1. health and fitness
2. disease
3. ibs and crohn's disease

Socioeconomic
Burden
and Major
Depression
Lewis
Kenneth
Objective:
f unction
economic
data
Area
B. Wells,
and
societal
gathered
Program
disorder
and
subjects
who
f inancial
bed
had
strain
days,
subjects
goal was
depressive
to evaluate
symptoms
(N=2,02
no disorder
poor
no
health
status.
strain,
limitations
in physical
self-ratings
of health
status,
syndromal
symptoms
and
those
depressive
levels
University
of California,
VA Medical
Sciences,
Center;
University
ration,
San
the
with
Diego;
the
of California,
symptoms,
depression.
combined
Service,
and
Angeles;
and
the
Santa
Monica,
Calif.
Address
reprint
requests
Department
of Psychiatry,
9500
(ilman
Dr.,
UCSD
San
Diego
Biohehavioral
Rand
to
La
Am
]
Psychiatry
I 53:
Research
Center
grant
Benjamin
of the Rand
the statistical
analysis.
1 1 , November
1 996
with
importance
Corpo-
Dr. Judd,
Jolla,
CA
92093-0603.
Supported
by NINtH
Clinical
and the Rohrer
Fund.
The authors
thank
Bernadette
for her assistance
in conducting
on I 0 domains
with
depressive
strain,
subjects
losses,
bed
Ml-1-30914
Corporation
and
of functional
symptoms
social
with
major
days,
high
than
irritability,
and
activity
days,
restricted
Conclusions:
the
prevalence
Catchment
Schedule,
the
into three groups:
subjects
(N= I 02), and no depres-
functioning,
or major
depression
with
no disorder.
1 1 and April
16,
of Psychiatry,
Psychiatry
of Psychiatry
Los
more
financial
major
symptoms
did subjects
received
Jan.
the Department
Department
or job
12-month
Epidemiologic
Interview
compared
subjects
ofhousehold
Significantly
nterest
in subthreshold
depression
has increased
since the report
of Wells et al. ( 1 ) that subthreshold
depressive
symptoms
in medical
and psychiatric
outpatients
were
associated
with
significant
levels
of psychosocial
dysfunction.
Epidemiologic
studies
have confirmed
that,
compared
with
no depressive
symptoms,
subthreshold
depressions,
variously
defined,
are associated
with
more
disability
and work
days
lost (1-3),
greaten
use of services
for mental
health
problems
(3),
1995;
revisR)ns
3, 1996.
From
were
more
I
Received
Juls- 12,
accepted
June
Using
ECA site
depression
emphasizes
the clinical
and public
health
iiito these symptoms.
(AmJ
Psychiatry
1996;
153:1411-1417)
1996;
Method:
the Los Angeles
(N=2 70), major
major
M.D.
depression
levels
or job functioning,
and poor
health
status.
no significant
differences
were found
between
with subsyndromal
depressive
I 0 functional
domains
than
subsyndromal
H. Rapaport,
(NIMH)
Diagnostic
in physical
reported
M.D.,
Health
NIMH
high
as limitations
disorder
conditions.
Significantly
reported
Mark
Symptoms
Population
the association
between
impairment
in daily
as well as major
depression
to determine
the
1). The groups
Results:
as well
and
with
of these
P. Paulus,
and
Institute
of Mental
on responses
to the
the 2,393
subjects
from
depressive
symptoms
well-being.
Martin
M.P.H.,
significance
or symptoms
outcome
M.D.,
M.D.,
by the National
(ECA),
based
authors
divided
with subsyndromal
sive
L. Judd,
The authors’
subsyndromal
and
of Subsyndromal
Depressive
in a Sample
of the General
Except
subjects
Significantly
than
of financial
for lower
with submore
people
reported
impairment
in eight
The high
1 -year
prevalence
associated
need
for
functional
additional
of
of
impairment,
investigations
poor self-ratings
of emotional
health
( 1 3, 4), and more
lifetime
suicide
attempts
(5).
To clarify
the clinical
significance
of subthreshold
depressive
symptoms,
we defined
a more
restrictive
condition,
subsyndromal
depressive
symptoms,
as at least
two on more current
depressive
symptoms,
present
for
most or all of the time,
lasting
for at least 2 weeks,
in
individuals
who did not meet criteria
for major
depres,
sion
on dysthymia.
Using
this
definition,
we
found
that
I 1.8%
of 9,160
community
residents
in the National
Institute
of Mental
Health
(NIMH)
Epidemiologic
Catchment
Area
Program
(ECA)
met criteria
for subsyndromal
depressive
symptoms
during
a 1-year
period
(6). Two-thirds
of the subjects
with subsyndromal
depressive
symptoms
were women;
the most common
depressive
symptoms
were insomnia,
fatigue,
and recurrent thoughts
of death.
Significantly
more subjects
with
subsyndromal
depressive
symptoms
than other
subjects
reported
lifetime
disability
benefits,
raising
the question
of whether
subsyndromal
depressive
symptoms
are associated
with
sufficient
disruption
in the domains
of
1411
SOCIOECONOMIC
BURDEN
OF DEPRESSION
everyday
function
to have demonstrable
public
health,
societal,
and economic
impact.
