

array of adverse life course outcomes that might be conceptualized as soci-

etal costs of these disorders, including reduced educational attainment,

early marriage, marital instability, and low occupational and financial

status (Kessler et al., 1997; Kessler, Foster, Saunders, & Stang, 1995;

Kessler, Walters, & Forthofer, 1998; Lee et al., 2009). It is unclear if these

associations are causal, that is, if interventions to treat early- onset mental

disorders would prevent the subsequent onset of the adverse outcomes

with which they are associated. From a public health perspective, carrying

out long- term interventions to evaluate this issue would be valuable. Even

in the absence of this evidence, though, the available data from the WMH

surveys show that mental disorders, and especially early- onset mental dis-

orders, are associated with substantially reduced life changes in terms of

physical health and achievements in a variety of role domains.

The workplace costs of mental disorders to employers

A large part of the societal burden of mental disorders is the costs of

mental disorders due to reduced rates of labor force participation (Zhang,

Zhao, & Harris, 2009), elevated rates of unemployment among people in

the labor force (Chatterji, Alegria, Lu, & Takeuchi, 2007), and under-

employment among those who are employed (Kessler et al., 2008). All

of these associations are documented in the WMH surveys. It is important

to note that mental disorders also have costs to employers, including

high rates of sporadic absenteeism (Kessler & Frank, 1997) and

disability- related work leave (Kessler et al., 1999), as well as low levels of

on- the-job work performance (Berndt et al., 1998). The most commonly

used approach to study these labor market costs is the human capital

approach (Tarricone, 2006), which is based on the observation that wages

and salaries are paid in direct return for productive services. This makes

The burden of mental disorders worldwide 31

earnings a good indicator of the human capital accumulated by the indi-

vidual and earnings- equivalent time forgone because of an illness a good

representation of the indirect costs of that illness to the employer.

Although a considerable body of empirical research has used the human

capital approach to document the adverse societal effects of mental dis orders,

this research has been carried out largely in a small number of high- income

countries (Chatterji et al., 2007; Kessler et al., 1999). However, the data on

prevalence of mental disorders presented earlier in this chapter make it clear

that mental disorders are common throughout the world. Based on this

observation, we used the WMH data to estimate the human capital costs of

specific disorders on role functioning in workplace settings (de Graaf et al.,

2008; Kessler et al., 2006). The results are striking. In the U.S. WMH survey,

for example, 6.4% of workers had an episode of major depressive disorder

(MDD) in the year of the survey, resulting in an average of over five weeks of

lost work productivity (Kessler et al., 2006). Given the salaries of these

workers, the annual human capital loss to employers in the U.S. labor force

associated with MDD was estimated to be in excess of $36 billion. A similar

result was found in a WMH analysis that estimated the workplace costs of

adult attention deficit hyperactivity disorder (ADHD) in 10 WMH surveys

(de Graaf et al., 2008). ADHD was associated with an average of 22 days

excess lost productivity per worker with this disorder across the ten WMH

countries that assessed it. Comparable analyses for other disorders are in

progress in ongoing WMH investigations.

The effects of mental disorders on earnings

We also estimated the effects of mental disorders on earnings. We focused on

serious mental illness (SMI) because previous research has shown that earn-

ings and long- term work incapacity are both much more strongly related to

SMI than to less serious forms of mental illness (Kessler et al., 2008; Shiels,

Gabbay, & Ford, 2004). WMH respondents were asked to report their per-

sonal earnings in the 12 months before interview. In order to facilitate

pooling of results across countries, earnings reports were divided by the

median earnings in the country. These transformed scores were then used as

outcomes in a pooled regression model estimated simultaneously across all

countries. The regression analysis used a dummy variable for SMI as the pre-

dictor of primary interest. The outcome was the continuous earnings score

(appropriately transformed to address the problem of a highly skewed distri-

bution in most countries). Control variables included sociodemographics

(age, sex), country, substance disorders, and interactions between sex and all

other predictors. The sex interactions were included because previous

research has shown that the predictors of earnings differ betweeen males and

females (Kessler et al., 2008; Rice & Miller, 1998).