To address
this question,
we analyzed
the database
from the Los Angeles
site of the ECA,
where
the investigators
systematically
probed
domains
of daily
function not obtained
at other
ECA sites. Results
of respondents
with
subsyndromal
depressive
symptoms
were
compared
with
those
of respondents
who had no depressive
disorder
or symptoms.
Because
the severe
impairment
associated
with major
depression
has been established
unequivocally
(1-4,
7), subjects
with
major
depression
were included
for reference
comparison
of
disability.
Two a priori
hypotheses
were tested:
I ) Significantly
more subjects
with subsyndromal
depressive
symptoms
than subjects
with no depressive
disorder
or
symptoms
will report
functional
impairment.
2) Significandy
more
subjects
with
major
depression
than
subjects with subsyndromal
depressive
symptoms
or no depressive
disorder
or symptoms
will report
functional
impairment.
Militating
against
hypothesis
2 are empinical reports
that subjects
with subsyndromal
depressive
symptoms,
even though
significantly
less impaired,
are
relatively
comparable
in impairment
to subjects
with
major
depression
(1-4).
METHOD
chotomized
into
irritable
more
ciated
with
current
medical
illnesses
and
should
not
be side
effects
drug or alcohol
consumption
or problems;
symptoms
associated
these conditions
are excluded.
The Los Angeles
ECA interviews
included
questions
probing
domains
of functional
outcomes
and well-being.
Details
of the
tent and methodology
of these
measures
have
been
reported
where
(15-18)
but will he summarized
here.
Domains
of Daily
of
with
10
conelse-
Function
1. High
or low social
irritability
in previous
6 months.
Respondents were asked
how often over the last 6 months
they had felt upset,
irritated,
or angry
in six types
of social
relationships:
with
friends,
with spouse,
with other adults
in the household,
with family
members
and relatives
not living with the respondent,
with boss or supervisor
at work,
and with co-workers.
Response
categories
were
not at all,
less than once a month,
once a month,
a few times a month,
and once
a week or more.
Responses
were averaged
for each respondent
across
all applicable
categories
and then,
due to high
skewness,
were
di-
1412
irritable
once
less than
once
a month
or 2) being
a month.
2. High
or low
household
strain.
This
domain
of function
is
the employment
strain
scale
of Pearlin
and Lieberman
(19) but applied
to household
responsibilities
by Golding
(20).
Respondents
were
asked
to answer
yes or no to four items:
“I have
more
home
(and family)
duties
than
I am able to do,”
“1 spend
too
many
hours
taking
care
of my home
(and
family),”
“I do more
work
around
the house
than I should
have to do,” and “ Household
and family
responsibilities
keep
me from
doing
things
I would
really like to do. “ Dichotomous
scoring
recommended
by Golding
based
(20)
on
was
used,
in which
responses
were
scored
four items
was endorsed.
3. High or low social
contacts
tacts
(visits
or telephone
calls)
in previous
during
the
family
(but
as
high
when
any
of
the
members,
or other
in household)
reports
of
adults
not
nzonth.
Number
last month
with
including
any
of con-
friends,
persons
living
were summed.
Skewed
response
distribution
scoring:
a rating
of high was given to reports
in the last month,
and a rating
of low was
dichotomized
more
contacts
fewer
than
13
in the
last
required
of 13 or
given
to
month.
4. Major
financial
loss in previous
6 months.
This
domain
was
scored
present
if one or more
of the following
five items
occurred
during
the last 6 months:
job loss, someone
important
to the respondent was out of work
for 1 month
or more,
loss of home
or anything
else important,
or the financial
situation
of someone
the respondent
depended
on got much
worse.
5. High
or low financial
strain.
In this functional
domain,
developed
by Pearlin
and Lieberman
(19) and Golding
(20), respondents
were
asked
to answer
yes or no to five negative
or positive
items.
The negative
items
were
“It is difficult
for me to afford
the kind of
medical
care I should
have,”
“It is difficult
for me to pay all the
monthly
bills,”
and
“It is difficult
to afford
needed
furniture
or
household
appliances.”
The
positive
items
were
“I have
enough
money
The sampling
methodology,
human
subject
consent
procedures,
study design,
survey
methods,
and demographic
characteristics
of the
ECA sample
have been reported
elsewhere
(8-10).
ECA DSM-III
diagnoses
were
developed
from
responses
to the NIMH
Diagnostic
Interview
Schedule
(DIS), previously
described
in the literature
( I 1, 12).
A Spanish-language
version
of the DIS was also used at the Los Angeles
site ofthe
ECA (13, 14).
At three
ECA sites, including
Los Angeles,
data were collected
for
each
individual
mental
symptom
probed
by the DIS and the time
frames
when
symptoms
occurred
(i.e.,
1 month,
6 months,
1 year, or
lifetime).
The DIS specifies
a standardized
threshold
to determine
if a
symptom
is clinically
significant.
For example,
depressive
symptoms
must
have been present
for at least 2 weeks
and substantial
enough
that the respondent
had either
talked
to a health
professional
or had
taken
medications
because
of it or felt the symptom
had interfered
substantially
with his or her everyday
activities.
Depressive
symptoms
identified
by the DIS are clinically
substantial
in severity
and duration,
extending
well beyond
normal
mood
vicissitudes.
The DIS requires
that depressive
or other
mental
symptoms
should
not be asso-
1) being
than
to
pay
for
the
kind
enough
described
money
to pay for
by Golding
(20),
positive
item
any
negative
was
endorsed
item
was
of
food
I should
have”
the kind
of clothes
low financial
strain
and
high
and
“I
have
I should
have. “ As
was present
if any
financial
strain
was
present
if
endorsed.
6. Talked
to someone
about
most
personal
problems,
fears,
and
hopes
in previous
6 months.
Respondents
were asked
to answer
yes
or no to the question,
“In the last 6 months,
is there
someone
you
have talked
to about
sour most personal
problems,
hopes,
or fears?”
7. Any day u’ith restricted
activity
due to physical
illness
in previous
2 weeks.
Respondents
were
asked,
“During
the last 2 weeks,
how
many
days
did physical
illness
or your
physical
condition
make
you cut down
on things
you would
like to do, such as getting
around
or having
visitors?”
Scores
were
dichotomized
as present
or absent.
8. Any day in bed due to physical
illness
in previous
2 weeks.
Respondents
were asked,
“ During
the last 2 weeks,
how many
days did
you
stay
in bed
all
or most
of the
day
because
of feeling
physical
pain
or illness?”
Scoring
was dichotomized
as present
or absent.
9. Any chronic
limitation
in physical
or job functioning
due to
physical
illness.
Respondents
were
asked
to answer
yes or no to
seven
items,
similar
to the Hispanic
Health
and Nutrition
Examination
Survey
( 1 7, I 8): “ Does your physical
health
or physical
condition
make
it difficult
for you
to carry
out your
daily
activities?”
“ Does
your
physical
health
or physical
condition
keep
you from
taking
the
or physical
“Does
your
kind
ofjoh
condition
physical
you
limit
health
would
like?”
the work
or physical
“Does
your
you can
condition
do
physical
on
make
health
your
job?”
it difficult
for you to get around
(outside
the home,
in this place)?”
“Does
your physical
health
or physical
condition
prevent
you from getting
the kind
of job you would
like?”
“Are
you able to dress
yourself
completely,
or do you need help?”
and “Do you have any problems
going
to the bathroom,
like not being
able to get there
in time,
or
needing
help to get there?”
Due to skewing
of responses,
scoring
was
dichotomized
as present
if one or more
items
were
endorsed
or
absent
if no items
were
endorsed.
1 0. General
health status. Respondents
were asked
one item from
the Rand
Health
Insurance
Experiment:
“Would
you say that your
health is excellent,
good,
fair, or poor?”
This
item was
scored
as
1=excellent,
2=good,
3=fair,
or 4=poor.
Am
]
Psychiatry
1 53:1
1 November
,
1996
JUDD,
TABLE 1. Demographic Characteristics
of 2,393 Subjects in the Los An geles Site of the NIMH
by Depression Status Into Three Mutually Exclusive Groups
Subjects
With
Depressive
Symptoms
Characteristic
No
or
Married
Single
with
without
ET AL.
Area Survey Divided
Depressive
Symptoms
With
Subjects
(N=270)
Depression
N
%
Major
(N=102)
N
%
1,014
1,007
50.2
49.8
96
174
35.6
64.4
35
67
34.3
65.7
children
children
children
575
467
242
children
718
28.5
23.1
12.0
35.5
52
51
43
122
19.3
18.9
15.9
45.2
14
12
24
52
13.7
11.8
23.5
51.0
433
643
21.4
31.8
78
48
28.9
17.8
27
15
26.5
945
46.8
144
53.3
60
58.8
Mean
SD
Mean
SD
Mean
SD
41.33
17.06
37.29
14.88
36.18
11.00
12.14
3.83
12.53
without
Single
Subsyndromal
%
Catchment
WELLS,
With
Subjects
Disorder
(N=2,021)
N
Gender
Male
Female
Marital
status
Married
with
Epidemiologic
PAULUS,
Ethnicity
Native
Hispanic
Immigrant
Hispanic
Non-Hispanic
Age
white
(sears)
Education
(total
Description
been
sion,
defined
toms
in respondents
with
by the
those
no
NIMH Epidemiologic Catchment Area Survey
of the Los Angeles
ECA site sample
detail
elsewhere
(IS,
16). Using
the
we divided
the 2,393
subjects
into three mutually
those
with subsyndromal
symptomatic
depresin
DIS data,
subgroups:
greater
presence
not
with
of at least
meeting
major
depressive
criteria
depression
disorder
two
or
or more
for major
as defined
depressive
symp-
depression
or dys-
by DSM-III;
and
Domain
symptoms.
High
a 12-month
period,
270 subjects
met criteria
for subsyndromal
depressive
symptoms,
102 met DSM-III
criteria
for major
depression,
and 2,02 1 were assigned
to the subgroup
subjects
with no
depressive
disorder
or symptoms.
The demographic
characteristics
of
the three groups
are given
in table
1. The groups
with subsyndromal
High
symptoms
two-to-one
preponderance
and
depressive
disorder
major
or
had
depression
of women
symptoms
to men,
had
and
approximately
the
a balanced
group
sex
a
with
no
to the complex
Due
two
the effects
of depression
of
variance
ables
status
design
and
approaches
on functional
(MANOVA)
characterizing
pression
sampling
statistical
distinct
(SAS)
(i.e.,
no
status.
was
functional
carried
status
depressive
the multiple
were used
out
that
as dependent
disorder
comparisons
to determine
First, a multiple
or
included
measures
symptoms,
functioning
a0
means
effect
of
predictions
Am
]
for
Psychiatry
the prevalence
153:1
sampling
of disturbances
1, November
design
was
used
in functional
1996
to
status
obtain
due
Taylor
each
was
calculated
(lifetime
multistate
or symptoms
den
<0.0001
10.72
1.52
<0.000l
n.s.