SMI was associated with an enormous reduction in earnings: 32% of

the median within- country earnings in high- income countries and 33% of

32 R. C. Kessler et al.

median within- country earnings in lower- middle-income countries (Levin-

son et al., 2010). Decomposition showed that 39% of this total association

in high- income countries and 27% in lower- middle-income countries was

due to the reduced probability of having any earnings among people with

SMI. That is, people with SMI were significantly less likely to be employed.

A larger component of the total association, 49% of the total in high-

income countries and 66% in lower- middle-income countries, was due to

lower mean level of earnings among people with SMI who had any earn-

ings. That is, employed people with SMI were found to have significantly

lower earnings than other employed people. Further analysis showed that

part of this effect was due to people with SMI having lower education than

other people, but a significant residual association still existed between

SMI and low earnings even after controlling for education.

The cost- effectiveness of treatment

Costs as large as those documented above raise the question of whether

expansion of detection and treatment efforts and increases in treatment

quality improvement initiatives might reduce the adverse effects of mental

disorders to an extent that makes treatment cost- effective either from a soci-

etal an employer perspective. An effectiveness trial carried out in conjunction

with the WMH survey in the U.S. evaluated this question experimentally

(Wang, Simon et al., 2007). A large sample of workers was screened for

major depressive disorder (MDD) and randomized to either a model outreach

and best- practices treatment intervention or to usual care. The intervention

group at six and 12 months had significantly higher job retention than con-

trols, as well as significantly more hours worked than controls (equivalent to

an annual two weeks more work). The financial benefits of these intervention

effects (in terms of hiring and training costs, disability payment, and salaries

paid for sickness absence days) were substantially higher than the costs of

treatment, demonstrating that an expansion of workplace screening, detec-

tion, and treatment of worker mental disorders could represent a human

capital investment opportunity for employers. Replications of this study are

currently underway in other WMH countries. Extensions of the intervention

to consider treatment of bipolar depression and adult ADHD are also under-

way. Ongoing analyses of the WMH data are also searching for other inter-

vention targets that can be used to evaluate the effects of treatment in

reducing the burdens associated with mental disorders.

Conclusions

The data reviewed in this chapter illustrate that mental disorders are

commonly occurring in the general population, often have an early

age- of-onset, and often are associated with significant adverse societal costs.

We also presented some evidence that at least part of this burden can be

The burden of mental disorders worldwide 33

reversed with best- practices treatment. The latter finding argues much more

persuasively than the naturalistic survey findings that mental dis orders are

actual causes rather than merely correlates of impaired role functioning.

Based on these results, we can safely conclude that mental disorders are

common and consequential from a societal perspective throughout the world.

However, as reported elsewhere, the WMH data show that only a small

minority of people with even seriously impairing mental disorders receive

treatment in most countries and that even fewer receive high- quality treat-

ment (Wang, Aguilar- Gaxiola et al., 2007). This situation has to change. A

good argument can be made based on the WMH results that an expansion of

treatment would be a human capital investment opportunity from the

employer’s perspective, and the same argument might be made from a societal

perspective. Ongoing WMH analyses are continuing to refine the naturalistic

analyses of the adverse effects of mental disorders in an effort to target exper-

imental interventions that can demonstrate the value of expanded treatment

in addressing the enormous global burden of mental disorders.

Acknowledgments

Preparation of this chapter was supported, in part, by the following grants

from the U.S. Public Health Service: U01MH060220, R01DA012058,

R01MH070884, R01DA016558. Portions of this paper appeared previously

in Kessler, R. C., Aguilar-Gaxiola, S., Alonso, J., et al. (2009). The global

burden of mental disorders: An update from the WHO World Mental Health

(WMH) Surveys. Epidemiologia E Psichiatria Sociale, 18(1), 23–33, © 2009

Cambridge University Press, and Alonso, J., Petukhova, M., Vilagut, G., et al.

(2010). Days out of role due to common physical and mental conditions:

Results from the WHO World Mental Health Surveys. Molecular Psychiatry,

doi:10.1038/mp.2010.101, © 2010 Nature Publishing Group. Both used

with permission. A complete list of WMH publications can be found at www.

hcp.med.harvard.edu/wmh/. The views and opinions expressed in this report

are those of the authors and should not be construed to represent the views

of any of the sponsoring organizations, agencies, or Governments.

Declaration of interest

Kessler has been a consultant for GlaxoSmithKline Inc., Pfizer Inc., Wyeth-

Ayerst, Sanofi- Aventis, Kaiser Permanente, and Shire Pharmaceuticals; has

served on advisory boards for Eli Lilly & Company and Wyeth- Ayerst;

and has had research support for his epidemiological studies from Eli Lilly

and Company, Pfizer Inc., Ortho- McNeil Pharmaceuticals Inc., Sanofi-

Aventis, Merck, Shire, and Bristol- Myers Squibb.

34 R. C. Kessler et al.

References

Alonso, J., Petukhova, M., Vilagut, G., Chatterji, S., Heeringa, S., Üstün, T. B., et al.

(2010). Days out of role due to common physical and mental conditions: Results

from the WHO World Mental Health Surveys. Molecular Psychiatry, doi:10.1038/

mp.2010.

Alonso, J., Vilagut, G., Chatterji, S., Heeringa, S., Schoenbaum, M., Üstün, T. B., et al.

(2010). Including information about comorbidity in estimates of disease burden:

Results from the WHO World Mental Health Surveys. Psychological Medicine.

doi:10.1017/S0033291710001212.

Berndt, E. R., Finkelstein, S. N., Greenberg, P. E., Howland, R. H., Keith, A., Rush, A.

J., et al. (1998). Workplace performance effects from chronic depression and its

treatment. Journal of Health Economics, 17(5), 511–535.

Berto, P., D’Ilario, D., Ruffo, P., Di Virgilio, R., & Rizzo, F. (2000). Depression: Cost-

of-illness studies in the international literature, a review. Journal of Mental Health

Policy and Economics, 3(1), 3–10.

Brown, T. A., & Barlow, D. H. (2005). Dimensional versus categorical classification

of mental disorders in the fifth edition of the Diagnostic and Statistical Manual of

Mental Disorders and beyond: Comment on the special section. Journal of Abnor-

mal Psychology, 114(4), 551–556.

Chang, S. M., Hahm, B. J., Lee, J. Y., Shin, M. S., Jeon, H. J., Hong, J. P., et al.

(2008). Cross- national difference in the prevalence of depression caused by the diag-

nostic threshold. Journal of Affective Disorders, 106(1–2), 159–167.

Chatterji, P., Alegria, M., Lu, M., & Takeuchi, D. (2007). Psychiatric disorders and

labor market outcomes: Evidence from the National Latino and Asian American

Study. Health Economics, 16(10), 1069–1090.

Connor, K. M., & Davidson, J. R. (2001). SPRINT: A brief global assessment of post-

traumatic stress disorder. International Clinical Psychopharmacology, 16(5),

279–284.

de Graaf, R., Kessler, R. C., Fayyad, J., ten Have, M., Alonso, J., Angermeyer, M., et

al. (2008). The prevalence and effects of adult attention- deficit/hyperactivity dis-

order (ADHD) on the performance of workers: Results from the WHO World

Mental Health Survey Initiative. Occupational and Environmental Medicine,

65(12), 835–842.

Druss, B. G., Hwang, I., Petukhova, M., Sampson, N. A., Wang, P. S., & Kessler, R.

C. (2008). Impairment in role functioning in mental and chronic medical disorders

in the United States: Results from the National Comorbidity Survey Replication.

Molecular Psychiatry, 14(7), 728–737.

Endicott, J., Spitzer, R. L., Fleiss, J. L., & Cohen, J. (1976). The global assessment

scale. A procedure for measuring overall severity of psychiatric disturbance.

Archives of General Psychiatry, 33(6), 766–771.

Greenberg, P. E., & Birnbaum, H. G. (2005). The economic burden of depression in

the US: Societal and patient perspectives. Expert Opinion on Pharmacotherapy,

6(3), 369–376.