13.67
27.46
21.95
<0.0001
<0.0001
<0.0001
28.39
<0.0001
19.95
<0.000
31.93
<0.0001
<0.0001
(N=2,021),
or major
subsyndromal
status
(21,
the adjustment
linearization,
dependent
1
subsyndromal
de-
depression
(N=l02).
depressive
symptoms
and
was
variable,
22).
of
The
STATA
standard
used
a separate
for
error
this
Huber
statistical
packby
calculations
regression
analysis.
regression
analysis
(22).
Candidate
variables
included
education,
age,
status,
ethnicity,
gender,
current
chronic
medical
condition,
mental
disorder,
comorbid
alcohol
use disorder
( I 2-month
and lifetime),
and the three
subsyndromal
morbid
22.32
depression
were significantly
higher
(indicating
than
the scores
of subjects
with
no depressive
(p<O.OS,
Tukey’s
Studentized
Range
test).
ofdepression
enables
For
sis
the
p
21.16
disorder
with major
impairment)
or symptoms
which
illness1’
status1’
depressive
subjects
greater
disorder
all
subsyndro-
2 weeksh
pressive
symptoms
(N=270),
hThe scores
of subjects
with
age,
de-
in previous
due to physical
health
to the
van-
illness
General
marital
comorbid
for
irnitahilityb
Any day in bed due to physical
illness
in
previous
2 weeksb
Any chronic
limitation
in physical
or job
analysis
and
social
mal depressive
symptoms,
or major depression)
as the independent
measure
to determine
whether
depression
status
had a significant
effect on functional
status.
This
analysis
included
one-way
ANOVAs
and Tukey’s
Studentized
Range
post hoc comparisons
to determine
the effect of depression
status
on the individual
variables
charactenizing functional
status.
Second,
a multinomial
logistic
regression
analyadapted
(MANOVA)
F (df=2,
2255)
Function
physical
Methods
involved,
of Depression
Status
household
strainh
High social
contacts
Talked
to someone
about
most personal
problems,
fears, and hopes in last 6 months1’
Major
financial
loss in last 6 monthsh
High
financial
stnainb
Any day with restricted
activity
due to
distribution.
Compared
with the group
with no depressive
disorder
or symptoms,
the group
with subsyndromal
depressive
symptoms
had fewer
marned
people
and the group
with
major
depression
even
fewer.
The
groups
with subsyndromal
depressive
symptoms
and major
depression were,
on the average,
younger
and better
educated.
There
was a
slightly
higher
prevalence
of non-Hispanic
whites
and native
Hispanics and fewer
immigrant
Hispanics
in the groups
with subsyndromal
depressive
symptoms
and major
depression
than
in the group
with no
depressive
disorder
or symptoms.
Statistical
of
Effect
those
During
depressive
3.84
TABLE 2. Effect of Depression Statusa on 10 Domains of Function
and Well-Being
in 2,393
Subjects in the Los Angeles Site of the
characteristics
described
12-month
exclusive
thymia;
4.58
of Sample
Demographic
have
11.34
years)
14.7
mental
only),
(lifetime
comorbid
diagnostic
depressive
disorder
bipolar
only),
drug use disorder
(12-month
and lifetime),
groups (no depressive
disorder
or symptoms,
symptoms,
and major depression).
The covariable
disorder
cannabis
included
(lifetime
abuse
or
manic
only),
episode,
atypical
dependence,
dysthymia
bipolar
disor-
schizophrenia,
1413
SOCIOECONOMIC
BURDEN
OF DEPRESSION
TABLE 3. Adjusted Percentagesa of Subjects in the Los Angeles Site of the NIMH Epidemio logic Catchment
Status, Who Reported Impairment in Social, Financial, and Physical or Role Function
Subjects
With
Subjects
With
No
Depressive
or
Depressive
Symptoms
(N=2,021)
of
%
SE
1
41
3
37
5
None
64
1
69
3
73
5
None
14
1
17
3
25
4
Between
SE
(13
50
about
Significant
to someone
most
personal
problems,
fears,
hopes
in last
Financial
and
6 months
status
Major
financial
last
6 months
loss
in
subjects
those
with
(t=3.1,
High
financial
strain
53
l
61
3
69
df=2289,
Between
6
1
18
2
24
3
between
restricted
activity
aPercentages
are
comorhid
adjusted
alcohol
hConfirmed
for
use
10
1
18
2
16
3
S
I
11
2
11
3
education,
age,
( 12-month
disorder
byTukey’s
Studentized
depression
were
major
Range
higher
marital
status,
and lifetime),
test
(pczO.OS).
(indicating
ethnicity,
gender,
and comorbid
In allcases,
greater
the
impairment)
than
for
contacts
the
following
(high
versus
(high
l13l
sonal
problems,
due
to physical
illness,
in physical
cellent,
good,
fair,
that
depression
tested
by
Tukey’s
poor).
The
to a t score
days
Three
already
status
(t=2.2,
df=2253,
subjects
with
with no depressive
(t=2.3,
df=2253,
tests
for
and
Only
domains
alpha
hoc
<0.05.
of disturbances
groups
depressive
predictions
comparison.
affects
rates
the
by using
Specifically,
by post
followed
obtained
for
subsyndromal
tested
the
standard
comparisons
limi-
status
(ex-
were
used
the
two
the
hypothesis
statis-
of function
Moreover,
errors
was
using
specific
in domains
with
no depressive
symptoms,
and
those
chronic
comparisons
of func-
disorder
or
major
depreswere
which
subjected
were
chronic
medical
conditions,
(12-month
and
with subsyndromal
of subjects
from
dis-
=O.O3(h
major
depression
disorder
or
p=0.02(t’
major
depression
with
no
by using
the regression
comorhid
lifetime).
depressive
depressive
Tukey’s
analysis
symptoms
disorder
test
and
mental
or
the
disorders,
and
subjects
symptoms.
t score
are reported
differences
as significant
ohin this
ir-
days
analyses
those
depressive
depressive
investigation.
activity
health
statistical
here.
significantly
Range
illness,
and general
were
social
current
use disorder
ofsubjects
tamed
low), social
about
per-
restricted
to physical
different
here
obtained:
(high versus
to someone
strain,
due
described
of prevalence
average
financial
hypotheses.
advanced
Studentized
tion
were
symptoms,
1414
bed
a MANOVA
predictions
sion.
loss,
were
strain
talking
or job functioning,
to address
the two
The hypotheses
approaches
variables
household
low l12l),
versus
financial
tations
tical
dependent
low),
drug
scores
to be significant
schizophreniform
disorder,
obsessive-compulsive
disorder,
phobias,
somatization
disorder,
panic
disorder,
antisocial
personality
disorder
(lifetime
only),
and anorexia
nervosa
(lifetime
only).
Based
on the calculated
logistic
regression
model,
predicted
prevaritabiliry
no
disorder
or
symptoms
(t=6.2,
df=2305,
p’zO.000
I)
Between
subjects
with subsyndromal
depressive
symptoms
and those with no depressive
disorder or symptoms
(t=3.7,
df=2296,
p<O.0001)
Between
subjects
with subsyndromal
depressive
symptoms
and those
with no depressive
disorder
or symptoms
(t=2.9,
df=2300,
p<0.004)1’
and between
subjects
with major
depression
and those with () depressive
disorder
or
symptoms
(t=2.3,
df=2300,
=O.O3)h
weeks
lences
with
and between
subjects with
and those with no depressive
day due to physical
illness in last 2 weeks
Any bed day due to physical illness in previous
2
and
or symptoms
Between
subjects
with subsyndromal
depressive
symptoms
and those
with no depressive
disorder
or symptoms
(t=4.3,
df=2305,
p<O.OOOl)i
illness
Any
depression
disorder
subsyndromal
those
or symptoms
symptoms
9
with
and
and those
Physical
or role limitations
Any chronic
limitations
in physical
or job functioning
due to physical
major
p=0.002)”
subjects
order
and
with
no depressive
symptoms
with
Difference
per month)
Talked
your
Depression
(N=102)
SE
%
contact
contacts
Major
Symptoms
(N=270)
()/
Function
Social
function
High social
Subjects
With
Subsyndromal
Disorder
Category
Area Survey, Divided by Depression
found
RESULTS
The MANOVA
for effects
of depression
status
on domains
of function
revealed
that significant
effects
were
present
for nine of the 1 0 domains
of everyday
function
(table 2).
Table
3 includes
results
of regression
analyses
for
seven of the 1 0 functional
domains.
Subjects
with major
depression
differed
significantly
from those with no depressive
disorder
or symptoms
on two financial
status
domains
and from
the group
with
subsyndromal
depressive
symptoms
on one financial
status
domain.
Significantly
more
of the subjects
with subsyndromal
depressive
symptoms
than
of those
with
no depressive
Am
J
Psychiatry
1 53:1
1 November
,
1996
JUDD,
disorder
on symptoms
physical
or
role
of the subjects
no depressive
in two
reported
limitation
of these
the
more
domains.
than of those with
reported
disability
were
no significant
difwith
major
depression
depressive
symptoms
in
There
domains
Compared
of the
No
regression
analy-
of household
strain
and social
with
the subjects
who
had
no
pressive
disorder
or symptoms,
the subjects
with
subsyndromal
reported
high levels of household
tability.
in three
significantly
with major
depression
disorder
or symptoms
ferences
between
the subjects
and those
with subsyndromal
these domains
of function.
Figure
1 compares
the results
ses on
tability.
impairment
domains;
significant
significantly
depressive
strain
and
differences
were
mnde-
more
of
symptoms
social
irni-
found
between
symptoms
(t=4.6,
df=2305,
The
full
Los
equal
Angeles
numbers
ECA
of
sample
included
Hispanic
and
appnoxi-
non-Hispanic
whites,
which
(IS,
16). Our
tested
in more
ported
findings
differs
from the general
U.S. population
conclusions
here,
therefore,
should
be
representative
samples;
however,
the neapply
for both
Hispanic
and non-His-
panic
and
groups,
ethnicity
was
controlled
for
in
the
analysis.