Greenberg, P. E., Sisitsky, T., Kessler, R. C., Finkelstein, S. N., Berndt, E. R., David-

son, J. R., et al. (1999). The economic burden of anxiety disorders in the 1990s.

Journal of Clinical Psychiatry, 60(7), 427–435.

Hambrick, J. P., Turk, C. L., Heimberg, R. G., Schneier, F. R., & Liebowitz, M. R.

The burden of mental disorders worldwide 35

(2004). Psychometric properties of disability measures among patients with social

anxiety disorder. Journal of Anxiety Disorders, 18(6), 825–839.

Harkness, J., Pennell, B. E., Villar, A., Gebler, N., Aguilar- Gaxiola, S., & Bilgen, I.

(2008). Translation procedures and translation assessment in the World Mental

Health Survey Initiative. In R. C. Kessler & T. B. Üstün (Eds.), The WHO World

Mental Health Surveys: Global perspectives on the epidemiology of mental dis orders

(pp. 91–113). New York: Cambridge University Press.

Haro, J. M., Arbabzadeh- Bouchez, S., Brugha, T. S., de Girolamo, G., Guyer, M. E.,

Jin, R., et al. (2006). Concordance of the Composite International Diagnostic Inter-

view Version 3.0 (CIDI 3.0) with standardized clinical assessments in the WHO

World Mental Health surveys. International Journal of Methods in Psychiatric

Research, 15(4), 167–180.

Harwood, H., Ameen, A., Denmead, G., Englert, E., Fountain, D., & Livermore, G.

(2000). The economic cost of mental illness, 1992. Rockville, MD: National Insti-

tute of Mental Health.

He, Y., Zhang, M., Lin, E. H., Bruffaerts, R., Posada- Villa, J., Angermeyer, M. C., et

al. (2008). Mental disorders among persons with arthritis: Results from the World

Mental Health Surveys. Psychological Medicine, 38(11), 1639–1650.

Heeringa, S. G., Wells, J. E., Hubbard, F., Mneimneh, Z., Chiu, W. T., & Sampson,

N. (2008). Sample Designs and Sampling Procedures. In R. C. Kessler & T. B. Üstün

(Eds.), The WHO World Mental Health Surveys: Global perspectives on the

epidemiology of mental disorders (pp. 14–32). New York: Cambridge University

Press.

Kessler, R.C., Aguilar- Gaxiola, S., Alonso, J., Chatterji, S., Lee, S., Ormel, J., et al.

(2009). The global burden of mental disorders: An update from the WHO World

Mental Health (WMH) Surveys. Epidemiologia E Psichiatria Sociale, 18(1), 23–33.

Kessler, R. C., Akiskal, H. S., Ames, M., Birnbaum, H., Greenberg, P., Hirschfeld, R.

et al. (2006). Prevalence and effects of mood disorders on work performance in a

nationally representative sample of U.S. workers. American Journal of Psychiatry,

163(9), 1561–1568.

Kessler, R. C., Amminger, G. P., Aguilar- Gaxiola, S., Alonso, J., Lee, S., & Üstün, T.

B. (2007). Age of onset of mental disorders: A review of recent literature. Current

Opinion in Psychiatry, 20(4), 359–364.

Kessler, R. C., Barber, C., Birnbaum, H. G., Frank, R. G., Greenberg, P. E., Rose, R.

M., et al. (1999). Depression in the workplace: Effects on short- term disability.

Health Affairs (Millwood), 18(5), 163–171.

Kessler, R. C., Berglund, P. A., Foster, C. L., Saunders, W. B., Stang, P. E., & Walters,

E. E. (1997). Social consequences of psychiatric disorders, II: Teenage parenthood.

American Journal of Psychiatry, 154(10), 1405–1411.

Kessler, R. C., Foster, C. L., Saunders, W. B., & Stang, P. E. (1995). Social

consequences of psychiatric disorders, I: Educational attainment. American Journal

of Psychiatry, 152(7), 1026–1032.