Significantly
more
subjects
with
subsyndromal
depressive
symptoms
than subjects
with no depressive
disorder
or symptoms
reported
disability
on seven
of 10
domains
of function:
high social irritability
(62%),
high
household
strain
(45%),
high financial
strain
(63%),
restricted
activity
days due to physical
illness
(18%),
bed days due to physical
illness
(11%),
chronic
limitation in physical
or job functioning
(18%),
and self-nating of health
as poor.
Although
one primary
study goal was to evaluate
associations
between
“pure”
subsyndromal
depression
and major
depression
and functional
impairment,
some
identified
impairment
could
be
associated
conditions.
The logistic
regression
signed
to control
for the comorbid
and current
( 12-month)
alcohol
and
Am
J Psychiatry
1 53: 1 1 , November
with
other
analysis
was deeffects
of lifetime
drug use disorders,
1 996
ET
AL.
EINO
DISORDER
OR
DSUBSYNDROMAL
DMAJOR
70
0
SYMPTOMS
DEPRESSIVE
DEPRESSIVE
DEPRESSION
(N=2,021)
SYMPTOMS
(N=270)
(N=102)
z
I-
60
0
w
50
40
w
0
z
30
20
10
[
c
HIGH SOCIAL
IRRITABILITY
HIGH HOUSEHOLD
STRAINb
c
“Percentages
are adjusted
for education,
age, marital
status,
ethnicity,
gender,
current
chronic
comorbid
medical
condition,
comorbid
mental disorder,
comorhid
alcohol
use disorder
( 12-month
and lifetime)
and comorhid
drug use disorder
( 12-month
and lifetime).
“Significant
difference
between
the group
with subsyndromal
depressive symptoms
and the other
two groups
(t=3.1,
df=2270,
p<O.OO2)
(significant
according
to Tukey’s
Studentized
Range
test, p<O.OS).
‘Significant
difference
between
the group
with subsyndromal
depressive symptoms
and the other
two groups
(t=3.3,
df=2274,
p<O.OO1)
(significant
according
to
Tukey’s
Studentized
Range
test,
p<0.OS).
mental
disorders,
and current
chronic
medical
conditions.
An additional
control
in reducing
the influence
of
comonbid
conditions
was provided
by the DIS itself,
which
requires
that when
a depressive
symptom
is recorded
it not
be associated
with a medical
illness
or be
a side effect
from
alcohol
or drug
use or disorders.
These
data
support
and extend
findings
from
six pre-
p=O.OS).
DISCUSSION
mately
WELLS,
FIGURE 1. Adjusted Percentages of Subjects With No Depressive
Disorder or Symptoms, Subsyndromal
Depressive Symptoms, or Major Depression Reporting High Levels of Household Strain or Social
Irrithbility’
w
the subjects
with
major
depression
and those
with no
depressive
disorder
or those with subsyndnomal
depressive symptoms.
Estimates
of general
health
status.
On the 4-point
health
scale (l=poon
to 4=excellent),
the subjects
with
subsyndromal
depressive
symptoms
gave themselves
a
significantly
poorer
mean
health
rating
(mean=2.91,
SE=0.05)
than did the subjects
with no depressive
disorder
or symptoms
(mean=3.13,
SE=0.02)
(t=4.S,
df=
2289,
p<O.OOI
). Respondents
with
major
depression
also gave themselves
a significantly
poorer
mean
health
rating
(mean=2.67,
SE=0.09)
than
did those
with
no
depressive
disorder
or symptoms
(t=5.03,
df=2289,
p<
0.0001
). The
respondents
with major
depression
gave
themselves
a significantly
poorer
mean
health
rating
than
did the subjects
with
subsyndromal
depressive
PAULUS,
vious
studies
reporting
significant
associations
between
subthreshold
depressive
symptoms
and greater
impairment
in everyday
function
(1-6).
There
is now strong
evidence
that subsyndromal
depressive
symptoms
are
associated
with substantial
disability
in many
spheres
of human
activity,
placing
notable
burdens
on individuals and the society.
Compared
with respondents
who had no depressive
disorder
on symptoms,
individuals
with major
depression
reported
mains.
impairment
Significantly
sion
reported
(25%),
major
more
reported
in
more
five
of
subjects
financial
loss
10
functional
with
major
in the
last
do-
depres6 months
from high financial
strain
(69%),
more
on job limitations
due to illness
(24%),
more
reported
bed days (1 1%), and their selfratings
of health
were poorest.
These
data
strengthen
the confluence
of scientific
evidence
that major
depression is associated
unquestionably
with severe
pervasive
disability
has
in many
a palpable
2, 4, 7, 23).
jon
suffered
physical
depression
aspects
impact
For
example,
reported
of
daily
on society
24%
functioning,
and
of the
limitations
the
which
economy
subjects
in physical
with
( 1,
ma-
on job
1415
SOCIOECONOMIC
BURDEN
OF
DEPRESSION
functioning,
and this is consistent
with annual
cost estimates
for mood
disorders,
which
found
that the two
largest
cost categories
were increased
absenteeism
and
reduced
worker
productivity
(23).
No
large consistent
differences
in impairment
in the
domains
of function
emerged
between
the subjects
with
subsyndromal
depressive
symptoms
and those with majon depression,
apart
from
their
self-ratings
of their
health
status,
which
could
have
occurred
by chance
alone.
This
does
not
support
one
of
our
a priori
hy-
potheses-that
more
subjects
with
major
depression
than
those
with
subsyndromal
depressive
symptoms
will report
impairment.
However,
every other
study using different
outcome
measures
than
those
reported
here, such as absenteeism,
use of health
services,
use of
public
assistance,
and suicide
attempts,
have found
significantly
more adverse
outcome
associated
with major
depression
than
subsyndromal
depressive
symptoms
(1-6).
We did find that higher
absolute
percentages
of
subjects
with
major
depression
than
of
subjects
Further,
the
functional
measurements
used
with
with
and
which
found
that
the
impairments
in
subjects
subsyndromal
depressive
symptoms
and those
major
depression
are qualitatively
comparable
more
similar
to each
other
than
they are to any
impairment
found
in
subjects
without
(24),
we
characteristics
found
of
approximately
sample
that
the
demographic
subjects
with
subsyndromal
one
met
criteria
10 people
in
for
of
subsyndromal
depressive
in this
subsyndromal
symptoms
major
depression
are needed,
vestigations,
because
no
data
that
subsyndromal
depressive
treatment
that
or if they
effective
should
de-
symptoms
may
tional
disability
venely
disabling
of
not
at all.
It is possible
subsyndromal
only
reduce
identified
hut
future
in unipolar
especially
treatment
inare available
indicating
symptoms
respond
to
be treated
treatment
depressive
the
associated
also
episodes
func-
prevent
of major
more
se-
depression
or
dysthymia.
REFERENCES
I . \X7ells KB, Stevart
A, Hays RI),
‘vI, Berry
5, (ireeiifield
S. \X7arcJ:
of depressed
JAMA
2.
Broadhead
ability
WE,
Blazer
and
survey.
The
from
DG,
days
MA,
Rogers
W, 1)aniels
and well-being
functioning
the
(george
lost
JAMA
from
1990;
J, Weissnun
Johnson
social
Medical
Outcomes
Study.
Tse
1K:
1)cprcssion,
in a prospective
Klcrn#{236}an (ii.:
with
associated
I 992;
JAMA
1K,
vork
dis-
cpideiiio-
264:2524-2528
\t’\t,
morbidity
conimunity.
4.
results
ltirnam
262:914-919
days
logic
3.
patients:
1989;
267:
Service
and
utilization
depressive
svmptonis
in
the
1478-1483
RL, Kroenke
K, Linzcr
NI, Hahn
SR. WilliamsJB,
deGruy
Brodv
D, I)avics
i\I: Health-related
quality
of life in priinarv
care
patients
vith
mental
disorders:
results
from
the
PRIN1E-Nil)
1 000 study.
JANIA
I 995;
274: 1 5 1 1-15 17
I-lowarth
E, Johnson
J, Klerman
UI.,
\Veissman
i’\I: I)epressive
Spitzer
FV III,
5.
syiiptoms
6.
7.
as
relative
and
major
depression.
Arch
Judd
1.1., Rapaport
1,
toniatic
1)epression.
‘1iiitz
J, Nlintz
TH,
and functional
slOil
attributable
(jen
risk
Psychiatry
factors
1 992;
for
49:8
first-onset
17-823
Paulus
NI, Brown
J: Suhsvndromal
Svnipj (;lii
Psychiatry
I 994
55(April
suppl(
Aruda
Nil, Hwang
SS: Treatment
of deprescapacity
to work.
Arch
Gen Psychiatry
1 992;
49:761-768
8.
Regier
DA,
\tversJK,
RE,
Eaton
iTient
Area
and
depressive
symptoms
significantly
more
closely
resembled those of subjects
with major
depression
than those
of nondepressed
medical
outpatients.
For example,
we
found
that 41 % of outpatients
with subsyndnomal
depnessive
symptoms
had family
histories
of depression,
a
rate nearly
as high as that for subjects
with major
depression
(59%).
We have proposed
that subsyndromal
depressive
symptoms
represent,
respectively,
either
prodromal
manifestations
or
residual
depressive
symptoms
from
impending
or incompletely
resolved
major
depression
on dysthymia
(25). Subsyndromal
depressive
symptoms
may also be a low-grade
chronic
condition,
placing
patients
at higher
risk for future
major
depressive
episodes
(25). This has led us to conclude
that
suhsyndnomal
depressive
symptoms
are a clinically
meaningful,
suhthreshold
variant
of unipolar
major
depression.
1416
ies
subsyndromal
depressive
symptoms
on major
depression.
It is not fully clear what
subsyndromal
symptomatic
depression
represents
clinically,
but we do know
from
the studies
of Broadhead
et al. (2) and Howarth
et at.
(5) that subthreshold
depressive
symptoms
are associated with greaten
risk for future
major
depressive
episodes.
In a large sample
of medical
and psychiatric
outpatients
clinical
that
pressive
symptoms
during
1 year,
when
combined
with
the associated
functional
impairment
reported
by us
and others
(1-6),
supports
the conclusion
that subsyndromal
depressive
symptoms
are a clinical
condition
of
public
health
significance.
Prospective,
controlled
stud-
in this
study
were primarily
dichotomous;
it is apparent
that
there are differences
between
subsyndrornal
depressive
symptoms
and major
depression
that we did not have
the power
to detect.