Kessler, R. C., & Frank, R. G. (1997). The impact of psychiatric disorders on work

loss days. Psychological Medicine, 27(4), 861–873.

Kessler, R. C., Heeringa, S., Lakoma, M. D., Petukhova, M., Rupp, A. E., Schoen-

baum, M., et al. (2008). Individual and societal effects of mental disorders on earn-

ings in the United States: Results from the national comorbidity survey replication.

American Journal of Psychiatry, 165(6), 703–711.

Kessler, R. C., & Üstün, T. B. (2004). The World Mental Health (WMH) Survey

36 R. C. Kessler et al.

Initiative Version of the World Health Organization (WHO) Composite Interna-

tional Diagnostic Interview (CIDI). International Journal of Methods in Psychi-

atric Research, 13(2), 93–121.

Kessler, R. C., Walters, E. E., & Forthofer, M. S. (1998). The social consequences

of psychiatric disorders, III: Probability of marital stability. American Journal of

Psychiatry, 155(8), 1092–1096.

Lee, S., Tsang, A., Breslau, J., Aguilar- Gaxiola, S., Angermeyer, M., Borges, et al.

(2009). Mental disorders and termination of education in high- income and low-

and middle- income countries: Epidemiological study. British Journal of Psychia-

try, 194(5), 411–417.

Leon, A. C., Olfson, M., Portera, L., Farber, L., & Sheehan, D. V. (1997). Assess-

ing psychiatric impairment in primary care with the Sheehan Disability Scale.

International Journal of Psychiatry in Medicine, 27(2), 93–105.

Levinson, D., Lakoma, M., Petukhova, M., Schoenbaum, M., Zaslavsky, A. M.,

Angermeyer, M., et al. (2010). The associations of serious mental illness with

earnings in the WHO World Mental Health surveys. British Journal of Psychia-

try, 197, 114-121.

Lopez, A. D., & Mathers, C. D. (2007). Inequalities in health status: Findings from

the 2001 Global Burden of Disease study. In S. Matlin (Ed.), The global forum

update on research for health, volume 4 (pp. 163–175). London: Pro- Brook Pub-

lishing Limited.

Maetzel, A., & Li, L. (2002). The economic burden of low back pain: A review of

studies published between 1996 and 2001. Best Practice and Research Clinical

Rheumatology, 16(1), 23–30.

Matsunaga, H., & Seedat, S. (2007). Obsessive- compulsive spectrum disorders:

Cross- national and ethnic issues. CNS Spectrums, 12(5), 392–400.

Merikangas, K. R., Ames, M., Cui, L., Stang, P. E., Üstün, T. B., Von Korff, M., &

Kessler, R. C. (2007). The impact of comorbidity of mental and physical con-

ditions on role disability in the US adult household population. Archives of

General Psychiatry, 64(10), 1180–1188.

Murray, C. J. L., & Lopez, A. D. (1996). The Global Burden of Disease: A compre-

hensive assessment of mortality and disability from diseases, injuries and risk

factors in 1990 and projected to 2020. Cambridge, MA: Harvard University Press.

Murray, C. J. L., Lopez, A. D., Mathers, C. D., & Stein, C. (2001). The Global

Burden of Disease 2000 Project: Aims, methods and data sources. Geneva:

World Health Organization.

Mylle, J., & Maes, M. (2004). Partial posttraumatic stress disorder revisited.

Journal of Affective Disorders, 78(1), 37–48.

Ormel, J., Petukhova, M., Chatterji, S., Aguilar- Gaxiola, S., Alonso, J., Anger-

meyer, M. C., et al. (2008). Disability and treatment of specific mental and phys-

ical disorders across the world. British Journal of Psychiatry, 192(5), 368–375.

Ormel, J., Von Korff, M., Burger, H., Scott, K., Demyttenaere, K., Huang, Y. Q.,

et al. (2007). Mental disorders among persons with heart disease – results from

World Mental Health surveys. General Hospital Psychiatry, 29(4), 325–334.

Pallanti, S., Bernardi, S., & Quercioli, L. (2006). The Shorter PROMIS Question-

naire and the Internet Addiction Scale in the assessment of multiple addictions in

a high- school population: Prevalence and related disability. CNS Spectrums,

11(12), 966–974.