It should
be noted
that our findings
are also consistent
with those of previous
investigations
(1-4),
fact
with
subsyndnomal
depressive
symptoms
reported
impairment
in five of 10 functional
domains,
hut the differences were not sufficiently
robust
to survive
formal
testing.
The
community
study
Kranier
M,
EN,
Robins
WW,
locke
LIZ: The NI’IH
Program:
historical
context,
population
characteristics.
Blazer
1)G,
Epidemiologic
major
Gen
Arch
Hough
Catch-
and
objectives,
Psychiatry
I 984;
4 1
934-941
9.
Eaton
WW,
(;eorge
L,
Holzer
CE III, Von Korff
1,
Boyd
JH, Kessler
Burnam
sign
of
and
measurement
the
Anthony
LG,
A,
Epidemiologic
Catchment
of error.
Arch
Area
Gen
JC,
Locke
Helzer
lIZ:
surveys:
Psychiatry
the
I 984;
JE,
The
de-
control
4 1 :942-
948
10.
11.
12.
Eaton
W, Kessler
1. (eds(:
Epideniiologie
Field
Methods
in Psychiatrv:
The NINIH
Epidemiologic
Catchment
Area
Program.
Orlando,
Fla, Acadeniic
Press,
I 985
Robins
EN, Helzer
JE, Cottler
1., (iolding
E: National
Institute
of Mental
Health
1)iagnostic
Interview
Schedule,
version
III, revised.
St Louis,
Washington
University,
Department
of Psvchiatry, 1989
Rohiiis
EN, Helzer
JE, Croughan
l. Ratcliff
KS: The National
Institute
of \lental
Health
Diagnostic
Interview
Schedule:
its history,
38
I 3.
characteristics,
Karno
‘v1, Burnam
velopnient
tote
of
of the
Niental
Psychiatry
14.
and
validity.
Arch
NIA,
Escohar
J,
Ccii
Psychiatry
I 98 1 ; 38:
1-389
Burnam
Spanish-language
Health
1 983;
isIA,
40:
Karim
Am
1)iagnostic
1 1 83-1
NI, Hoogh
j Psychiatry
Hough
RL,
version
of the
Interview
Eaton
\V\V:
National
Schedule.
I)eInsti-
Arch
Gen
188
RI.,
Escoharjl,
Forsythe
1 53: 1 1 , November
AB: The
1996
JUDD,
I 9.
Spanish
Diagnostic
Interview
Schedule:
reliability
and comparison with clinical
diagnoses.
Arch Gen Psychiatry
1 983; 40: 1 1891196
Wells KB, Golding
JM, Burnam
MA: Psychiatric
disorder
and
limitations
in physical
functioning
in a sample
of the Los Angeles
general
population.
Am J Psychiatry
1988;
145:712-717
Wells
KB, Golding
JM,
Burnam
MA:
Psychiatric
disorder
in a
sample
of the general
population
with and without
chronic
medical conditions.
AmJ
Psychiatry
1988;
145:976-981
Data
Collection
Forms
of the Hispanic
HANES:
Adult Sample
Person
Questionnaire
Form PHS 6206.
Washington,
DC, US Department
of Health
and Human
Services,
1982
Wilder
CS: Health
Characteristics
of Persons
With
Chronic
Activity
Limitation:
Data
From
the National
Health
Survey,
Series
10, Number
I 12: Department
of Health,
Education
and Welfare
Publication
HRA
77-1539.
Rockville,
Md, National
Center
for
Health Statistics,
1976
Pearlin
LI, Lieberman
MA: Social
sources
of emotional
distress.
20.
Golding
Am
J
15.
16.
17.
18.
Res
Community
JM:
Psychiatry
Ment
Division
Health
1979;
of household
I 53: 1 1 November
,
1:217-248
labor,
strain
1996
and
depressive
21.
PAULUS,
WELLS,
symptoms
among
Mexican
Americans
and non-Hispanic
Psychol Women
Q 1990; 14:103-1
17
Schluchter
MD: Analysis
of incomplete
multivariate
linear
models
with
structured
covariance
matrices.
ET AL.
whites.
data
Stat
using
Med
1988;
22.
23.
24.
25.
7:317-324
PJ: The behavior
of maximum
likelihood
estimates
under
nonstandard
conditions,
in Fifth
Berkeley
Symposium,
1965.
Berkeley,
University
of California
Press,
1967,
pp 22 1-233
Greenberg
PE, Stiglin
L, Finkelstein
5, Berndt
E: The economic
burden
of depression
in 1990.
J Clin Psychiatry
1993;
54:405418
Sherbourne
CD, Wells
KB, Hays
RD, Rogers
W, Burnam
MA,
Judd LL: Subthreshold
depression
and depressive
disorder:
clinical characteristics
of general
medical
and mental
health
specialty
outpatients.
Am J Psychiatry
1994;
151:1777-1784
Judd
L: Clinical
characteristics
of subsyndromal
symptomatic
depression,
in Critical
Issues
in the Treatment
of Affective
Disorders,
vol 9. Edited
by Langer
SZ, Brunello
N, Racagni
G,
Mendlewicz
J. Basel, Switzerland,
Karger (International
Academy for Biomedical
and Drug
Research),
1994,
pp 67-74
Huber
1417