Pennell, B. E., Mneimneh, Z., Bowers, A., Chardoul, S., Wells, J. E., Viana, M. C.,

et al. (2008). Implementation of the World Mental Health Surveys. In R. C.

The burden of mental disorders worldwide 37

Kessler & T. B. Üstün (Eds.), The WHO World Mental Health Surveys: Global

perspectives on the epidemiology of mental disorders (pp. 33–57). New York:

Cambridge University Press.

Reed, S. D., Lee, T. A., & McCrory, D. C. (2004). The economic burden of allergic

rhinitis: A critical evaluation of the literature. Pharmacoeconomics, 22(6),

345–361.

Rice, D. P., Kelman, S., Miller, L. S., & Dunmeyer, S. (1990). The economic costs

of alcohol and drug abuse and mental illness: 1985. Washington, DC: US

Department of Health and Human Services.

Rice, D. P., & Miller, L. S. (1998). Health economics and cost implications of

anxiety and other mental disorders in the United States. British Journal of Psy-

chiatry, Supplement(34), 4–9.

Ruscio, A. M., Chiu, W. T., Roy- Byrne, P., Stang, P. E., Stein, D. J., Wittchen, H.

U., & Kessler, R. C. (2007). Broadening the definition of generalized anxiety dis-

order: Effects on prevalence and associations with other disorders in the National

Comorbidity Survey Replication. Journal of Anxiety Disorders, 21(5), 662–676.

Shiels, C., Gabbay, M. B., & Ford, F. M. (2004). Patient factors associated with

duration of certified sickness absence and transition to long- term incapacity.

British Journal of General Practice, 54(499), 86–91.

Silverman, W. K., & Moreno, J. (2005). Specific phobia. Child and Adolescent Psy-

chiatric Clinics of North America, 14(4), 819–843, ix–x.

Simon, G. E., Goldberg, D. P., Von Korff, M., & Ustun, T. B. (2002). Understand-

ing cross- national differences in depression prevalence. Psychological Medicine,

32(4), 585–594.

Skeppar, P., & Adolfsson, R. (2006). Bipolar II and the bipolar spectrum. Nordic

Journal of Psychiatry, 60(1), 7–26.

Somers, J. M., Goldner, E. M., Waraich, P., & Hsu, L. (2006). Prevalence and inci-

dence studies of anxiety disorders: A systematic review of the literature. Cana-

dian Journal of Psychiatry, 51(2), 100–113.

Tarricone, R. (2006). Cost- of-illness analysis. What room in health economics?

Health Policy, 77(1), 51–63.

Üstün, T. B., & Sartorius, N. (Eds.). (1995). Mental illness in general health care:

An international study. New York: Wiley.

Wang, P. S., Aguilar- Gaxiola, S., Alonso, J., Angermeyer, M. C., Borges, G.,

Bromet, E. J., et al. (2007). Use of mental health services for anxiety, mood, and

substance disorders in 17 countries in the WHO world mental health surveys.

Lancet, 370(9590), 841–850.

Wang, P. S., Simon, G. E., Avorn, J., Azocar, F., Ludman, E. J., McCulloch, J., et al.

(2007). Telephone screening, outreach, and care management for depressed workers

and impact on clinical and work productivity outcomes: A randomized controlled

trial. Journal of the American Medical Association, 298(12), 1401–1411.

Waraich, P., Goldner, E. M., Somers, J. M., & Hsu, L. (2004). Prevalence and inci-

dence studies of mood disorders: A systematic review of the literature. Canadian

Journal of Psychiatry, 49(2), 124–138.

Wittchen, H. U., & Jacobi, F. (2005). Size and burden of mental disorders in

Europe: A critical review and appraisal of 27 studies. European Neuropsycho-

pharmacology, 15(4), 357–376.

Zhang, X., Zhao, X., & Harris, A. (2009). Chronic diseases and labour force par-

ticipation in Australia. Journal of Health Economics, 28(1), 91–108.